Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Childhood ADHD and Executive Functioning: Unique Predictions of Early Adolescent Depression

Childhood ADHD and Executive Functioning: Unique Predictions of Early Adolescent Depression Given the increasing prevalence of adolescent depression, identification of its early predictors and elucidation of the mecha- nisms underlying its individual differences is imperative. Controlling for baseline executive functioning (EF), we tested separate ADHD dimensions (i.e., inattention, hyperactivity-impulsivity) as independent predictors of early adolescent depres- sion, including temporally-ordered causal mediation by academic functioning and social problems, using structural equation modeling. At baseline, participants consisted of 216 children (67% male) ages 6–9 years old with (n = 112) and without (n = 104) ADHD who subsequently completed Wave 2 and 3 follow-ups approximately two and four years later, respectively. Predictors consisted of separate parent and teacher ratings of childhood ADHD and laboratory-based assessments of key EF domains. At Wave 2, parents and teachers completed normed rating scales of youth academic and social functioning; youth completed standardized assessments of academic achievement. At Wave 3, youth self-reported depression. Baseline inattention positively predicted early adolescent depression whereas childhood hyperactivity-impulsivity and EF did not. Neither academic nor social functioning significantly mediated predictions of depression from baseline ADHD and EF. We consider prediction of early adolescent depression from inattention, including directions for future intervention and preven- tion research. Keywords ADHD · Executive functioning · Depression · SEM In the United States, the prevalence of major depressive increased precipitously since 2010 (Twenge et al., 2017), disorder has risen consistently since 1940 and currently including suicide being the second leading cause of death incurs over $210 billion annually, a 21% increase since 2005 among 10–24-year-old individuals (Center for Disease Con- (Greenberg et al., 2015). The increase in prevalence coin- trol and Prevention, 2016). To reduce its considerable clini- cides with an earlier age of onset (Birmaher & Brent, 2007): cal and public health burden, characterizing predictors of 2% of children and up to 8% of adolescents meet diagnos- early adolescent depression is a priority. tic criteria for major depressive disorder with an additional Although understudied relative to co-occurring external- 5–10% of youth experiencing subclinical symptoms. Cru- izing problems, cross-sectional associations of ADHD with cially, 30% of youth with major depressive disorder expe- depression (i.e., heterotypic comorbidity) are well-established rienced suicidal ideation in the past 12 months (Avenevoli (Humphreys et al., 2013; Meinzer et al., 2014; Seymour et al., et  al., 2015). Moreover, depression and suicidality have 2014). Further, even with control of demographic and clinical factors (e.g., maternal depression), childhood and adolescent ADHD each uniquely predict depression through young adult- * Michelle C. Fenesy hood (Chronis-Tuscano et al., 2010; Meinzer et al., 2016), and * Steve S. Lee children with ADHD frequently experience recurrent depres- stevelee@psych.ucla.edu sive episodes (Chronis-Tuscano et al., 2010). ADHD is opti- mally conceptualized as the quantitative extreme of inattention Department of Child & Adolescent Psychiatry, New York- Presbyterian Hospital, Columbia University Irving Medical and hyperactivity-impulsivity, but ADHD is typically exam- Center, 622 W. 168th St., New York, NY 10032, USA ined dichotomously, thereby preventing specific inferences Department of Psychology, University of California, about the relative association of inattention and hyperactiv- Box 951563, 1285 Franz Hall, Los Angeles, CA 90095-1563, ity-impulsivity with respect to depression. In key exceptions, USA Vol.:(0123456789) 1 3 754 Research on Child and Adolescent Psychopathology (2022) 50:753–770 inattention—not hyperactivity-impulsivity—was uniquely consider the prospective association of EF with depression associated with depression in 5–10-year-old youth (Fenesy & during the transition from childhood to early adolescence Lee, 2019; Humphreys et al., 2013), although hyperactivity- using latent variable approaches. Specifically, this evidence impulsivity predicted depression indirectly via emotion regu- will inform the potential need/utility of EF and/or behavio- lation (Seymour et al., 2014). In addition to the independent ral interventions prior to depression onset in adolescence. prediction of depression from separable dimensions of ADHD, To clarify whether ADHD and EF independently predict there is a pressing need to identify additional factors exacerbat- youth depression, ADHD and EF must be examined con- ing or mitigating risk. currently. ADHD principally reflects executive dysfunction Several lines of evidence suggest the plausibility of exec- (Barkley, 1997; Castellanos & Tannock, 2002) with meta- utive functioning (EF) as a risk factor with respect to youth analytic evidence that youth with ADHD exhibit impaired depression. Although a precise definition of EF is somewhat EF in community-based and clinical samples (Willcutt intractable (Barkley, 2012), there is general consensus that et  al., 2005). Although EF is commonly proposed as a EF encapsulates related cognitive processes (i.e., inhibitory cognitive endophenotype or risk factor for ADHD (Gau & control, working memory, set shifting) associated with the Shang, 2010; Nigg et al., 2004), controversy remains as prefrontal cortex that support goal-directed behaviors (Bar- there is considerable heterogeneity in EF among individu- kley, 2012; Miyake et al., 2000; Nigg, 2017). Importantly, als with ADHD and not all children with a diagnosis of individuals with depression consistently exhibit poor emo- ADHD exhibit EF deficits (Kofler et al., 2019; Willcutt, tion regulation, which is directly supported by EF among et al., 2005). Further, the broader field of EF research seeks adolescents and adults (Gyurak et al., 2012; Joormann & to clarify whether individual differences in EF represent Gotlib, 2010; Wagner et al., 2015). Further, emotion regula- a risk-factor for psychopathology, a correlate of psycho- tion tasks are sensitive to differential patterns of activation pathology, or a result of psychopathology across multiple in prefrontal cortex regions associated with EF that suppress disorders (Snyder et al., 2015). Therefore, it is reasonable emotional responses from the limbic system (Zelazo & Cun- to evaluate EF as a correlate of ADHD. The present study ningham, 2007). Moreover, rumination (i.e., engaging in aims to test the prospective prediction of depression from repetitive thinking patterns that amplify emotions) is central ADHD and EF, as controlling for comorbidities is recom- to the onset and maintenance of depression (McLaughlin mended when evaluating EF in relation to depressive symp- & Nolen-Hoeksema, 2011); elevated rumination mediated toms (McClintock et al., 2010). the cross-sectional association between impaired EF (i.e., Relatively few studies simultaneously examined the set shifting) and depressive symptoms in typically develop- association of ADHD and EF with youth depression. ing adolescents (Dickson et al., 2017). Finally, individuals For example, working memory, a key dimension of EF, with poor EF often experience difficulties with goal attain- was impaired among adolescents with ADHD + depres- ment and planning across domains, thus catalyzing stress sion relative to youth with depression alone (Roy et al., generation and potentially subsequent depression (Snyder 2017), but depression and ADHD were exclusively youth & Hankin, 2016). Overall, there is growing evidence that self-reported, despite the utility of multiple informants EF is correlated with youth depression, although findings (Martel et al., 2017). Moreover, controlling for inatten- are frequently based on cross-sectional designs (e.g., Favre tion and hyperactivity-impulsivity, EF dimensions (i.e., et al., 2008) and consist of broad age ranges (e.g., child- working memory, mental flexibility, inhibition) were hood through adolescence), preventing specific inferences unrelated to parent- or self-reported depression among about EF as a correlate versus risk factor for early-adoles- 9–16-year-old youth (Øie et al., 2016). However, latent cent depression. Further, separable EF dimensions require variable approaches improve measurement error; latent rigorous measurement strategies to adequately capture its EF was positively associated with parent-rated depres- multidimensionality. Despite their empirical separability, sion among 5–10-year-old children (Fenesy & Lee, 2019). boundaries among EF dimensions may be less distinct in Among hospitalized inpatients, youth with depression and children, giving rise to a unified set of cognitive processes EF deficits had elevated ADHD compared to children with (Lee et al., 2013; Wiebe et al., 2011), thus necessitating depression alone (Weber et al., 2018), although directional careful methodological and developmental considerations. inferences were unclear, further underscoring the need for Indeed, when derived from a single latent variable, poor multi-method/informant longitudinal studies. preschool EF prospectively predicted early school-age In addition to examining the independent prediction depression symptoms, controlling for baseline depression of depression from childhood ADHD and EF, it is neces- (Nelson et al., 2018). However, this study did not extend sary to investigate ADHD diagnostic status as a moderator into early adolescence, an important limitation given the of the prospective association between childhood EF and precipitous increase in depression secondary to pubertal early adolescent depression to evaluate possible interactive onset (Avenevoli et al., 2015). There is a pressing need to effects. ADHD and EF interact to predict other key outcomes 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 755 such as academic achievement. For example, elevated inat- evaluating risk for depression. We also tested whether tentive symptoms and poor EF predicted special education academic and social functioning temporally mediated (Diamantopoulou et al., 2007). Thus, inclusion of ADHD predictions from inattention, hyperactivity-impulsivity, diagnostic status as a moderator will reveal whether the and EF. Consistent with recommendations for SEM, if prospective association between EF and depression varies initial models did not fit the data, we thoughtfully respec- according to ADHD diagnostic status. Testing independent ified the models (Weston & Gore Jr., 2006; Violato & and interactive prediction of depression from ADHD and EF Hecker, 2007.) We hypothesized that inattention would will elucidate their contributions, informing how specific positively predict whereas EF would inversely predict assessment and intervention approaches must be tailored. early adolescent depression. Lastly, we expected that aca- Beyond predictive models, innovations in intervention demic and social functioning would each significantly require elucidation of underlying mechanisms of predictions mediate predictions of adolescent depression from inat- of psychopathology from EF and ADHD (Meinzer & Chronis- tention, hyperactivity-impulsivity, and EF. Tuscano, 2017; Snyder et al., 2015; Zhou et al., 2012), a critical next step in EF research (Snyder et al., 2019). Both dual failure and competency-based models propose academic and social difficulties as key factors in the development of youth depres- Methods sion (Cole, 1991; Patterson & Stoolmiller, 1991). Whereas the dual failure model hypothesized that academic failure and peer Participants rejection mediated predictions of depression from early conduct problems (McCarty et al., 2008; Patterson & Stoolmiller, 1991), Participants were 216 children (67% male) with (n = 112) competency-based models contend that children internalize and without (n = 104) ADHD and their families. Drawn negative feedback from the environment, which adversely from a large metropolitan city in the Western U.S., children affects their self-esteem and increases vulnerability to depres- were recruited from referrals from pediatric offices, men- sion (Harter & Marold, 1994). Additionally, ADHD and EF tal health service agencies, and local schools. Inclusion each predicted academic achievement (Best et al., 2011; Miller criteria were English fluency and living with a biological et al., 2012) and social functioning (Diamantopoulou et al., caregiver at least halftime; exclusion criteria consisted of 2007; Huang-Pollock et al., 2009), representing plausible medi- an IQ below 70 or seizure, autism spectrum, or other neu- ators of depression from ADHD and EF. In particular, these rological condition. Table 1 summarizes key demographic theoretically-justified mediators should be subjected to strin- data for participants. The sample was ethnically diverse gent tests of causal mediation wherein constructs are tempo- (51.93% Caucasian; 8.33% African American; 12.04% rally ordered relative to predictors and outcomes (MacKinnon Hispanic; 3.70% Asian; 22.22% Mixed; 1.85% Other/ & Fairchild, 2009). Unknown) and 29.95% of families had an annual income Overall, the extant literature does not consider sepa- of $70,000 or less. At baseline (i.e., Wave 1), participat- rable dimensions of ADHD in conjunction with EF, and ing children were 6- to 9-years-old (M = 7.39, SD = 1.07). prospective studies are necessary to determine whether Families were invited to a follow-up approximately two EF represents a predictor versus correlate of early ado- years later (i.e., Wave 2) at which point youth ranged in lescent depression. Lastly, identification of underlying age from 7- to 13- years old (M = 9.68, SD = 1.27). A mechanisms is necessary to identify targets for interven- final follow-up (i.e., Wave 3) was conducted about two tion to reduce the risk of youth depression. To address years after Wave 2 when they were 9- to 15-years old these significant limitations hindering innovations in (M = 12.07, SD = 1.30). Data from all three waves were understanding the antecedents and mediators of early utilized in the current study. adolescent depression, the present study employed Of the 216 Wave 1 participants, 89.35% (n = 193) and structural equation modeling (SEM) to evaluate inat- 80.10% (n = 173) participated in the Wave 2 and 3 assess- tention, hyperactivity-impulsivity, and EF as independ- ment, respectively. There were no significant differences in ent, prospective predictors of early adolescent depres- child age, sex, race, or baseline ADHD symptoms between sion controlling for key covariates (see Data Analytic children who participated at Wave 1 and Wave 2 or Wave 1 Procedure). Additionally, we tested ADHD diagnostic and Wave 3 (ps > 0.15). Youth who participated at Waves 2 status as a moderator to examine whether the prospec- and/or 3 had a higher mean IQ (ps < 0.01) than those who tive prediction of depression from EF differs for chil- did not complete follow-ups. Overall, missing data ranged dren with and without ADHD. Determining whether the from approximately 72% on a youth self-rated depression prospective association between EF and depression var- measure to 0% on baseline ADHD data from a structured ries according to group status would inform clinicans diagnostic interview. Maximum likelihood estimation for whom the screening of EF is most important when addressed missing data (described further below). 1 3 756 Research on Child and Adolescent Psychopathology (2022) 50:753–770 quickly as possible to sequentially connect numbered cir- Procedures cles from 1 to 15 without making errors. Part B (TMT-B) involves alphabetically and numerically alternating between Families initially completed a phone screen to determine eligibility. Parents and teachers of eligible participants were connecting numbers 1 through 13 and letters A through L. The time (min) to complete TMT-B reflects set shifting, mailed rating scales of child functioning and were invited to a lab-based assessment conducted by well-trained gradu- with longer completion times indicating worse EF (Reitan & Wolfson, 1992). TMT-B significantly differentiated youth ate students in clinical psychology or B.A. level staff. After obtaining parental consent and youth assent, parents com- with ADHD and controls (Martel et al., 2007) and loaded onto a latent set shifting factor in children (Arán Filippetti pleted structured diagnostic interviews and rating scales to assess youth social-emotional functioning, whereas youth & Richaud, 2017). To aid in interpretation, completion time was reverse coded so that higher scores represented better completed standardized tests of cognition, EF, and aca- demic achievement as well as measures of social-emotional EF across all tests. Participants completed the Children’s Version of the functioning. Researchers requested that children complete a medication washout for the day of testing, however this Golden Stroop (Golden et al., 2003) to assess inhibitory con- trol. In the first condition, participants read as many words was optional. A total of 32 children with ADHD at baseline were prescribed medication, and 13 of these children took (i.e., red, blue, green) as possible in 45 s. The second condi- tion required naming different colors of ink (i.e., red, blue, a simulant medication on the day of Wave 1 testing (6.02% of the total sample; N = 216). Wave 2 and Wave 3 follow-up green) in the same 45 s timeframe. In the third and final con- dition (i.e., Color-Word [C-W]), the color names are printed assessments consisted of parallel procedures to assess simi- lar constructs. The University of California, Los Angeles in discordant colors. The total score on the C-W condition is the number of ink colors named in 45 s. The Stroop C-W IRB approved all study procedures. shows strong criterion validity in children and adults (Arán Filippetti & Richaud, 2017; Miyake et al., 2000). The raw Measures C-W score was used to estimate inhibitory control. Children completed the Digit Span subtest from the Wave 1 Predictors Wechsler Intelligence Scale for Children-IV (Wechsler, 2003). In the forward condition, the examiner presents a ADHD. We administered the ADHD module of the Diag- string of numbers aloud; participants must recall the num- bers in the proper order. The backward condition requires nostic Interview Schedule for Children (DISC-IV) Parent Edition (Shaffer et al., 2000) to parents to assess Diagnos- recalling the numbers in the reverse order. The Digit Span Backwards (DSB) raw score was used to estimate working tic and Statistical Manual of Mental Disorders DSM-IV ADHD. The DISC-IV is a computer-assisted, structured memory given that it loaded more strongly on a latent work- ing memory factor of EF relative to the forward condition interview of symptoms, onset, and impairment with strong psychometric properties. The ADHD module has high inter- (Arán Filippetti & Richaud, 2017). Normed scores do not exist for the raw backwards portion of the Digit Span subtest. nal consistency (α = 0.84; Schaffer et al., 2000). Given its superior predictive validity (Fergusson & Horwood, 1995), Therefore, raw scores were utilized for all EF measures and we controlled for age on the latent EF variable. we analyzed the number of inattention (0–9) and hyperactiv- ity-impulsivity symptoms (0–9) from the DISC-IV. Parents Wave 2 Mediators and teachers also rated ADHD symptoms on the Disruptive Behavior Disorder Rating Scale (DBD). Responses ranged Academic Functioning. Youth completed the Word Reading from 0 (not at all) to 3 (very much; Pelham et al., 1992), yielding separate inattention and hyperactivity-impulsivity and Math Reasoning subtests from the Wechsler Individual Achievement Test Second Edition (WIAT-II; Wechsler, totals (0–27). The inattention and hyperactivity-impulsiv- ity scales demonstrated high reliability within the present 2002), which assessed phonological awareness/decoding and mathematical problem solving, respectively. Separate sample (parent-rated inattention  α = 0.94; teacher-rated inattention α = 0.93; parent-rated hyperactivity-impulsivity Word Reading and Math Reasoning standard scores were employed. α = 0.92; teacher-rated hyperactivity-impulsivity α = 0.94). Raw scores were used for inattention and hyperactivity- Parents and teachers completed the Child Behavior Checklist (CBCL) and Teacher Report Form (TRF), respec- impulsivity variables as adjusted norms are not provided for these measures. tively (Achenbach & Rescorla, 2001). The CBCL and TRF consist of 113-items with behaviors rated from 0 (not true) Executive Functioning (EF). For set shifting, we admin- istered the child version of the Trail Making Test (Reitan to 2 (very true/often true). The CBCL includes School Competence items, ranging from 0–6, to assess grades, & Wolfson, 1992). On Part A, participants drew lines as 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 757 class placement, grade repetition, and other problems in the measure due to time constraints. We used the age- and the school setting. Items from the school competence scale sex-adjusted Affective Problems T-score as another meas- differentiate between youth from clinic referred and non- ure of self-reported depression. According to the manual, referred samples and has an acceptable reliability according this subscale demonstrates satisfactory reliability (α = 0.81; to the manual (α = 0.63; Achenbach & Rescorla, 2001). The Achenbach & Rescorla, 2001). All three variables for the TRF Academic Performance scale is the mean of the child’s proposed Wave 3 latent depression factor provided adjusted performance across academic subjects with a test–retest reli- values; therefore, we utilized T-scores. ability of r = 0.90 (Achenbach & Rescorla, 2001). We used the CBCL School Competence and the TRF Academic Per- formance T-scores to create a latent academic functioning Data Analytic Procedures variable. All four observed variables for the proposed aca- demic functioning latent factor provided adjusted values; We employed SEM to test Wave 1 ADHD (i.e., inattention, thus, T-scores were utilized. hyperactivity-impulsivity) and EF as predictors of Wave Social Problems. Parents and teachers completed the 3 depression (Fig. 1A). Next, we tested ADHD diagnostic Dishion Social Preference Scale, a three-item (5-point status as a moderator to evaluate whether the prospective metric) measure of peer acceptance, rejection, and being prediction of depression from EF differs according to ADHD ignored (Dishion, 1990). Negative social preference was diagnostic status. We finally examined academic function- calculated by subtracting the reject from the accept rating ing and social problems as mediators of the prospective and then reversing scoring the difference (Humphreys et al., prediction of Wave 3 depression from baseline inattention, 2013; Lee & Hinshaw, 2006). In addition to the academic hyperactivity-impulsivity, and EF (Fig. 1B). To test these functioning scales, the CBCL and TRF yield parallel Social proposed models, we implemented full information maxi- Problems scales, which are reliable (α = 0.82 for both) and mum likelihood (FIML; Enders, 2010), which performs well valid (Achenbach & Rescorla, 2001). These two variables even with extreme missingness (e.g., 50%; Schlomer et al., were coupled with negative social preference to collectively 2010). We addressed FIML requirements that data are Miss- estimate social problems. Because the Dishion Social Prefer- ing at Random (MAR) or Missing Completely at Random ence Scale does not provide age and sex adjusted norms, we (MCAR) and multivariate normality (Pritikin et al., 2018). used the raw scores for all four social variables for consist- Little’s Test suggested that the data violated the assumption ency across this proposed latent variable. of MCAR. Thus, we determined appropriate auxiliary vari- ables to include to ensure data were missing at random and Wave 3 Youth Self‑Report Outcomes to improve power (Enders, 2010). Auxiliary variables are not central to the specific research question; rather, auxiliary Depression. Youth completed the 27-item Children’s variables may be highly correlated with missingness on the Depression Inventory (CDI; Kovacs, 1992), rating descrip- study variables or the included variables (Enders, 2010). We tions from the past two weeks (e.g., “I feel like crying every tested the correlation between potential auxiliary variables day,” “I feel like crying many days,” “I feel like crying once and variables relevant to the analyses as well as missingness in a while”). Each item is scored from 0 to 2, providing age- on variables for the analyses. Several auxiliary variables sig- and sex-adjusted T-scores. The CDI has shown strong con- nificantly correlated with missingness and/or variables for vergent validity with internalizing and disruptive behavior analyses and were included in all SEM models (Table 2). For (Timbremont et al., 2004). The CDI demonstrated accept- example, Wave 1 WIAT-II Math Reasoning standard score able reliability within our sample (α = 0.85). is not a variable in our analyses; however, it was determined The 47-item Revised Children’s Anxiety and Depression to be an appropriate auxiliary variable because it was sig- Scale (RCADS) includes five normed subscales (Chorpita nificantly correlated with: Wave 2 WIAT-II Word Reading et al., 2000). Items were rated from 0 to 3, reflecting never, standard score, Wave 2 WIAT-II Math Reasoning standard sometimes, often, or always, respectively. The sex- and score, Wave 2 CBCL School Competence T-score missing- grade-adjusted T-score from the Major Depressive Dis- ness, Wave 2 CBCL Social Problems raw score missingness, order subscale, which previously correlated with the CDI and Wave 2 Parent Negative Social Preference missingness. (Chorpita et al., 2000) and shows satisfactory internal con- For predictive models, Wave 2 academic and social mediators sistency within this sample (α = 0.80), was analyzed. were implemented as auxiliary variables. Although Mardia’s The 113-item Youth Self Report (YSR; Achenbach & test of Skewness [191.46, χ (969) = 936.96, p = 0.76] did Rescorla, 2001) is parallel to the CBCL and TRF. The YSR not violate the criteria for multivariate normality, Mardia’s is normed on a sample of 11–18-year-old youth (Achenbach test of Kurtosis [296.59, χ (1) = 7.02, p < 0.01] violated & Rescorla, 2001). Although 154 youth in this study were the assumption for the simplest model; therefore, we imple- 11 years or older at Wave 3, only 61 participants completed mented maximum likelihood robust procedures in Mplus 1 3 758 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Wave 1 Inattention Wave 1 Wave 3 Hyperactivity/ Depression Impulsivity Wave 1 Executive Functioning A) Wave 1 Wave 2 Inattention Academic Fx. Wave 1 Wave 3 Hyperactivity/ Depression Impulsivity Wave 1 Wave 2 Executive Social Problems Functioning B) Fig. 1 A) Proposed predictive model, testing the prospective predic- sion and baseline inattention, hyperactivity/impulsivity, and executive tion of Wave 3 self-reported depression from Wave 1 inattention, functioning. For A and B, circles indicate latent variables. Table  4 hyperactivity-impulsivity, and executive functioning. B) Proposed includes observed variables proposed to derive latent variables. mediational model, testing Wave 2 academic functioning and social Covariates are not included problems as mediators of the relationship between Wave 3 depres- 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 759 (Muthén & Muthén, 1998–2010) to address non-normality status. Such measures are reliably associated with pubertal and accommodate missing data. development as evaluated by the Tanner Scales (Carskadon Prior to running SEM, we conducted separate confirma- & Acebo, 1993; Petersen et al., 1988) For the present study, a tory factor analyses on each Wave 1 (i.e., inattention, hyper- quantitative score described by Carskadon and Acebo (1993) activity-impulsivity, EF), Wave 2 (i.e., academic function- was implemented. Each characteristic was rated from 1- 4 ing, social problems), and Wave 3 (i.e., depression) latent (1 = not yet started changing, 2 = has barely started chang- variables (see preliminary analyses). For proposed latent ing, 3 = change is definitely underway, 4 = change seems variables with three indicator variables (i.e., inattention, completed) with the exception of a menstruation item for hyperactivity-impulsivity, EF, depression) we evaluated girls, which was coded as 1 for no and 4 for yes. The aver- factor loadings to ensure that they exceeded the guidelines age of the items was used to measure puberty for the current of at least 0.3 (Brown, 2014). For latent factors with four study (M = 2.66, SD = 0.73). Finally, if inattention signifi- indicator variables (i.e., academic achievement, social cantly predicted depression, we added anxiety as a covariate problems), fit indices were examined. For predictive and on the latent inattention factor to strengthen specific infer - mediational models, multiple fit indices were evaluated. A ences (Pliszka, 2019). In addition to completing the CBCL non-significant chi-square and comparative Fit Index (CFI) at Wave 2 to assess youth academic and social functioning values ≥ 0.95 indicate good fit. R oot mean square error of (see above for details), parents also completed this measure at approximation (RMSEA) estimates model fit with control of baseline (i.e., Wave 1). The T-score from the Wave 1 Anxiety sample size and per degrees of freedom where values ≤ 0.06 Problems subscale was used to assess child anxiety. Baseline are acceptable (Hu & Bentler, 1999). Finally, a value of 0.08 anxiety was included as a covariate on the latent inattention or less for standardized root mean square residual (SRMR) variable in models where inattention emerged as a significant suggests good model fit (Hu & Bentler,  1999). predictor. Models conservatively accounted for age, sex, baseline depression, SES, and pubertal status. Age and sex were accounted for as the depression outcome observed variables Results adjusted for these factors. Additionally, we controlled for baseline age on the EF latent factor as EF improves with Preliminary Analyses child development. Baseline depression, SES, and pubertal status were included as covariates. Consistent with other Co-occurrence between ADHD and Depression. To pro- studies (e.g., Lawson & Farah, 2017), we utilized family vide additional support for the use of dimensional concep- income and parent education to approximate SES. To cap- tualization of ADHD, we tested for significant differences in ture family income, we used a binary variable where $75,000 Wave 1 and Wave 3 depression based upon baseline ADHD or less was coded as 0 and $75,001 or more was coded as 1 diagnostic status. Because considerable discrepancies due to the fact that 66% of the sample with observed data exist between parent- and self-reported youth internalizing on this variable had an income of more than $75,000. We symptoms (De Los Reyes et al., 2015; Johnston & Murray, also considered mother and father education in estimating 2003; Lewis et al., 2014), youth self-reported depression SES. Education level was categorized as follows: 1 = eighth measures were examined. On the CDI at baseline, 11 of 114 grade or less, 2 = some high school, 3 = high school graduate participants with complete data on this measure were at or or GED, 4 = some college or post-high school, 5 = college above a T-score of 60 (i.e., high average and above). Seven graduate, 6 = advanced graduate or professional degree. The of those 11 participants met diagnostic criteria for ADHD average value of mother and father education, which ranged on the DISC-IV. This did not represent a significant differ - from 1.5-6 (M = 4.92, SD = 0.91), was used; a single value ence in Wave 1 depression according to baseline ADHD from mother or father was utilized in circumstances when status [χ (1) = 1.08, p = 0.30]. At Wave 3, a total of 149 education data was present for one parent. Income and parent participating youth completed the CDI. Seven of those par- education variables were included in all SEM models. Chil- ticipants met or exceeded a T-score of 60, with four of those dren completed the CDI (Kovacs, 1992), as described above seven youth meeting diagnostic criteria for ADHD at Wave at baseline in addition to Wave 3. Each item is scored from 0 1. Again, there was not a significant difference in Wave 3 to 2, with 2 representing higher depression severity. The child depression on the CDI according to baseline ADHD diag- age and sex adjusted T-score was used; norms for seven-year- nostic status [χ (1) = 1.08, p = 0.30]. Similarly, there were olds were used to calculate T-scores for six-year-old children no significant differences between Wave 3 depression and in the current sample. At Wave 3, children completed the baseline ADHD diagnosis according to youth self-report on Pubertal Development Scale (Petersen et al., 1988), a six- the RCADS depression subscale [χ (1) = 1.82, p = 0.18]. item self-report questionnaire used to assess pubertal status Specifically, on the Wave 3 RCADS, two out of 148 youth in males and females and noninvasively assesses pubertal met or exceeded the clinical cutoff (T-score above 70) for 1 3 760 Research on Child and Adolescent Psychopathology (2022) 50:753–770 depression, both of whom had a diagnosis of ADHD at base- inattention, hyperactivity-impulsivity, and EF as prospective line. The fact that we did not observe a significant differ - predictors of early adolescent depression, we modified the ence between those with and without ADHD at baseline and model as is recommended within SEM (Weston & Gore Jr., Wave 1 or Wave 3 depression demonstrates the importance 2006; Violato & Hecker, 2007). Due to concerns regarding of utilizing a dimensional conceptualization of ADHD when collinearity, we next tested a model evaluating total ADHD examining ADHD as a risk factor for depression. symptoms and EF as predictors of Wave 3 depression. Confirmatory Factor Analyses. Factor loadings from Model 2. We created three total ADHD symptom scores all confirmatory factor analyses are provided in Table  4. by summing the inattention and hyperactivity-impulsivity For the inattention, hyperactivity-impulsivity, and EF latent dimensions on the DISC-IV (M = 7.88, SD = 5.55), parent variables, standardized beta coefficients exceeded the rec- DBD (M = 20.46, SD = 14.00), and teacher DBD (M = 17.31, ommended cutoff of 0.3 (Brown, 2014). The latent vari- SD = 15.25). Confir matory factor analyses revealed that ables for Wave 2 academic functioning and social problems these ADHD variables had acceptable factor loadings consisted of four observed variables each. The academic (> 0.49). We regressed the latent Wave 3 self-reported functioning factor demonstrated good fit across multi- depression outcome on Wave 1 ADHD total symptom and ple indices (Table 4): χ (2) = 4.42, p = 0.10, CFI = 0.98, EF latent variables including described covariates. This and SRMR = 0.03, although the RMSEA value was 0.08. model improved fit (Table  5): χ (59) = 78.70, p = 0.04, Because RMSEA cutoffs are vulnerable to poor fit in mod- CFI = 0.97, RMSEA = 0.04, SRMR = 0.07. Although the els with few degrees of freedom (Kenny et al., 2015), but chi-square was significant, this measure of fit is often unre- other fit indices for this academic functioning latent vari- liable (Vandenberg, 2006), so we prioritized other fit indi- able were acceptable, a latent variable was implemented for ces. For this alternative structural model, CFI, RMSEA, academic functioning. In contrast, the social problems latent SRMR all demonstrated good fit where ADHD symp- variable showed poor model fit [χ (2) = 43.51, p < 0.001, toms (β = 0.23, SE = 0.10, p = 0.01), but not EF (β = 0.02, CFI = .71, RMSEA = 0.33 and SRMR = 0.09]. Therefore, SE = 0.13, p = 0.87), positively predicted depression. Model we created two composite variables to estimate parent- and 2 accounted for 5.3% of the variance in Wave 3 depression. teacher-rated social problems, respectively (see Table  3) Model 3. To improve specificity, we next tested EF by z-scoring the CBCL Social Problems raw score and the and inattention as predictors of Wave 3 depression with Dishion Negative Social Preference rating followed by cal- identical covariates. The model fit the data well (Table  5) culating the average z-score; the same approach yielded a [χ (59) = 82.79, p < 0.02, CFI = 0.96, RMSEA = 0.04, teacher-rated social functioning composite. Higher scores SRMR = 0.07] where inattention positively predicted Wave 3 reflected worse social functioning. early adolescent depression (β = 0.33, SE = 0.09, p < 0.001), but EF did not (β = 0.06, SE = 0.12, p = 0.60). In addition Prediction of Early Adolescent Depression to demonstrating good fit to the data, Model 3 accounted for 9.0% of the variance in the Wave 3 latent depression To review, controlling for baseline depression, SES, and variable. pubertal status, we tested childhood inattention, hyper- Model 4. Hyperactivity-impulsivity and EF were also activity-impulsivity, and EF as independent predictors of tested as predictors of depression. Although the model early adolescent depression; we also controlled for baseline showed good fit (Table  5) [χ (59) = 72.43 p < 0.11, age on the EF factor (depression outcomes were adjusted C F I = 0 . 9 8 , R M SE A = 0 . 0 3 , S RM R = 0 . 0 7 ], it o n ly for age/grade and sex). When inattention significantly pre- accounted for 2.0% of the variance in Wave 3 depression. dicted depression, we also controlled for Wave 1 anxiety on Additionally, neither hyperactivity-impulsivity (β = 0.10, inattention. SE = 0.11, p = 0.36) nor EF (β = -0.04, SE = 0.13 p = .77) Model 1. We regressed a latent depression variable from predicted depression, suggesting overall that inattention is Wave 3 on Wave 1 inattention, hyperactivity-impulsivity, the primary risk factor for later depression. and EF latent variables. Key indices suggested model Model 5. We reproduced Model 3 (i.e., inattention and misspecification [χ (93) = 221.02, p < 0.001, CFI = 0.90, EF predicting depression) but conservatively added base- RMSEA = 0.08, SRMR = 0.08; Table 5]. Additionally, the line anxiety (i.e., CBCL Anxiety Problems). Even with latent variable covariance matrix was not positive definite, control of anxiety on the latent inattention variable, the reflecting the high correlation between the inattention and model showed good fit (Table  5) [χ (72) = 99.09, p < 0.01, hyperactivity-impulsivity latent variables. Wave 1 inat- CFI = 0.96, RMSEA = 0.04, SRMR = 0.08] and accounted tention positively predicted Wave 3 depression (β = 0.49, for 10.8% of the variance in Wave 3 depression. Base- SE = 0.15, p = 0.001), but hyper activity-im pulsivity line inattention continued to predict depression (β = 0.35, (β = -0.24, SE = 0.17, p = 0.14) and EF (β = 0.06, SE = 0.12, SE = 0.09, p < 0.001; Fig. 2) whereas EF did not (β = 0.10, p = 0.60) did not. To further examine the initial aim of testing SE = 0.12, p = 0.40). 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 761 DISC DBD-P DBD-T Inattention Inattention Inattention Parent Parent Wave 1 CDITotal Wave 3 *** .92 (.03) *** Income Education T-Score Puberty .89 (.03) *** .43 (.07) *** CBCL Anxiety .36 (.06) T-Score Wave 1 Inattention .03 (.10) -.04 (.11)-.01 (.12) -.01 (.09) *** .35 (.09) CDITotal T-Score *** .66 (.09) *** Wave 3 RCADS .98 (.10) Depression *** Depression -.56 (.08) T-Score *** .76 (.11) YSR Affective T-Score .10 (.12) Wave 1 Executive *** .63 (.05) Functioning Wave 1 Age *** .51 (.06) *** .73 (.05) *** .58 (.06) TMT Stroop WISC-IVDS B CW Backward Fig. 2 Represents Model 5 with standardized estimates. Standard- with additional control of anxiety on inattention. The model accounts *** ized errors are in parentheses. Solid lines indicate significant relation- for 10.8% of the variance in Wave 3 depression. p ≤ 0.001. ** * ships among variables. Inattention positively predicts early adolescent p ≤ 0.01. p ≤ 0.05 depression with control of SES, baseline depression, and puberty Multi-group Analysis. We conducted multi-group analy- A model consisting of Wave 2 academic function- ing and social problems as mediators of predictions from ses based on Wave 1 ADHD diagnostic status to determine whether predictions of youth depression from baseline EF Wave 1 latent inattention, hyperactivity-impulsivity, and EF showed poor fit [χ (194) = 255.03 p < 0.01, CFI = 0.93, differed according to group. However, this model had poor fit (SRMR = 0.22), suggesting that EF predicted depression RMSEA = 0.04, SRMR = 0.09]. Examination of model pathways revealed no mediated effects. Wave 1 inattention similarly among youth with and without ADHD, though concerns related to power for multigroup models (Kline, inversely predicted Wave 2 academic functioning (β = -0.83, SE = 0.34, p = 0.01) and positively predicted Wave 2 par- 2015) limit conclusions. ent-rated social problems (β = 0.35, SE = 0.17 p = 0.03). However, neither Wave 2 academic achievement (β = 0.20, Early Adolescent Depression: Mediation by Academic Functioning and Social Problems SE = 0.23, p = 0.37) nor Wave 2 parent-rated social problems (β = 0.05, SE = 0.13, p = 0.71) predicted Wave 3 depression. Using SEM, we tested academic functioning and social Baseline hyperactivity-impulsivity and EF did not predict any mediators. Finally, none of the mediators (i.e., academic problems as temporally-ordered mediators of predictions of Wave 3 youth self-reported depression from baseline inatten- functioning, parent-rated social problems, teacher-rated social problems) predicted Wave 3 depression. There was tion, hyperactivity-impulsivity, and EF. Specifically, a Wave 2 latent academic functioning variable and two composite a significant direct effect from Wave 1 inattention to Wave 3 depression (β = 0.68, SE = 0.30, p = 0.02), but no signifi - social problems variables (i.e., parent-report, teacher-report) were entered as mediators, controlling for the same covari- cant direct effects from hyperactivity-impulsivity (β = -0.44, SE = 0.24, p = 0.06) or EF (β = 0.00, SE = 0.15, p = 0.37) to ates previously described. 1 3 762 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Table 1 Descriptive Statistics Variable M (SD) or % of Sample Range n of Demographics and Key Study Variables Wave 1 Age 7.39 (1.07) 6-9 216 Wave 2 Age 9.68 (1.27) 7-13 193 Wave 3 Age 12.07 (1.30) 9-15 172 Sex (% Male) 66.67 - 216 Race-Ethnicity (% Caucasian) 50.93 - 216 SES Income (% $75,001 or more) 65.99 - 197 Parent Education 4.92 (0.91) 1.5-6 202 Wave 1 Depression (CDI T-Score) 46.84 (7.47) 35-71 144 Wave 3 Puberty 2.27 (0.73) 1-3.8 143 Wave 1 Anxiety (CBCL Anxiety Problems T-Score) 56.21 (7.49) (50-75) 214 FSIQ 107.29 (14.24) 73-144 216 Wave 1 Inattention Symptoms Inattention Symptoms (DSIC-IV) 4.54 (3.16) 0-9 216 Inattention Symptoms (DBD Parent) 11.17 (7.63) 0-27 210 Inattention Symptoms (DBD Teacher) 9.35 (8.33) 0-27 150 Wave 1 Hyperactivity-Impulsivity Symptoms Hyperactivity-Impulsivity Symptoms (DSIC-IV) 3.34 (3.08) 0-9 216 Hyperactivity-Impulsivity Symptoms (DBD Parent) 9.32 (7.33) 0-27 209 Hyperactivity-Impulsivity Symptoms (DBD Teacher) 7.79 (8.38) 0-27 150 Wave 1 Executive Functioning TMT-B (min) -1.22 (0.81) -5.02 - -0.27 212 WISC-IV DSB Raw 5.86 (1.58) 0-10 210 Stroop C-W 21.42 (6.43) 4-41 201 Wave 2 Academic Functioning WIAT-II Word Reading Standard Score 107.58 (14.27) 53-141 184 WIAT-II Math Reasoning Standard Score 111.86 (16.82) 61-160 183 School Competence T-Score (CBCL) 44.79 (9.46) 24-55 182 Academic Performance T-Score (TRF) 49.27 (10.12) 35-65 95 Wave 2 Social Problems Negative Social Preference (Parent Dishion) -2.60 (1.85) -4-4 172 Negative Social Preference (Teacher Dishion) -2.18 (2.13) -4-4 92 Social Problems Raw (CBCL) 3.27 (3.48) 0-16 188 Social Problems Raw (TRF) 2.28 (2.67) 0-14 91 Wave 3 Depression CDI T-Score 44.14 (7.74) 34-75 149 RCADS T-Score 43.91 (10.55) 30-78 148 Affective Problems T-Score (YSR) 54.66 (7.49) 50-80 61 SES  =  Socioeconomic Status, CDI  =  Children's Depression Inventory, CBCL  =  Child Behavior Check- list, DISC-IV = Diagnostic Interview Schedule for Children Fourth Edition, DBD  =  Disruptive Behav- ior Disorder Rating Scale, TMT-B  =  Trail Making Test Part B, WISC-IV  =  Wechsler Intelligence Scale for Children-Fourth Edition, DSB  =  Digit Span Backwards, Stroop C-W = Stroop Color-Word Condi- tion, WIAT-II  =  Wechsler Individual Achievement Scale Second Edition, TRF  =  Teacher Report Form, RCADS = Revised Children's Anxiety and Depression Scale, YSR = Youth Self-Report Wave 3 depression emerged. Because no model pathways Discussion suggested significant mediation, respecified models were not investigated. We also tested a multigroup model with ADHD Identifying childhood predictors and mediators of depres- diagnostic status as a moderator of mediation by academic sion, prior to the acute vulnerability in adolescence, will and social functioning, but the model’s failure to converge accelerate innovations in prevention and intervention. prevented strong inferences. In a study of 216 youth ages 6–9, baseline inattention, 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 763 Table 2 Auxiliary Variables Demographic Wave 1 Wave 2 Wave 3 Child Age (Waves WIAT-II Word Reading SS DISC-IV Inattention Symptoms CDI Total Raw Score 2 and 3) Child Sex WIAT-II Math Reasoning SS DISC-IV HI Symptoms RCADS Depression Raw Score FSIQ CBCL School Competence T-score DBD-Parent Inattention Symptoms YSR Affective Problems Raw Score CBCL Social Problems T-score DBD-Parent HI Symptoms TRF Academic Performance T-score DBD-Teacher Inattention Symptoms TRF Social Problems T-score DBD-Teacher HI Symptoms TMT-B Stroop C-W Condition WISC-IV Digit Span Backwards RCADS Depression T-score FSIQ  =  Full Scale IQ, WIAT-II   =  Wechsler Individual Achievement Scale Second Edition; SS  =  Standard Score, CBCL  =  Child Behavior Checklist, TRF = Teacher Report Form, DISC-IV = Diagnostic Interview Schedule for Children Fourth Edition, HI = Hyperactivity-Impulsivity, DBD = Disruptive Behavior Disorder Rating Scale; TMT-B = Trail Making Test Part B, Stroop C-W =  Stroop Color-Word Condition, WISC- IV = Wechsler Intelligence Scale for Children-Fourth Edition, RCADS = Revised Children's Anxiety and Depression Scale, CDI  =  Children's Depression Inventory, YSR = Youth Self-Report hyperactivity-impulsivity, and EF were tested as simulta- competency-based model (Cole, 1991), self-esteem may be neous predictors of youth self-reported depression approx- primary given its proximity to depression. Therefore, inter- imately four years later. Controlling for SES, baseline pretations or appraisals of academic and social functioning depression, and puberty as well as age (on the EF latent may be more salient. For example, negative self-schema acti- factor), inattention positively predicted early adolescent vated by stressors and/or low mood potentiates depression, depression, even with added control of baseline anxiety in including through maladaptive attributional styles (Jacobs a model with inattention and EF as latent predictors. In et al., 2008). Given the centrality of multifinality to ADHD, contrast, academic functioning and social problems did not across multiple settings and functional contexts, we await mediate predictions of depression from baseline inattention, strong tests of cognitive factors (e.g., attributional style) and hyperactivity-impulsivity, and EF. competency-based constructs, as mediators of depression. The unique, robust association of early ADHD (inatten- Further, because adolescence confers heighted sensitivity to tion in particular) with early adolescent depression observed social relationships, academic and social functioning meas- in this study is well-aligned with previous evidence, includ- ured at an older age may mediate emerging depression later in ing predictions through young adulthood (Chronis-Tuscano development from childhood ADHD (Powell et al., 2020). In et al., 2010; Meinzer et al., 2016). The centrality of inatten- the current study, social measures were not derived from self- tion to depression was also reported previously, across mul- report; instead, social functioning was estimated from broader tiple informants, in a cross-sectional study (Fenesy & Lee, constructs (i.e., social acceptance, social problems) according 2019). Although inattention is transdiagnostic in nature, to informant report that may lack the precision necessary to inattention assessed via measures of ADHD across inform- capture successful peer interactions. For example, a multi- ants (i.e., parent, teacher) prospectively predicted depression ple mediation model testing whether individual social skills independent of co-occurring EF, anxiety, and baseline depres- (e.g., cooperation, empathy) collectively and uniquely medi- sion. Whereas youth with ADHD struggle with selective and ate predictions of depression from inattention, hyperactivity- sustained attention, anxiety is characterized by attentional impulsivity, and EF could identify particular aspects of social biases secondary to threat (Weissman et al., 2012). Further, behavior to target for intervention and reduce depression in directed attention is generally a top-down cognitive process, youth. Social skills training improves social functioning and but responses to threat are primarily mediated subcortically reduces the risk of depression in youth with autism spectrum (Nigg, 2017). Thus, inattention secondary to ADHD is evident disorder (Hotton & Coles, 2016), suggesting that continued across settings and contexts and contributes to impairments research examining social functioning as an underlying mech- across multiple domains, independently conferring risk for anism of depression in youth with neurodevelopmental disor- depression later in development. The hypothesis that symp- ders is worthwhile and will facilitate innovations in interven- toms of inattention and hyperactivity-impulsivity as well tion development. Overall, appropriate interventions targeting as EF deficits would predict multiple academic and social either a behavioral deficit or maladaptive cognitions could “failures” was not supported in this study. According to a subsequently be applied to mitigate the risk of depression. 1 3 764 Research on Child and Adolescent Psychopathology (2022) 50:753–770 1 3 Table 3 Correlations Among Covariates, Predictor, Mediator, and Outcome Variables Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Wave 1 Age - 2. Sex (% Male) -0.09 - 3. Income (1 = $75,001 or more) -0.05 0.12 - * *** 4. Parent Education -0.04 0.14 0.46 - * *** *** 5. Wave 1 Depression (CDI T-Score) 0.11 -0.17 -0.27 -0.32 - *** *** 6. Wave 3 Puberty 0.49 -0.35 -0.07 -0.02 0.17 - 7. Wave 1 Anxiety (CBCL Anxiety Problems T-Score) 0.04 0.07 -0.03 -0.03 0.06 0.06 - * * *** * 8. FSIQ -0.04 0.13 0.15 0.35 -0.17 -0.07 -0.06 - ** *** *** 9. Wave 1 Inattention Symptoms (DSIC-IV) -0.03 0.11 -0.09 -0.08 0.24 0.01 0.34 -0.24 - * *** *** *** 10. Wave 1 Inattention Symptoms (DBD P) -0.05 0.09 -0.06 -0.09 0.17 0.03 0.31 -0.26 0.82 - *** *** *** 11. Wave 1 Inattention Symptoms (DBD T) -0.10 0.15 -0.11 -0.04 0.14 0.06 0.08 -0.28 0.40 0.41 - ** ** *** *** *** *** 12. Wave 1 Hyperactivity-Impulsivity Symptoms (DSIC-IV) -0.19 0.19 0.10 -0.04 0.10 0.02 0.32 -0.03 0.58 0.60 0.28 - * ** *** *** *** *** *** 13. Wave 1 Hyperactivity-Impulsivity Symptoms (DBD P) -0.16 0.16 0.10 -.04 0.09 0.05 0.34 -0.08 0.61 0.75 0.34 0.87 - *** ** ** *** *** *** *** 14. Wave 1 Hyperactivity-Impulsivity Symptoms (DBD T) -0.31 0.21 0.10 -0.03 0.11 0.14 0.14 -0.14 0.21 0.28 0.64 0.49 0.49 *** * *** *** *** ** ** 15. Wave 1 TMT-B (min) 0.46 0.00 -0.02 0.05 -0.01 0.18 0.06 0.38 -0.27 -0.28 -0.15 -0.19 -0.18 *** *** * * 16. Wave 1 WISC-IV DSB Raw 0.29 0.02 0.08 0.10 -0.11 0.09 0.02 0.38 -0.12 -0.15 -0.12 -0.14 -0.09 *** ** *** *** *** ** * * 17. Wave 1 Stroop C-W 0.38 -0.04 0.09 0.06 -0.06 0.22 -0.04 0.31 -0.23 -0.27 -0.23 -0.15 -0.16 * ** *** *** *** *** 18. Wave 2 WIAT-II Word Reading Standard Score -0.08 0.05 0.14 0.16 -0.21 -0.08 0.04 0.57 -0.24 -0.24 -0.31 -0.06 -0.11 ** *** *** *** *** *** * ** 19. Wave 2 WIAT-II Math Reasoning Standard Score -0.10 0 .11 0.21 0.39 -0.16 -0.06 -0.07 0.76 -0.31 -0.36 -0.31 -0.14 -0.18 *** *** ** ** *** *** *** *** *** *** 20. Wave 2 School Competence T-Score (CBCL) -0.02 0.07 0.27 0.28 -0.21 0.01 -0.21 0.59 -0.48 -0.49 -0.38 -0.24 0.33 ** ** * *** ** ** *** 21. Wave 2 Academic Performance T-Score (TRF) 0.08 0.05 0.28 0.26 -0.26 0.16 -0.02 0.55 -0.31 -0.32 -0.39 -0.04 -0.07 *** * *** *** *** *** *** 22. Wave 2 Negative Social Preference (Parent Dishion) 0.07 0.01 -0.00 -0.07 0.10 0.08 0.26 -0.16 0.35 0.38 0.28 0.32 0.37 ** * * * * 23. Wave 2 Negative Social Preference (Teacher Dishion) 0.08 0.13 -0.13 -0.08 0.12 0.04 0.31 -0.16 0.22 0.20 0.28 0.23 0.23 * *** *** *** *** * *** *** 24. Wave 2 Social Problems Raw (CBCL) -0.01 0.01 -0.02 -0.16 0.10 0.03 0.39 -0.24 0.41 0.46 0.23 0.46 0.49 * * ** *** ** 25. Wave 2 Social Problems Raw (TRF) 0.02 0.13 -0.03 -0.03 0.01 0.00 0.19 -0.13 0.22 0.25 0.07 0.26 0.25 * ** *** *** * * * 26. Wave 3 CDI T-Score 0.17 -0.23 -0.13 -0.15 0.09 0.28 0.27 -0.17 0.11 0.17 0.22 -0.02 -0.00 *** * ** *** 27. Wave 3 RCADS T-Score 0.06 0.03 -0.02 -0.06 0.00 0.03 0.32 -0.17 0.24 0.28 0.12 0.10 0.14 28. Wave 3 Affective Problems T-Score (YSR) 0.04 0.07 0.02 0.09 0.23 0.17 0.14 0.13 0.07 0.14 0.05 0.13 0.11 Variable 14 15 16 17 18 19 20 21 22 23 24 25 26 27 14. Wave 1 Hyperactivity-Impulsivity Symptoms (DBDT) - 15. Wave 1 TMT-B (min) -0.14 - *** 16. Wave 1 WISC-IV DSB Raw -0.13 0.45 - *** *** 17. Wave 1 Stroop C-W -0.11 0.34 0.26 - * *** *** ** 18. Wave 2 WIAT-II Word Reading Standard Score -0.17 0.25 0.36 0.20 - *** *** *** *** 19. Wave 2 WIAT-II Math Reasoning Standard Score -0.11 0.39 0.39 0.26 0.61 - Research on Child and Adolescent Psychopathology (2022) 50:753–770 765 Critically, screening for childhood inattention may iden- tify youth vulnerable to later onset depression. These indi- viduals may benefit from prevention efforts (e.g., learning CBT skills to enhance mood) to mitigate depression onset and associated negative outcomes, in line with experimen- tal evidence (Brent et al., 2015). Consistent with pediat- ric treatment guidelines, our findings emphasize the need for ongoing care and monitoring of children with ADHD, especially as inattentive symptoms are likely to persist into adolescence and adulthood (Wolraich et al., 2019) and may increase the likelihood of co-occurring depression. Effective intervention for childhood ADHD requires treatment by a multidisciplinary team to properly assess and intervene prior to depression onset (e.g., medical providers, psychologists, educators; Barbaresi, 2020). Future studies should test the associations observed in the present study in youth diag- nosed with major depressive disorder or those at elevated risk (e.g., offspring of depressed mothers) as well. Parental depression, for example, is a robust risk factor for adolescent depression. If inattention symptoms predicted clinically sig- nificant depression in other populations, this would substan- tiate the rationale to screen inattention even in the absence of ADHD. Additionally, greater refinement of specific inatten- tion symptoms or combinations of symptoms that increase risk for later depression would enable clinicians to improve screening techniques. Surprisingly, with control of baseline ADHD, childhood EF did not predict early adolescent depression. Consist- ent with evidence that EF impairments remit following a depressive episode in adults and adolescents (Biringer et al., 2005; Maalouf et al., 2011), present findings suggest that childhood EF did not uniquely confer vulnerability for early adolescent depression. However, developmentally-sensitive socio-emotional and neurobiological changes critically con- textualize these findings. For example, most of the current participants had only begun their transition to adolescence, thus still undergoing well-characterized developmental unfolding of limbic systems and prefrontal regions under- lying emotion dysregulation and heightened reward sensi- tivity (Powers & Casey, 2015). Further, poor EF may be more acutely related to adolescent-onset depression because frontal networks supporting cognitive down-regulation of negative emotionality are still emerging. EF measures con- tinue to suffer from poor ecological validity whereas rating scales may prove more useful (Barkley & Murphy, 2011). Similarly, computerized EF measures, which include more accurate assessments of subtle variations in response time (e.g., NIH ToolBox; Zelazo et al., 2014), may show better predictive properties. Finally, we used a latent EF variable, derived from multiple EF domains, to reduce measurement error. However, differentiated measures of EF may yield more specific patterns of association. For example, better inhibition in adults supported reappraisal of negative stimuli 1 3 Table 3 (continued) Variable 14 15 16 17 18 19 20 21 22 23 24 25 26 27 ** *** *** *** *** 20. Wave 2 School Competence T-Score (CBCL) -0.22 0.33 0.25 0.26 0.56*** 0.60 - ** *** ** *** *** *** 21. Wave 2 Academic Performance T-Score (TRF) -0.07 0.30 0.38 0.28 0.50 0.57 0.68 - ** ** * *** 22. Wave 2 Negative Social Preference (Parent Dishion) 0.25 -0.10 -0.03 -0.19 -0.12 -0.15 -0.38 -0.18 - * ** *** ** *** 23. Wave 2 Negative Social Preference (Teacher Dishion) 0.26 -0.01 0.02 0.04 -0.04 -0.25 -0.34 -0.28 0.54*** - ** ** ** * *** *** * *** *** 24. Wave 2 Social Problems Raw (CBCL) 0.23 -0.18 -0.13 -0.18 -0.17 -0.31 -0.45 -0.21 0.67 0.37 - * ** *** ** *** *** *** 25. Wave 2 Social Problems Raw (TRF) 0.21 -0.23 0.03 -0.04 -0.12 -0.25 -0.33 -0.25 0.45 0.71 0.38 - ** * * 26. Wave 3 CDI T-Score -0.01 0.01 -0.00 0.05 -0.10 -0.15 -0.21 -0.15 0.18 0.16 0.18 0.06 - ** *** 27. Wave 3 RCADS T-Score 0.00 -0.10 -0.02 -0.09 -0.03 -0.12 -0.13 -0.07 0.23 0.20 0.07 0.16 0.64 - *** *** 28. Wave 3 Affective Problems T-Score (YSR) 0.06 0.13 0.22 0.17 0.21 0.11 -0.02 0.13 0.06 0.12 -0.09 -0.07 0.57 0.76 CDI  =  Children's Depression Inventory, CBCL  =  Child Behavior Checklist, DISC-IV  =  Diagnostic Interview Schedule for Children Fourth Edition, DBD P  =  Disruptive Behavior Disorder Rating Scale Parent Report, DBD T = Disruptive Behavior Disorder Rating Scale Teacher Report, TMT-B = Trail Making Test Part B, WISC-IV = Wechsler Intelligence Scale for Children- Fourth Edition, DSB = Digit Span Backwards; Stroop C-W = Stroop Color-Word Condition, WIAT-II = Wechsler Individual Achievement Scale Second Edition, TRF = Teacher Report Form; RCADS = Revised Children's Anxiety and Depression Scale, YSR = Youth Self-Report ***p ≤ 0.001; **p ≤ 0.01; *p ≤ 0.05 766 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Table 4 Factor Loadings for Factor Factor Loading SE z p Confirmatory Factor Analyses Wave 1 Inattention* Inattention Symptoms (DSIC-IV) 0.85 0.06 13.29 <0.001 Inattention Symptoms (DBD Parent) 0.96 0.06 15.40 <0.001 Inattention Symptoms (DBD Teacher) 0.44 0.07 6.07 <0.001 Wave 1 Hyperactivity-Impulsivity* Hyperactivity-Impulsivity Symptoms (DSIC-IV) 0.93 0.04 21.92 <0.001 Hyperactivity-Impulsivity Symptoms (DBD Parent) 0.93 0.04 23.78 <0.001 Hyperactivity-Impulsivity Symptoms (DBD Teacher) 0.53 0.07 14.95 <0.001 Wave 1 Executive Functioning* TMT-B (min) 0.76 0.08 9.39 <0.001 WISC-IV DSB Raw 0.58 0.07 7.36 <0.001 Stroop C-W 0.48 0.07 6.48 <0.001 Wave 2 Academic Functioning* WIAT-II Word Reading Standard Score 0.73 0.05 16.06 <0.001 WIAT-II Math Reasoning Standard Score 0.77 0.05 16.21 <0.001 School Competence T-Score (CBCL) 0.77 0.05 15.66 <0.001 Academic Performance T-Score (TRF) 0.81 0.06 14.61 <0.001 Wave 2 Social Problems Negative Social Preference (Parent Dishion) 0.85 0.07 12.26 <0.001 Negative Social Preference (Teacher Dishion) 0.67 0.12 5.67 <0.001 Social Problems Raw (CBCL) 0.74 0.06 11.82 <0.001 Social Problems Raw (TRF) 0.64 0.11 5.62 <0.001 Wave 3 Depression* CDI T-Score 0.70 0.05 13.39 <0.001 RCADS T-Score 0.92 0.06 14.90 <0.001 Affective Problems T-Score (YSR) 0.79 0.07 10.68 <0.001 DISC-IV = Diagnostic Interview Schedule for Children Fourth Edition; DBD = Disruptive Behavior Dis- order Rating Scale, TMT-B = Trail Making Test Part B, WISC-IV = Wechsler Intelligence Scale for Chil- dren-Fourth Edition, DSB  =  Digit Span Backwards, Stroop C-W  =  Stroop Color-Word Condition, WIAT- II  =  Wechsler Individual Achievement Scale Second Edition, CBCL  =  Child Behavior Checklist, TRF =    Teacher Report Form, CDI  =  Children's Depression Inventory, RCADS  =  Revised Children's Anxiety and Depression Scale, YSR = Youth Self-Report *Indicates Latent Factors Used in Analyses and effective emotion suppression that mitigated depressive limitations include a sample recruited for ADHD at baseline, symptoms (Joormann & Gotlib, 2010). but one not specifically designed to capture youth depression. The present study had several key strengths and limitations. Thus, few participants demonstrated clinically significant First, the SEM approach reduced measurement error and Type depression in early adolescence. At Wave 3, only one par- I error, which is particularly important for EF given its mul- ticipant had completed pubertal development, a key limiting tidimensionality. We also utilized temporally-ordered, multi- factor given the dramatic increase in psychopathology sec- method/informant data to test causal mediation. Important ondary to pubertal timing. These aspects of the study design, Table 5 Predictive Model Fit Model χ2 CFI RMSEA SRMR Indices Model 1 (Inattention, Hyperactivity-Impulsivity, EF) χ (93) = 221.02, p< 0.001 0.90 0.08 0.08 Model 2 (ADHD, EF) χ (59) = 78.70, p= 0.04 0.97 0.04 0.07 Model 3 (Inattention, EF) χ (59) = 82.79, p< 0.02 0.96 0.04 0.07 Model 4 (Hyperactivity-Impulsivity, EF) χ2(59) = 72.43 p< 0.11 0.98 0.03 0.07 Model 5 (Inattention with control of anxiety, EF) χ (72) = 99.09, p< 0.01 0.96 0.04 0.08 Latent predictor variables in parentheses. Bold text indicates good model fit based on the index 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 767 Informed Consent Prior to each wave of data collection, written con- in combination with the prospective longitudinal framework, sent was collected from parents/guardians and participating youth pro- likely contributed to the somewhat low variance accounted for vided written assent. the Wave 3 latent depression variable. Variance accounted for would likely increase if replicated in samples studying post- Conflicts of Interest The authors declare that they have no conflicts pubertal adolescents and if other risk factors highly predic- of interest. tive of youth depression (e.g., maternal depression; Hammen & Brennan, 2003) were included. Nevertheless, the fact that Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- childhood inattention significantly predicted early adolescent tion, distribution and reproduction in any medium or format, as long depression approximately four years later remains an important as you give appropriate credit to the original author(s) and the source, finding and may be particularly critical to study in a sample of provide a link to the Creative Commons licence, and indicate if changes youth who all meet diagnostic criteria for ADHD. With respect were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated to multigroup models, the present study may have been under- otherwise in a credit line to the material. If material is not included in powered to detect ee ff cts given 216 participants, missing data, the article's Creative Commons licence and your intended use is not and inclusion of several parameters (Kline, 2015); therefore, permitted by statutory regulation or exceeds the permitted use, you will results related to the multigroup model must be interpreted need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . with caution. Multigroup models examining whether the pro- spective prediction of depression from childhood EF varies according to ADHD diagnostic status should be evaluated in References larger samples to improve confidence in our results. Addi- tionally, consideration of sex and age moderators would be Achenbach, T. M., & Rescorla, L. (2001). ASBEA School-age Forms compelling, especially as the risk for depression increases dur- & Profiles . University of Vermont, Research Center for Children, ing adolescence and the prevalence of depression is higher in Youth, and Families. Arán Filippetti, V., & Richaud, M. C. (2017). A structural equation adolescent girls relative to boys (Thapar et al., 2012). Finally, modeling of executive functions, IQ and mathematical skills in a small number of participants with a diagnosis of ADHD primary students: Differential effects on number production, men- were on stimulant medication on the day of baseline testing tal calculus and arithmetical problems. Child Neuropsychology, (n = 13). Simulant medication improves performance on some 23(7), 864–888. https://doi. or g/10. 1080/ 09297 049. 2016. 11996 65 Avenevoli, S., Swendsen, J., He, J.-P., Burstein, M., & Merikangas, K. EF dimensions in children with ADHD (Barnett et al., 2001; R. (2015). Major Depression in the National Comorbidity Survey- Snyder et al., 2008). Medication use in ADHD is likely most Adolescent Supplement: Prevalence, Correlates, and Treatment. reflective of disorder severity due to the intervention selection Journal of the American Academy of Child & Adolescent Psychia- bias, and because the current study did not use random assign- try, 54(1), 37-44.e32. https:// doi. org/ 10. 1016/j. jaac. 2014. 10. 010 Barbaresi, W. J., Campbell, L., Diekroger, E. A., Froehlich, T. E., Liu, ment we are not positioned to make inferences about treatment Y. H., O’Malley, E., & Chan, E. (2020). Society for developmen- effects (Larzelere et al., 2004). tal and behavioral pediatrics clinical practice guideline for the We tested the prospective association of childhood ADHD assessment and treatment of children and adolescents with com- plex attention-deficit/hyperactivity disorder. Journal of Develop - and EF with depression in early adolescence, including medi- mental & Behavioral Pediatrics, 41, S35–S57. https://doi. or g/10. ation by academic functioning and social problems. Latent 1097/ DBP. 00000 00000 000770 inattention positively predicted youth self-reported depression, Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and but hyperactivity-impulsivity and EF did not. These findings executive functions: Constructing a unifying theory of ADHD. Psychological Bulletin, 121(1), 65–94. https:// doi. org/ 10. 1037/ support the centrality of inattention as a unique risk factor 0033- 2909. 121.1. 65 for depression, demonstrating the importance of screening Barkley, R. A. (2012). Executive functions: What they are, how they for inattention and interventions to improve inattention and work, and why they evolved. Guilford Press. reduce risk. Continued investigation of underlying mecha- Barkley, R. A., & Murphy, K. R. (2011). The Nature of Executive Function (EF) Deficits in Daily Life Activities in Adults with nisms between inattention and early adolescent depression ADHD and Their Relationship to Performance on EF Tests. will provide additional targets for intervention. Journal of Psychopathology and Behavioral Assessment, 33(2), 137–158. https:// doi. org/ 10. 1007/ s10862- 011- 9217-x Funding This work was supported by the National Institutes of Mental Barnett, R., Maruff, P., Vance, A., Luk, E. S., Costin, J., Wood, C., & Health predoctoral fellowship in Biobehavioral Issues in Mental and Pantelis, C. (2001). Abnormal executive function in attention defi- Physical Health (MH 15750 to Michelle C. Fenesy) and the National cit hyperactivity disorder: The effect of stimulant medication and Institutes of Health (R03AA020186-01 to Steve S. Lee). age on spatial working memory. Psychological Medicine, 31(6), 1107–1115. https:// doi. org/ 10. 1017/ S0033 29170 10041 72 Best, J. R., Miller, P. H., & Naglieri, J. A. (2011). Relations between execu- Compliance with Ethical Standards tive function and academic achievement from ages 5 to 17 in a large, representative national sample. Learning and Individual Differences, Ethical Approval Investigators obtained Institutional Review Board 21(4), 327–336. https:// doi. org/ 10. 1016/j. lindif. 2011. 01. 007 (IRB) approval from the University of California, Los Angeles. 1 3 768 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Biringer, E., Lundervold, A., Stordal, K., Mykletun, A., Egeland, J., Fergusson, D. M., & Horwood, L. J. (1995). Predictive Validity of Bottlender, R., & Lund, A. (2005). Executive function improve- Categorically and Dimensionally Scored Measures of Disruptive ment upon remission of recurrent unipolar depression. European Childhood Behaviors. Journal of the American Academy of Child Archives of Psychiatry and Clinical Neuroscience, 255(6), 373– & Adolescent Psychiatry, 34(4), 477–487. https://do i.or g/10. 10 97/ 380. https:// doi. org/ 10. 1007/ s00406- 005- 0577-700004 583- 19950 4000- 00015 Birmaher, B., & Brent, D. (2007). Practice Parameter for the Assess- Gau, S. S. F., & Shang, C. Y. (2010). Executive functions as endophe- ment and Treatment of Children and Adolescents With Depressive notypes in ADHD: Evidence from the Cambridge Neuropsycho- Disorders. Journal of the American Academy of Child & Adoles- logical Test Battery (CANTAB). Journal of Child Psychology and cent Psychiatry, 46(11), 1503–1526. https:// doi. org/ 10. 1097/ chi. Psychiatry, 51(7), 838–849. https://doi. or g/10. 1111/j. 1469- 7610. 0b013 e3181 45ae1c2010. 02215.x Brent, D. A., Brunwasser, S. M., Hollon, S. D., Weersing, V. R., Clarke, Golden, C. J., Freshwater, S. M., & Zarabeth, G. (2003). Stroop Color G. N., Dickerson, J. F., & Lynch, F. L. (2015). Effect of a cogni- and Word Test Children's Version for ages 5-14: A Manual for tive-behavioral prevention program on depression 6 years after Clinical and Experimental Uses. Stoelting. implementation among at-risk adolescents: A randomized clini- Greenberg, P. E., Fournier, A.-A., Sisitsky, T., Pike, C. T., & Kessler, cal trial. JAMA Psychiatry, 72(11), 1110–1118. https://doi. or g/10. R. C. (2015). The economic burden of adults with major depres- 1001/ jamap sychi atry. 2015. 1559 sive disorder in the United States (2005 and 2010). The Journal Brown, T. A. (2014). Confirmatory factor analysis for applied of Clinical Psychiatry, 76(2), 155–162. https:// doi. org/ 10. 4088/ research: Guilford Publications.jcp. 14m09 298 Carskadon, M. A., & Acebo, C. (1993). A self-administered rating Gyurak, A., Goodkind, M. S., Kramer, J. H., Miller, B. L., & Levenson, scale for pubertal development. Journal of Adolescent Health, R. W. (2012). Executive functions and the down-regulation and 14(3), 190–195. https:// doi. org/ 10. 1016/ 1054- 139x(93) 90004-9 up-regulation of emotion. Cognition & Emotion, 26(1), 103–118. Castellanos, F. X., & Tannock, R. (2002). Neuroscience of attention-https:// doi. org/ 10. 1080/ 02699 931. 2011. 557291 deficit/hyperactivity disorder: The search for endophenotypes. Hammen, C., & Brennan, P. A. (2003). Severity, chronicity, and timing Nature Reviews Neuroscience, 3, 617–628. https:// doi. org/ 10. of maternal depression and risk for adolescent offspring diagnoses 1038/ nrn896 in a community sample. Archives of General Psychiatry, 60(3), Center for Disease Control and Prevention (2016). National Vital Sta- 253–258. https:// doi. org/ 10. 1001/ archp syc. 60.3. 253 tistics Reports. Deaths: Leading Causes for 2014. Retrieved Harter, S., & Marold, D. R. (1994). The direction of the link between Chorpita, B. F., Yim, L., Moffitt, C., Umemoto, L. A., & Francis, S. E. self-esteem and affect: Beyond causal modeling. Disorders and (2000). Assessment of symptoms of DSM-IV anxiety and depres- dysfunctions of the self. In D. Cicchetti & S. L. Toth (Eds.), Roch- sion in children: A revised child anxiety and depression scale. ester symposium on developmental psychopathology (Vol. 5, pp. Behaviour Research and Therapy, 38(8), 835–855. https://doi. or g/ 333–370). New York: University of Rochester Press. 10. 1016/ S0005- 7967(99) 00130-8 Hotton, M., & Coles, S. (2016). The effectiveness of social skills train- Chronis-Tuscano, A., Molina, B. G., Pelham, W. E., et al. (2010). Very ing groups for individuals with autism spectrum disorder. Review early predictors of adolescent depression and suicide attempts in Journal of Autism and Developmental Disorders, 3(1), 68–81. children with attention-deficit/hyperactivity disorder. Archives of https:// doi. org/ 10. 1007/ s40489- 015- 0066-5 General Psychiatry, 67(10), 1044–1051. https://doi. or g/10. 1001/ Hu, L., & t., & Bentler, P. M. . (1999). Cutoff criteria for fit indexes in archg enpsy chiat ry. 2010. 127. covariance structure analysis: Conventional criteria versus new Cole, D. A. (1991). Preliminary support for a competency-based model alternatives. Structural Equation Modeling: A Multidisciplinary of depression in children. Journal of Abnormal Psychology, Journal, 6(1), 1–55. https://doi. or g/10. 1080/ 10705 51990 95401 18 100(2), 181–190. https:// doi. org/ 10. 1037/ 0021- 843X. 100.2. 181 Huang-Pollock, C., Mikami, A. Y., Pfiffner, L., & McBurnett, K. De Los Reyes, A., Augenstein, T. M., Wang, M., Thomas, S. A., Drabick, (2009). Can Executive Functions Explain the Relationship D. A. G., Burgers, D. E., & Rabinowitz, J. (2015). The validity Between Attention Deficit Hyperactivity Disorder and Social of the multi-informant approach to assessing child and adolescent Adjustment? Journal of Abnormal Child Psychology, 37(5), mental health. Psychological Bulletin, 141(4), 858–900. https://doi. 679–691. https:// doi. org/ 10. 1007/ s10802- 009- 9302-8 org/ 10. 1037/ a0038 498. Humphreys, K. L., Katz, S. J., Lee, S. S., Hammen, C., Brennan, P. A., Diamantopoulou, S., Rydell, A.-M., Thorell, L. B., & Bohlin, G. & Najman, J. M. (2013). The association of ADHD and depres- (2007). Impact of Executive Functioning and Symptoms of Atten- sion: Mediation by peer problems and parent–child difficulties in tion Deficit Hyperactivity Disorder on Children’s Peer Relations two complementary samples. Journal of Abnormal Psychology, and School Performance. Developmental Neuropsychology, 32(1), 122(3), 854–867. https:// doi. org/ 10. 1037/ a0033 895. 521–542. https:// doi. org/ 10. 1080/ 87565 64070 13609 81 Jacobs, R. H., Reinecke, M. A., Gollan, J. K., & Kane, P. (2008). Dickson, K. S., Ciesla, J. A., & Zelic, K. (2017). The role of executive Empirical evidence of cognitive vulnerability for depression functioning in adolescent rumination and depression. Cognitive among children and adolescents: A cognitive science and develop- Therapy and Research, 41(1), 62–72. https:// doi. org/ 10. 1007/ mental perspective. Clinical Psychology Review, 28(5), 759–782. s10608- 016- 9802-0https:// doi. org/ 10. 1016/j. cpr. 2007. 10. 006 Dishion, T. J. (1990). The peer context of troublesome child and ado- Johnston, C., & Murray, C. (2003). Incremental Validity in the Psycho- lescent behavior. In P. E. Leone (Ed.), Understanding troubled and logical Assessment of Children and Adolescents. Psychological troubling youth (pp. 128–153). Sage Publications. Assessment, 15(4), 496–507. https:// doi. org/ 10. 1037/ 1040- 3590. Enders, C. K. (2010). Applied missing data analysis. Guilford Press.15.4. 496. Favre, T., Hughes, C., Emslie, G., Stavinoha, P., Kennard, B., & Car- Joormann, J., & Gotlib, I. H. (2010). Emotion regulation in depression: mody, T. (2008). Executive functioning in children and adoles- Relation to cognitive inhibition. Cognition and Emotion, 24(2), cents with major depressive disorder. Child Neuropsychology, 281–298. https:// doi. org/ 10. 1080/ 02699 93090 34079 48 15(1), 85–98. https:// doi. org/ 10. 1080/ 09297 04080 25773 11. Kenny, D. A., Kaniskan, B., & McCoach, D. B. (2015). The Perfor- Fenesy, M. C., & Lee, S. S. (2019). Association of ADHD and Execu- mance of RMSEA in Models With Small Degrees of Freedom. tive Functioning With Childhood Depression. Journal of Atten- Sociological Methods & Research, 44(3), 486–507. https:// doi. org/ 10. 1177/ 00491 24114 543236 tion Disorders, 23(11), 1368–1378. https://doi. or g/10. 1177/ 10870 54717 723983. 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 769 Kline, R. B. (2015). Principles and practice of structural equation Clinical Psychology Review, 34(8), 595–607. https:// doi. org/ 10. modeling. Guilford publications.1016/j. cpr. 2014. 10. 002 Kofler, M. J., Irwin, L. N., Soto, E. F., Groves, N. B., Harmon, S. L., Meinzer, M. C., Pettit, J. W., Waxmonsky, J. G., Gnagy, E., Molina, & Sarver, D. E. (2019). Executive functioning heterogeneity in B. S. G., & Pelham, W. E. (2016). Does Childhood Attention- pediatric ADHD. Journal of Abnormal Child Psychology, 47(2), Dec fi it/Hyperactivity Disorder (ADHD) Predict Levels of Depres - 273–286. https:// doi. org/ 10. 1007/ s10802- 018- 0438-2 sive Symptoms during Emerging Adulthood? Journal of Abnor- Kovacs, M. (1992). Children’s depression inventory: Manual. North mal Child Psychology, 44(4), 787–797. https:// doi. org/ 10. 1007/ Tonawanda, NJ: Multi-Health Systems.s10802- 015- 0065-0 Lawson, G. M., & Farah, M. J. (2017). Executive function as a mediator Miller, M., Nevado-Montenegro, A. J., & Hinshaw, S. P. (2012). Child- between SES and academic achievement throughout childhood. hood Executive Function Continues to Predict Outcomes in Young International Journal of Behavioral Development, 41(1), 94–104. Adult Females with and Without Childhood-Diagnosed ADHD. https:// doi. org/ 10. 1177/ 01650 25415 603489 Journal of Abnormal Child Psychology, 40(5), 657–668. https:// Larzelere, R. E., Kuhn, B. R., & Johnson, B. (2004). The interven-doi. org/ 10. 1007/ s10802- 011- 9599-y tion selection bias: An underrecognized confound in intervention Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, research. Psychological Bulletin, 130(2), 289–303. https://doi. or g/ A., & Wager, T. D. (2000). The unity and diversity of executive 10. 1037/ 0033- 2909. 130.2. 289 functions and their contributions to complex “Frontal Lobe” tasks: Lee, K., Bull, R., & Ho, R. M. H. (2013). Developmental Changes in A latent variable analysis. Cognitive Psychology, 41(1), 49–100. Executive Functioning. Child Development, 84(6), 1933–1953. https:// doi. org/ 10. 1006/ cogp. 1999. 0734 https:// doi. org/ 10. 1111/ cdev. 12096 Muthén, L. K., & Muthén, B. O. (1998 –2010). Mplus user’s guide (6th Lee, S. S., & Hinshaw, S. P. (2006). Predictors of Adolescent Func- ed.). Los Angeles, CA: Author. tioning in Girls With Attention Deficit Hyperactivity Disorder Nelson, T. D., Kidwell, K. M., Nelson, J. M., Tomaso, C. C., Hankey, (ADHD): The Role of Childhood ADHD, Conduct Problems, and M., & Espy, K. A. (2018). Preschool Executive Control and Inter- Peer Status. Journal of Clinical Child & Adolescent Psychology, nalizing Symptoms in Elementary School. Journal of Abnormal 35(3), 356–368. https:// doi. org/ 10. 1207/ s1537 4424j ccp35 03_2. Child Psychology. https:// doi. org/ 10. 1007/ s10802- 017- 0395-1 Lewis, A. J., Bertino, M. D., Bailey, C. M., Skewes, J., Lubman, D. I., Nigg, J. T. (2017). Annual Research Review: On the relations among & Toumbourou, J. W. (2014). Depression and suicidal behavior self-regulation, self-control, executive functioning, effortful con- in adolescents: a multi-informant and multi-methods approach to trol, cognitive control, impulsivity, risk-taking, and inhibition for diagnostic classification. Frontiers in Psychology, 5(766), 1–24. developmental psychopathology. Journal of Child Psychology and https:// doi. org/ 10. 3389/ fpsyg. 2014. 00766. Psychiatry, 58(4), 361–383. https:// doi. org/ 10. 1111/ jcpp. 12675 Maalouf, F. T., Brent, D., Clark, L., Tavitian, L., McHugh, R. M., Nigg, J. T., Blaskey, L. G., Stawicki, J. A., & Sachek, J. (2004). Evaluat- Sahakian, B. J., & Phillips, M. L. (2011). Neurocognitive impair- ing the endophenotype model of ADHD neuropsychological defi- ment in adolescent major depressive disorder: state vs. trait illness cit: Results for parents and siblings of children with ADHD com- markers. Journal of Affective Disorders, 133(3), 625–632. https:// bined and inattentive subtypes. Journal of Abnormal Psychology, doi. org/ 10. 1016/j. jad. 2011. 04. 041. 113(4), 614–625. https:// doi. org/ 10. 1037/ 0021- 843X. 113.4. 614 MacKinnon, D. P., & Fairchild, A. J. (2009). Current directions in Øie, M., Hovik, K. T., Andersen, P. N., Czajkowski, P. N., & Skogli, mediation analysis. Current Directions in Psychological Science, E. W. (2016). Gender Differences in the Relationship Between 18(1), 16–20. https:// doi. org/ 10. 1111/j. 14678 721. 2009. 01598 Changes in ADHD Symptoms, Executive Functions, and Self- Martel, M. M., Markon, K., & Smith, G. T. (2017). Research Review: and Parent-Report Depression Symptoms in Boys and Girls With Multi-informant integration in child and adolescent psychopathol- ADHD: A 2-Year Follow-Up Study. Journal of Attention Disor- ogy diagnosis. Journal of Child Psychology and Psychiatry, 58(2), ders. https:// doi. org/ 10. 1177/ 10870 54716 664407 116–128. https:// doi. org/ 10. 1111/ jcpp. 12611 Patterson, G. R., & Stoolmiller, M. (1991). Replications of a dual fail- Martel, M. M., Nikolas, M., & Nigg, J. T. (2007). Executive Function ure model for boys’ depressed mood. Journal of Consulting and in Adolescents With ADHD. Journal of the American Academy Clinical Psychology, 59(4), 491–498. https:// doi. org/ 10. 1037/ of Child & Adolescent Psychiatry, 46(11), 1437–1444. https://doi. 0022- 006X. 59.4. 491 org/ 10. 1097/ chi. 0b013 e3181 4cf953 Pelham, W. E., Gnagy, E. M., Greenslade, K. E., & Milich, R. (1992). McCarty, C. A., Mason, W. A., Kosterman, R., Hawkins, J. D., Lengua, Teacher Ratings of DSM-III-R Symptoms for the Disruptive L. J., & McCauley, E. (2008). Adolescent School Failure Predicts Behavior Disorders. Journal of the American Academy of Child Later Depression Among Girls. Journal of Adolescent Health, & Adolescent Psychiatry, 31(2), 210–218. https://do i.or g/10 .1097 / 43(2), 180–187. https://doi. or g/10. 1016/j. jadoh ealt h.2008. 01. 023 00004 583- 19920 3000- 00006 McClintock, S. M., Husain, M. M., Greer, T. L., & Cullum, C. M. Petersen, A. C., Crockett, L., Richards, M., & Boxer, A. (1988). A self- (2010). Association between depression severity and neurocogni- report measure of pubertal status: Reliability, validity, and initial tive function in major depressive disorder: A review and synthesis. norms. Journal of Youth and Adolescence, 17(2), 117–133. https:// Neuropsychology, 24(1), 9–34. https://doi. or g/10. 1037/ a0017 336 doi. org/ 10. 1007/ BF015 37962 McLaughlin, K. A., & Nolen-Hoeksema, S. (2011). Rumination as Pliszka, S. R. (2019). ADHD and Anxiety: Clinical Implications. a transdiagnostic factor in depression and anxiety. Behaviour Journal of Attention Disorders, 23(3), 203–205. https:// doi. org/ Research and Therapy, 49(3), 186–193. https://doi. or g/10. 1016/j. 10. 1177/ 10870 54718 817365 brat. 2010. 12. 006 Powell, V., Riglin, L., Hammerton, G., Eyre, O., Martin, J., Anney, R., Meinzer, M. C., & Chronis-Tuscano, A. (2017). ADHD and the devel- Rice, F. (2020). What explains the link between childhood ADHD opment of depression: Commentary on the prevalence, proposed and adolescent depression? Investigating the role of peer relation- mechanisms, and promising interventions. Current Develop- ships and academic attainment. European Child & Adolescent mental Disorders Reports, 4(1), 1–4. https:// doi. or g/ 10. 1007/ Psychiatry 1–11. https:// doi. org/ 10. 1007/ s00787- 019- 01463-w s40474- 017- 0106-1 Powers, A., & Casey, B. (2015). The adolescent brain and the emer- Meinzer, M. C., Pettit, J. W., & Viswesvaran, C. (2014). The co- gence and peak of psychopathology. Journal of Infant, Child, and occurrence of attention-deficit/hyperactivity disorder and unipolar Adolescent Psychotherapy, 14(1), 3–15. https:// doi. org/ 10. 1080/ depression in children and adolescents: A meta-analytic review. 15289 168. 2015. 10048 89 1 3 770 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Pritikin, J. N., Brick, T. R., & Neale, M. C. (2018). Multivariate normal Organizational Research Methods, 9(2), 194–201. https://doi. or g/ maximum likelihood with both ordinal and continuous variables, 10. 1177/ 10944 28105 285506. and data missing at random. Behavior Research Methods. https:// Violato, C., & Hecker, K. G. (2007). How to use structural equation doi. org/ 10. 3758/ s13428- 017- 1011-6 modeling in medical education research: A brief guide. Teaching Reitan, R. M., & Wolfson, D. (1992). Neuropsychological evaluation and Learning in Medicine, 19(4), 362–371. https:// doi. org/ 10. of older chilren. Neuropsychology Press.1080/ 10401 33070 15426 85 Roy, A., Oldehinkel, A. J., & Hartman, C. A. (2017). Cognitive Func- Wagner, S., Müller, C., Helmreich, I., Huss, M., & Tadić, A. (2015). A tioning in Adolescents with Self-Reported ADHD and Depres- meta-analysis of cognitive functions in children and adolescents sion: Results from a Population-Based Study. Journal of Abnor- with major depressive disorder. European Child & Adolescent Psy- mal Child Psychology, 45(1), 69–81. https:// doi. or g/ 10. 1007/ chiatry, 24(1), 5–19. https:// doi. org/ 10. 1007/ s00787- 014- 0559-2 s10802- 016- 0160-x Weber, E. B., Studeny, J., Kavanaugh, B. C., Cook, N. E., Gaudet, C. Schlomer, G. L., Bauman, S., & Card, N. A. (2010). Best practices E., McCurdy, K., & Holler, K. A. (2018). Pediatric Depression for missing data management in counseling psychology. Journal Symptoms, Executive Functioning Weaknesses, and Associated of Counseling Psychology, 57(1), 1–10. https:// doi. org/ 10. 1037/ Neuropsychological and Psychiatric Outcomes. Journal of Child a0018 082 and Family Studies, 27, 1661–1670. htt ps :// d oi. o r g/ 1 0. 10 07/ Seymour, K. E., Chronis-Tuscano, A., Iwamoto, D. K., Kurdziel, G., & s10826- 017- 0999-7 MacPherson, L. (2014). Emotion Regulation Mediates the Associ- Wechsler, D. (2002). Wechsler Individual Achievement Test—2nd ation Between ADHD and Depressive Symptoms in a Community Edition. Examiner’s Manual. San Antonio, TX: Psychological Sample of Youth. Journal of Abnormal Child Psychology, 42(4), Corportation. 611–621. https:// doi. org/ 10. 1007/ s10802- 013- 9799-8 Wechsler, D. (2003). Wechsler Intelligence Scale for Children, Fourth Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab-Stone, Edition: Administration and Scoring Manual. Psychological M. E. (2000). NIMH Diagnostic Interview Schedule for Children Corporation. Version IV (NIMH DISC-IV): Description, Differences From Weissman, A. S., Chu, B. C., Reddy, L. A., & Mohlman, J. (2012). Previous Versions, and Reliability of Some Common Diagnoses. Attention mechanisms in children with anxiety disorders and in Journal of the American Academy of Child & Adolescent Psy- children with attention deficit hyperactivity disorder: Implications chiatry, 39(1), 28–38. https:// doi. org/ 10. 1097/ 00004 583- 20000 for research and practice. Journal of Clinical Child & Adolescent 1000- 00014 Psychology, 41(2), 117–126. https:// doi. org/ 10. 1080/ 15374 416. Snyder, H. R., Friedman, N. P., & Hankin, B. L. (2019). Transdi-2012. 651993 agnostic mechanisms of psychopathology in youth: Executive Weston, R., & Gore, P. A., Jr. (2006). A brief guide to structural equa- functions, dependent stress, and rumination. Cognitive Ther- tion modeling. The Counseling Psychologist, 34(5), 719–751. apy and Research, 43(5), 834–851. https:// doi. or g/ 10. 1007/ https:// doi. org/ 10. 1177/ 00110 00006 286345 s10608- 019- 10016-z Wiebe, S. A., Sheffield, T., Nelson, J. M., Clark, C. A. C., Chevalier, Snyder, H. R., & Hankin, B. L. (2016). Spiraling out of control: Stress N., & Espy, K. A. (2011). The structure of executive function in generation and subsequent rumination mediate the link between 3-year-olds. Journal of Experimental Child Psychology, 108(3), poorer cognitive control and internalizing psychopathology. Clini- 436–452. https:// doi. org/ 10. 1016/j. jecp. 2010. 08. 008 cal Psychological Science, 4(6), 1047–1064. https:// doi. org/ 10. Willcutt, E. G., Doyle, A. E., Nigg, J. T., Faraone, S. V., & Pennington, 1177/ 21677 02616 633157. B. F. (2005). Validity of the Executive Function Theory of Atten- Snyder, A. M., Maruff, P., Pietrzak, R. H., Cromer, J. R., & Snyder, tion-Deficit/Hyperactivity Disorder: A Meta-Analytic Review. P. J. (2008). Effect of treatment with stimulant medication on Biological Psychiatry, 57(11), 1336–1346. https:// doi. or g/ 10. nonverbal executive function and visuomotor speed in children 1016/j. biops ych. 2005. 02. 006 with attention deficit/hyperactivity disorder (ADHD). Child Wolraich, M. L., Hagan, J. F., Allan, C., Chan, E., Davison, D., Earls, Neuropsychology, 14(3), 211–226. htt ps:// d oi. or g / 10. 10 80/ M., & Zurhellen, W. (2019). Clinical practice guideline for the 09297 04070 12200 05 diagnosis, evaluation, and treatment of attention-deficit/hyper - Snyder, H. R., Miyake, A., & Hankin, B. L. (2015). Advancing under- activity disorder in children and adolescents. Pediatrics, 144(4). standing of executive function impairments and psychopathology: https:// doi. org/ 10. 1542/ peds. 2019- 2528 Bridging the gap between clinical and cognitive approaches. Fron- Zelazo, P. D., Anderson, J. E., Richler, J., Wallner-Allen, K., Beaumont, tiers in Psychology, 6, 328. https:// doi. org/ 10. 3389/ fpsyg. 2015. J. L., Conway, K. P., & Weintraub, S. (2014). NIH Toolbox Cogni- 00328 tion Battery (CB): Validation of Executive Function Measures in Thapar, A., Collishaw, S., Pine, D. S., & Thapar, A. K. (2012). Depres- Adults. Journal of the International Neuropsychological Society, sion in adolescence. The Lancet, 379(9820), 1056–1067. https:// 20(6), 620–629. https:// doi. org/ 10. 1017/ S1355 61771 40004 72 doi. org/ 10. 1016/ S0140- 6736(11) 60871-4 Zelazo, P. D., & Cunningham, W. A. (2007). Executive function: mech- Timbremont, B., Braet, C., & Dreessen, L. (2004). Assessing Depres- anisms underlying emotion regulation. In J. J. Gross (Ed.), Hand- sion in Youth: Relation Between the Children’s Depression Inven- book of emotion regulation (p. 135–158). The Guilford Press. tory and a Structured Interview. Journal of Clinical Child & Zhou, Q., Chen, S. H., & Main, A. (2012). Commonalities and Differ - Adolescent Psychology, 33(1), 149–157. https:// doi. org/ 10. 1207/ ences in the Research on Children’s Effortful Control and Execu- S1537 4424J CCP33 01_ 14 tive Function: A Call for an Integrated Model of Self-Regulation. Twenge, J. M. J., Thomas, E., Rogers, M. L., & Martin, G. N. (2017). Child Development Perspectives, 6(2), 112–121. https:// doi. org/ Increases in Depressive Symptoms, Suicide-Related Outcomes, 10. 1111/j. 1750- 8606. 2011. 00176.x and Suicide Rates Among U.S. Adolescents After 2010 and Links to Increased New Media Screen Time. Clinical Psychological Sci- Publisher's Note Springer Nature remains neutral with regard to ence, 6(1), 3–17. https:// doi. org/ 10. 1177/ 21677 02617 723376. jurisdictional claims in published maps and institutional affiliations. Vandenberg, R. J. (2006). Introduction: statistical and methodological myths and urban legends: Where, pray tell, did they get this idea? 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Research on Child and Adolescent Psychopathology Springer Journals

Childhood ADHD and Executive Functioning: Unique Predictions of Early Adolescent Depression

Loading next page...
 
/lp/springer-journals/childhood-adhd-and-executive-functioning-unique-predictions-of-early-tkvQqnaDVu

References (101)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
0091-0627
eISSN
2730-7174
DOI
10.1007/s10802-021-00845-6
Publisher site
See Article on Publisher Site

Abstract

Given the increasing prevalence of adolescent depression, identification of its early predictors and elucidation of the mecha- nisms underlying its individual differences is imperative. Controlling for baseline executive functioning (EF), we tested separate ADHD dimensions (i.e., inattention, hyperactivity-impulsivity) as independent predictors of early adolescent depres- sion, including temporally-ordered causal mediation by academic functioning and social problems, using structural equation modeling. At baseline, participants consisted of 216 children (67% male) ages 6–9 years old with (n = 112) and without (n = 104) ADHD who subsequently completed Wave 2 and 3 follow-ups approximately two and four years later, respectively. Predictors consisted of separate parent and teacher ratings of childhood ADHD and laboratory-based assessments of key EF domains. At Wave 2, parents and teachers completed normed rating scales of youth academic and social functioning; youth completed standardized assessments of academic achievement. At Wave 3, youth self-reported depression. Baseline inattention positively predicted early adolescent depression whereas childhood hyperactivity-impulsivity and EF did not. Neither academic nor social functioning significantly mediated predictions of depression from baseline ADHD and EF. We consider prediction of early adolescent depression from inattention, including directions for future intervention and preven- tion research. Keywords ADHD · Executive functioning · Depression · SEM In the United States, the prevalence of major depressive increased precipitously since 2010 (Twenge et al., 2017), disorder has risen consistently since 1940 and currently including suicide being the second leading cause of death incurs over $210 billion annually, a 21% increase since 2005 among 10–24-year-old individuals (Center for Disease Con- (Greenberg et al., 2015). The increase in prevalence coin- trol and Prevention, 2016). To reduce its considerable clini- cides with an earlier age of onset (Birmaher & Brent, 2007): cal and public health burden, characterizing predictors of 2% of children and up to 8% of adolescents meet diagnos- early adolescent depression is a priority. tic criteria for major depressive disorder with an additional Although understudied relative to co-occurring external- 5–10% of youth experiencing subclinical symptoms. Cru- izing problems, cross-sectional associations of ADHD with cially, 30% of youth with major depressive disorder expe- depression (i.e., heterotypic comorbidity) are well-established rienced suicidal ideation in the past 12 months (Avenevoli (Humphreys et al., 2013; Meinzer et al., 2014; Seymour et al., et  al., 2015). Moreover, depression and suicidality have 2014). Further, even with control of demographic and clinical factors (e.g., maternal depression), childhood and adolescent ADHD each uniquely predict depression through young adult- * Michelle C. Fenesy hood (Chronis-Tuscano et al., 2010; Meinzer et al., 2016), and * Steve S. Lee children with ADHD frequently experience recurrent depres- stevelee@psych.ucla.edu sive episodes (Chronis-Tuscano et al., 2010). ADHD is opti- mally conceptualized as the quantitative extreme of inattention Department of Child & Adolescent Psychiatry, New York- Presbyterian Hospital, Columbia University Irving Medical and hyperactivity-impulsivity, but ADHD is typically exam- Center, 622 W. 168th St., New York, NY 10032, USA ined dichotomously, thereby preventing specific inferences Department of Psychology, University of California, about the relative association of inattention and hyperactiv- Box 951563, 1285 Franz Hall, Los Angeles, CA 90095-1563, ity-impulsivity with respect to depression. In key exceptions, USA Vol.:(0123456789) 1 3 754 Research on Child and Adolescent Psychopathology (2022) 50:753–770 inattention—not hyperactivity-impulsivity—was uniquely consider the prospective association of EF with depression associated with depression in 5–10-year-old youth (Fenesy & during the transition from childhood to early adolescence Lee, 2019; Humphreys et al., 2013), although hyperactivity- using latent variable approaches. Specifically, this evidence impulsivity predicted depression indirectly via emotion regu- will inform the potential need/utility of EF and/or behavio- lation (Seymour et al., 2014). In addition to the independent ral interventions prior to depression onset in adolescence. prediction of depression from separable dimensions of ADHD, To clarify whether ADHD and EF independently predict there is a pressing need to identify additional factors exacerbat- youth depression, ADHD and EF must be examined con- ing or mitigating risk. currently. ADHD principally reflects executive dysfunction Several lines of evidence suggest the plausibility of exec- (Barkley, 1997; Castellanos & Tannock, 2002) with meta- utive functioning (EF) as a risk factor with respect to youth analytic evidence that youth with ADHD exhibit impaired depression. Although a precise definition of EF is somewhat EF in community-based and clinical samples (Willcutt intractable (Barkley, 2012), there is general consensus that et  al., 2005). Although EF is commonly proposed as a EF encapsulates related cognitive processes (i.e., inhibitory cognitive endophenotype or risk factor for ADHD (Gau & control, working memory, set shifting) associated with the Shang, 2010; Nigg et al., 2004), controversy remains as prefrontal cortex that support goal-directed behaviors (Bar- there is considerable heterogeneity in EF among individu- kley, 2012; Miyake et al., 2000; Nigg, 2017). Importantly, als with ADHD and not all children with a diagnosis of individuals with depression consistently exhibit poor emo- ADHD exhibit EF deficits (Kofler et al., 2019; Willcutt, tion regulation, which is directly supported by EF among et al., 2005). Further, the broader field of EF research seeks adolescents and adults (Gyurak et al., 2012; Joormann & to clarify whether individual differences in EF represent Gotlib, 2010; Wagner et al., 2015). Further, emotion regula- a risk-factor for psychopathology, a correlate of psycho- tion tasks are sensitive to differential patterns of activation pathology, or a result of psychopathology across multiple in prefrontal cortex regions associated with EF that suppress disorders (Snyder et al., 2015). Therefore, it is reasonable emotional responses from the limbic system (Zelazo & Cun- to evaluate EF as a correlate of ADHD. The present study ningham, 2007). Moreover, rumination (i.e., engaging in aims to test the prospective prediction of depression from repetitive thinking patterns that amplify emotions) is central ADHD and EF, as controlling for comorbidities is recom- to the onset and maintenance of depression (McLaughlin mended when evaluating EF in relation to depressive symp- & Nolen-Hoeksema, 2011); elevated rumination mediated toms (McClintock et al., 2010). the cross-sectional association between impaired EF (i.e., Relatively few studies simultaneously examined the set shifting) and depressive symptoms in typically develop- association of ADHD and EF with youth depression. ing adolescents (Dickson et al., 2017). Finally, individuals For example, working memory, a key dimension of EF, with poor EF often experience difficulties with goal attain- was impaired among adolescents with ADHD + depres- ment and planning across domains, thus catalyzing stress sion relative to youth with depression alone (Roy et al., generation and potentially subsequent depression (Snyder 2017), but depression and ADHD were exclusively youth & Hankin, 2016). Overall, there is growing evidence that self-reported, despite the utility of multiple informants EF is correlated with youth depression, although findings (Martel et al., 2017). Moreover, controlling for inatten- are frequently based on cross-sectional designs (e.g., Favre tion and hyperactivity-impulsivity, EF dimensions (i.e., et al., 2008) and consist of broad age ranges (e.g., child- working memory, mental flexibility, inhibition) were hood through adolescence), preventing specific inferences unrelated to parent- or self-reported depression among about EF as a correlate versus risk factor for early-adoles- 9–16-year-old youth (Øie et al., 2016). However, latent cent depression. Further, separable EF dimensions require variable approaches improve measurement error; latent rigorous measurement strategies to adequately capture its EF was positively associated with parent-rated depres- multidimensionality. Despite their empirical separability, sion among 5–10-year-old children (Fenesy & Lee, 2019). boundaries among EF dimensions may be less distinct in Among hospitalized inpatients, youth with depression and children, giving rise to a unified set of cognitive processes EF deficits had elevated ADHD compared to children with (Lee et al., 2013; Wiebe et al., 2011), thus necessitating depression alone (Weber et al., 2018), although directional careful methodological and developmental considerations. inferences were unclear, further underscoring the need for Indeed, when derived from a single latent variable, poor multi-method/informant longitudinal studies. preschool EF prospectively predicted early school-age In addition to examining the independent prediction depression symptoms, controlling for baseline depression of depression from childhood ADHD and EF, it is neces- (Nelson et al., 2018). However, this study did not extend sary to investigate ADHD diagnostic status as a moderator into early adolescence, an important limitation given the of the prospective association between childhood EF and precipitous increase in depression secondary to pubertal early adolescent depression to evaluate possible interactive onset (Avenevoli et al., 2015). There is a pressing need to effects. ADHD and EF interact to predict other key outcomes 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 755 such as academic achievement. For example, elevated inat- evaluating risk for depression. We also tested whether tentive symptoms and poor EF predicted special education academic and social functioning temporally mediated (Diamantopoulou et al., 2007). Thus, inclusion of ADHD predictions from inattention, hyperactivity-impulsivity, diagnostic status as a moderator will reveal whether the and EF. Consistent with recommendations for SEM, if prospective association between EF and depression varies initial models did not fit the data, we thoughtfully respec- according to ADHD diagnostic status. Testing independent ified the models (Weston & Gore Jr., 2006; Violato & and interactive prediction of depression from ADHD and EF Hecker, 2007.) We hypothesized that inattention would will elucidate their contributions, informing how specific positively predict whereas EF would inversely predict assessment and intervention approaches must be tailored. early adolescent depression. Lastly, we expected that aca- Beyond predictive models, innovations in intervention demic and social functioning would each significantly require elucidation of underlying mechanisms of predictions mediate predictions of adolescent depression from inat- of psychopathology from EF and ADHD (Meinzer & Chronis- tention, hyperactivity-impulsivity, and EF. Tuscano, 2017; Snyder et al., 2015; Zhou et al., 2012), a critical next step in EF research (Snyder et al., 2019). Both dual failure and competency-based models propose academic and social difficulties as key factors in the development of youth depres- Methods sion (Cole, 1991; Patterson & Stoolmiller, 1991). Whereas the dual failure model hypothesized that academic failure and peer Participants rejection mediated predictions of depression from early conduct problems (McCarty et al., 2008; Patterson & Stoolmiller, 1991), Participants were 216 children (67% male) with (n = 112) competency-based models contend that children internalize and without (n = 104) ADHD and their families. Drawn negative feedback from the environment, which adversely from a large metropolitan city in the Western U.S., children affects their self-esteem and increases vulnerability to depres- were recruited from referrals from pediatric offices, men- sion (Harter & Marold, 1994). Additionally, ADHD and EF tal health service agencies, and local schools. Inclusion each predicted academic achievement (Best et al., 2011; Miller criteria were English fluency and living with a biological et al., 2012) and social functioning (Diamantopoulou et al., caregiver at least halftime; exclusion criteria consisted of 2007; Huang-Pollock et al., 2009), representing plausible medi- an IQ below 70 or seizure, autism spectrum, or other neu- ators of depression from ADHD and EF. In particular, these rological condition. Table 1 summarizes key demographic theoretically-justified mediators should be subjected to strin- data for participants. The sample was ethnically diverse gent tests of causal mediation wherein constructs are tempo- (51.93% Caucasian; 8.33% African American; 12.04% rally ordered relative to predictors and outcomes (MacKinnon Hispanic; 3.70% Asian; 22.22% Mixed; 1.85% Other/ & Fairchild, 2009). Unknown) and 29.95% of families had an annual income Overall, the extant literature does not consider sepa- of $70,000 or less. At baseline (i.e., Wave 1), participat- rable dimensions of ADHD in conjunction with EF, and ing children were 6- to 9-years-old (M = 7.39, SD = 1.07). prospective studies are necessary to determine whether Families were invited to a follow-up approximately two EF represents a predictor versus correlate of early ado- years later (i.e., Wave 2) at which point youth ranged in lescent depression. Lastly, identification of underlying age from 7- to 13- years old (M = 9.68, SD = 1.27). A mechanisms is necessary to identify targets for interven- final follow-up (i.e., Wave 3) was conducted about two tion to reduce the risk of youth depression. To address years after Wave 2 when they were 9- to 15-years old these significant limitations hindering innovations in (M = 12.07, SD = 1.30). Data from all three waves were understanding the antecedents and mediators of early utilized in the current study. adolescent depression, the present study employed Of the 216 Wave 1 participants, 89.35% (n = 193) and structural equation modeling (SEM) to evaluate inat- 80.10% (n = 173) participated in the Wave 2 and 3 assess- tention, hyperactivity-impulsivity, and EF as independ- ment, respectively. There were no significant differences in ent, prospective predictors of early adolescent depres- child age, sex, race, or baseline ADHD symptoms between sion controlling for key covariates (see Data Analytic children who participated at Wave 1 and Wave 2 or Wave 1 Procedure). Additionally, we tested ADHD diagnostic and Wave 3 (ps > 0.15). Youth who participated at Waves 2 status as a moderator to examine whether the prospec- and/or 3 had a higher mean IQ (ps < 0.01) than those who tive prediction of depression from EF differs for chil- did not complete follow-ups. Overall, missing data ranged dren with and without ADHD. Determining whether the from approximately 72% on a youth self-rated depression prospective association between EF and depression var- measure to 0% on baseline ADHD data from a structured ries according to group status would inform clinicans diagnostic interview. Maximum likelihood estimation for whom the screening of EF is most important when addressed missing data (described further below). 1 3 756 Research on Child and Adolescent Psychopathology (2022) 50:753–770 quickly as possible to sequentially connect numbered cir- Procedures cles from 1 to 15 without making errors. Part B (TMT-B) involves alphabetically and numerically alternating between Families initially completed a phone screen to determine eligibility. Parents and teachers of eligible participants were connecting numbers 1 through 13 and letters A through L. The time (min) to complete TMT-B reflects set shifting, mailed rating scales of child functioning and were invited to a lab-based assessment conducted by well-trained gradu- with longer completion times indicating worse EF (Reitan & Wolfson, 1992). TMT-B significantly differentiated youth ate students in clinical psychology or B.A. level staff. After obtaining parental consent and youth assent, parents com- with ADHD and controls (Martel et al., 2007) and loaded onto a latent set shifting factor in children (Arán Filippetti pleted structured diagnostic interviews and rating scales to assess youth social-emotional functioning, whereas youth & Richaud, 2017). To aid in interpretation, completion time was reverse coded so that higher scores represented better completed standardized tests of cognition, EF, and aca- demic achievement as well as measures of social-emotional EF across all tests. Participants completed the Children’s Version of the functioning. Researchers requested that children complete a medication washout for the day of testing, however this Golden Stroop (Golden et al., 2003) to assess inhibitory con- trol. In the first condition, participants read as many words was optional. A total of 32 children with ADHD at baseline were prescribed medication, and 13 of these children took (i.e., red, blue, green) as possible in 45 s. The second condi- tion required naming different colors of ink (i.e., red, blue, a simulant medication on the day of Wave 1 testing (6.02% of the total sample; N = 216). Wave 2 and Wave 3 follow-up green) in the same 45 s timeframe. In the third and final con- dition (i.e., Color-Word [C-W]), the color names are printed assessments consisted of parallel procedures to assess simi- lar constructs. The University of California, Los Angeles in discordant colors. The total score on the C-W condition is the number of ink colors named in 45 s. The Stroop C-W IRB approved all study procedures. shows strong criterion validity in children and adults (Arán Filippetti & Richaud, 2017; Miyake et al., 2000). The raw Measures C-W score was used to estimate inhibitory control. Children completed the Digit Span subtest from the Wave 1 Predictors Wechsler Intelligence Scale for Children-IV (Wechsler, 2003). In the forward condition, the examiner presents a ADHD. We administered the ADHD module of the Diag- string of numbers aloud; participants must recall the num- bers in the proper order. The backward condition requires nostic Interview Schedule for Children (DISC-IV) Parent Edition (Shaffer et al., 2000) to parents to assess Diagnos- recalling the numbers in the reverse order. The Digit Span Backwards (DSB) raw score was used to estimate working tic and Statistical Manual of Mental Disorders DSM-IV ADHD. The DISC-IV is a computer-assisted, structured memory given that it loaded more strongly on a latent work- ing memory factor of EF relative to the forward condition interview of symptoms, onset, and impairment with strong psychometric properties. The ADHD module has high inter- (Arán Filippetti & Richaud, 2017). Normed scores do not exist for the raw backwards portion of the Digit Span subtest. nal consistency (α = 0.84; Schaffer et al., 2000). Given its superior predictive validity (Fergusson & Horwood, 1995), Therefore, raw scores were utilized for all EF measures and we controlled for age on the latent EF variable. we analyzed the number of inattention (0–9) and hyperactiv- ity-impulsivity symptoms (0–9) from the DISC-IV. Parents Wave 2 Mediators and teachers also rated ADHD symptoms on the Disruptive Behavior Disorder Rating Scale (DBD). Responses ranged Academic Functioning. Youth completed the Word Reading from 0 (not at all) to 3 (very much; Pelham et al., 1992), yielding separate inattention and hyperactivity-impulsivity and Math Reasoning subtests from the Wechsler Individual Achievement Test Second Edition (WIAT-II; Wechsler, totals (0–27). The inattention and hyperactivity-impulsiv- ity scales demonstrated high reliability within the present 2002), which assessed phonological awareness/decoding and mathematical problem solving, respectively. Separate sample (parent-rated inattention  α = 0.94; teacher-rated inattention α = 0.93; parent-rated hyperactivity-impulsivity Word Reading and Math Reasoning standard scores were employed. α = 0.92; teacher-rated hyperactivity-impulsivity α = 0.94). Raw scores were used for inattention and hyperactivity- Parents and teachers completed the Child Behavior Checklist (CBCL) and Teacher Report Form (TRF), respec- impulsivity variables as adjusted norms are not provided for these measures. tively (Achenbach & Rescorla, 2001). The CBCL and TRF consist of 113-items with behaviors rated from 0 (not true) Executive Functioning (EF). For set shifting, we admin- istered the child version of the Trail Making Test (Reitan to 2 (very true/often true). The CBCL includes School Competence items, ranging from 0–6, to assess grades, & Wolfson, 1992). On Part A, participants drew lines as 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 757 class placement, grade repetition, and other problems in the measure due to time constraints. We used the age- and the school setting. Items from the school competence scale sex-adjusted Affective Problems T-score as another meas- differentiate between youth from clinic referred and non- ure of self-reported depression. According to the manual, referred samples and has an acceptable reliability according this subscale demonstrates satisfactory reliability (α = 0.81; to the manual (α = 0.63; Achenbach & Rescorla, 2001). The Achenbach & Rescorla, 2001). All three variables for the TRF Academic Performance scale is the mean of the child’s proposed Wave 3 latent depression factor provided adjusted performance across academic subjects with a test–retest reli- values; therefore, we utilized T-scores. ability of r = 0.90 (Achenbach & Rescorla, 2001). We used the CBCL School Competence and the TRF Academic Per- formance T-scores to create a latent academic functioning Data Analytic Procedures variable. All four observed variables for the proposed aca- demic functioning latent factor provided adjusted values; We employed SEM to test Wave 1 ADHD (i.e., inattention, thus, T-scores were utilized. hyperactivity-impulsivity) and EF as predictors of Wave Social Problems. Parents and teachers completed the 3 depression (Fig. 1A). Next, we tested ADHD diagnostic Dishion Social Preference Scale, a three-item (5-point status as a moderator to evaluate whether the prospective metric) measure of peer acceptance, rejection, and being prediction of depression from EF differs according to ADHD ignored (Dishion, 1990). Negative social preference was diagnostic status. We finally examined academic function- calculated by subtracting the reject from the accept rating ing and social problems as mediators of the prospective and then reversing scoring the difference (Humphreys et al., prediction of Wave 3 depression from baseline inattention, 2013; Lee & Hinshaw, 2006). In addition to the academic hyperactivity-impulsivity, and EF (Fig. 1B). To test these functioning scales, the CBCL and TRF yield parallel Social proposed models, we implemented full information maxi- Problems scales, which are reliable (α = 0.82 for both) and mum likelihood (FIML; Enders, 2010), which performs well valid (Achenbach & Rescorla, 2001). These two variables even with extreme missingness (e.g., 50%; Schlomer et al., were coupled with negative social preference to collectively 2010). We addressed FIML requirements that data are Miss- estimate social problems. Because the Dishion Social Prefer- ing at Random (MAR) or Missing Completely at Random ence Scale does not provide age and sex adjusted norms, we (MCAR) and multivariate normality (Pritikin et al., 2018). used the raw scores for all four social variables for consist- Little’s Test suggested that the data violated the assumption ency across this proposed latent variable. of MCAR. Thus, we determined appropriate auxiliary vari- ables to include to ensure data were missing at random and Wave 3 Youth Self‑Report Outcomes to improve power (Enders, 2010). Auxiliary variables are not central to the specific research question; rather, auxiliary Depression. Youth completed the 27-item Children’s variables may be highly correlated with missingness on the Depression Inventory (CDI; Kovacs, 1992), rating descrip- study variables or the included variables (Enders, 2010). We tions from the past two weeks (e.g., “I feel like crying every tested the correlation between potential auxiliary variables day,” “I feel like crying many days,” “I feel like crying once and variables relevant to the analyses as well as missingness in a while”). Each item is scored from 0 to 2, providing age- on variables for the analyses. Several auxiliary variables sig- and sex-adjusted T-scores. The CDI has shown strong con- nificantly correlated with missingness and/or variables for vergent validity with internalizing and disruptive behavior analyses and were included in all SEM models (Table 2). For (Timbremont et al., 2004). The CDI demonstrated accept- example, Wave 1 WIAT-II Math Reasoning standard score able reliability within our sample (α = 0.85). is not a variable in our analyses; however, it was determined The 47-item Revised Children’s Anxiety and Depression to be an appropriate auxiliary variable because it was sig- Scale (RCADS) includes five normed subscales (Chorpita nificantly correlated with: Wave 2 WIAT-II Word Reading et al., 2000). Items were rated from 0 to 3, reflecting never, standard score, Wave 2 WIAT-II Math Reasoning standard sometimes, often, or always, respectively. The sex- and score, Wave 2 CBCL School Competence T-score missing- grade-adjusted T-score from the Major Depressive Dis- ness, Wave 2 CBCL Social Problems raw score missingness, order subscale, which previously correlated with the CDI and Wave 2 Parent Negative Social Preference missingness. (Chorpita et al., 2000) and shows satisfactory internal con- For predictive models, Wave 2 academic and social mediators sistency within this sample (α = 0.80), was analyzed. were implemented as auxiliary variables. Although Mardia’s The 113-item Youth Self Report (YSR; Achenbach & test of Skewness [191.46, χ (969) = 936.96, p = 0.76] did Rescorla, 2001) is parallel to the CBCL and TRF. The YSR not violate the criteria for multivariate normality, Mardia’s is normed on a sample of 11–18-year-old youth (Achenbach test of Kurtosis [296.59, χ (1) = 7.02, p < 0.01] violated & Rescorla, 2001). Although 154 youth in this study were the assumption for the simplest model; therefore, we imple- 11 years or older at Wave 3, only 61 participants completed mented maximum likelihood robust procedures in Mplus 1 3 758 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Wave 1 Inattention Wave 1 Wave 3 Hyperactivity/ Depression Impulsivity Wave 1 Executive Functioning A) Wave 1 Wave 2 Inattention Academic Fx. Wave 1 Wave 3 Hyperactivity/ Depression Impulsivity Wave 1 Wave 2 Executive Social Problems Functioning B) Fig. 1 A) Proposed predictive model, testing the prospective predic- sion and baseline inattention, hyperactivity/impulsivity, and executive tion of Wave 3 self-reported depression from Wave 1 inattention, functioning. For A and B, circles indicate latent variables. Table  4 hyperactivity-impulsivity, and executive functioning. B) Proposed includes observed variables proposed to derive latent variables. mediational model, testing Wave 2 academic functioning and social Covariates are not included problems as mediators of the relationship between Wave 3 depres- 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 759 (Muthén & Muthén, 1998–2010) to address non-normality status. Such measures are reliably associated with pubertal and accommodate missing data. development as evaluated by the Tanner Scales (Carskadon Prior to running SEM, we conducted separate confirma- & Acebo, 1993; Petersen et al., 1988) For the present study, a tory factor analyses on each Wave 1 (i.e., inattention, hyper- quantitative score described by Carskadon and Acebo (1993) activity-impulsivity, EF), Wave 2 (i.e., academic function- was implemented. Each characteristic was rated from 1- 4 ing, social problems), and Wave 3 (i.e., depression) latent (1 = not yet started changing, 2 = has barely started chang- variables (see preliminary analyses). For proposed latent ing, 3 = change is definitely underway, 4 = change seems variables with three indicator variables (i.e., inattention, completed) with the exception of a menstruation item for hyperactivity-impulsivity, EF, depression) we evaluated girls, which was coded as 1 for no and 4 for yes. The aver- factor loadings to ensure that they exceeded the guidelines age of the items was used to measure puberty for the current of at least 0.3 (Brown, 2014). For latent factors with four study (M = 2.66, SD = 0.73). Finally, if inattention signifi- indicator variables (i.e., academic achievement, social cantly predicted depression, we added anxiety as a covariate problems), fit indices were examined. For predictive and on the latent inattention factor to strengthen specific infer - mediational models, multiple fit indices were evaluated. A ences (Pliszka, 2019). In addition to completing the CBCL non-significant chi-square and comparative Fit Index (CFI) at Wave 2 to assess youth academic and social functioning values ≥ 0.95 indicate good fit. R oot mean square error of (see above for details), parents also completed this measure at approximation (RMSEA) estimates model fit with control of baseline (i.e., Wave 1). The T-score from the Wave 1 Anxiety sample size and per degrees of freedom where values ≤ 0.06 Problems subscale was used to assess child anxiety. Baseline are acceptable (Hu & Bentler, 1999). Finally, a value of 0.08 anxiety was included as a covariate on the latent inattention or less for standardized root mean square residual (SRMR) variable in models where inattention emerged as a significant suggests good model fit (Hu & Bentler,  1999). predictor. Models conservatively accounted for age, sex, baseline depression, SES, and pubertal status. Age and sex were accounted for as the depression outcome observed variables Results adjusted for these factors. Additionally, we controlled for baseline age on the EF latent factor as EF improves with Preliminary Analyses child development. Baseline depression, SES, and pubertal status were included as covariates. Consistent with other Co-occurrence between ADHD and Depression. To pro- studies (e.g., Lawson & Farah, 2017), we utilized family vide additional support for the use of dimensional concep- income and parent education to approximate SES. To cap- tualization of ADHD, we tested for significant differences in ture family income, we used a binary variable where $75,000 Wave 1 and Wave 3 depression based upon baseline ADHD or less was coded as 0 and $75,001 or more was coded as 1 diagnostic status. Because considerable discrepancies due to the fact that 66% of the sample with observed data exist between parent- and self-reported youth internalizing on this variable had an income of more than $75,000. We symptoms (De Los Reyes et al., 2015; Johnston & Murray, also considered mother and father education in estimating 2003; Lewis et al., 2014), youth self-reported depression SES. Education level was categorized as follows: 1 = eighth measures were examined. On the CDI at baseline, 11 of 114 grade or less, 2 = some high school, 3 = high school graduate participants with complete data on this measure were at or or GED, 4 = some college or post-high school, 5 = college above a T-score of 60 (i.e., high average and above). Seven graduate, 6 = advanced graduate or professional degree. The of those 11 participants met diagnostic criteria for ADHD average value of mother and father education, which ranged on the DISC-IV. This did not represent a significant differ - from 1.5-6 (M = 4.92, SD = 0.91), was used; a single value ence in Wave 1 depression according to baseline ADHD from mother or father was utilized in circumstances when status [χ (1) = 1.08, p = 0.30]. At Wave 3, a total of 149 education data was present for one parent. Income and parent participating youth completed the CDI. Seven of those par- education variables were included in all SEM models. Chil- ticipants met or exceeded a T-score of 60, with four of those dren completed the CDI (Kovacs, 1992), as described above seven youth meeting diagnostic criteria for ADHD at Wave at baseline in addition to Wave 3. Each item is scored from 0 1. Again, there was not a significant difference in Wave 3 to 2, with 2 representing higher depression severity. The child depression on the CDI according to baseline ADHD diag- age and sex adjusted T-score was used; norms for seven-year- nostic status [χ (1) = 1.08, p = 0.30]. Similarly, there were olds were used to calculate T-scores for six-year-old children no significant differences between Wave 3 depression and in the current sample. At Wave 3, children completed the baseline ADHD diagnosis according to youth self-report on Pubertal Development Scale (Petersen et al., 1988), a six- the RCADS depression subscale [χ (1) = 1.82, p = 0.18]. item self-report questionnaire used to assess pubertal status Specifically, on the Wave 3 RCADS, two out of 148 youth in males and females and noninvasively assesses pubertal met or exceeded the clinical cutoff (T-score above 70) for 1 3 760 Research on Child and Adolescent Psychopathology (2022) 50:753–770 depression, both of whom had a diagnosis of ADHD at base- inattention, hyperactivity-impulsivity, and EF as prospective line. The fact that we did not observe a significant differ - predictors of early adolescent depression, we modified the ence between those with and without ADHD at baseline and model as is recommended within SEM (Weston & Gore Jr., Wave 1 or Wave 3 depression demonstrates the importance 2006; Violato & Hecker, 2007). Due to concerns regarding of utilizing a dimensional conceptualization of ADHD when collinearity, we next tested a model evaluating total ADHD examining ADHD as a risk factor for depression. symptoms and EF as predictors of Wave 3 depression. Confirmatory Factor Analyses. Factor loadings from Model 2. We created three total ADHD symptom scores all confirmatory factor analyses are provided in Table  4. by summing the inattention and hyperactivity-impulsivity For the inattention, hyperactivity-impulsivity, and EF latent dimensions on the DISC-IV (M = 7.88, SD = 5.55), parent variables, standardized beta coefficients exceeded the rec- DBD (M = 20.46, SD = 14.00), and teacher DBD (M = 17.31, ommended cutoff of 0.3 (Brown, 2014). The latent vari- SD = 15.25). Confir matory factor analyses revealed that ables for Wave 2 academic functioning and social problems these ADHD variables had acceptable factor loadings consisted of four observed variables each. The academic (> 0.49). We regressed the latent Wave 3 self-reported functioning factor demonstrated good fit across multi- depression outcome on Wave 1 ADHD total symptom and ple indices (Table 4): χ (2) = 4.42, p = 0.10, CFI = 0.98, EF latent variables including described covariates. This and SRMR = 0.03, although the RMSEA value was 0.08. model improved fit (Table  5): χ (59) = 78.70, p = 0.04, Because RMSEA cutoffs are vulnerable to poor fit in mod- CFI = 0.97, RMSEA = 0.04, SRMR = 0.07. Although the els with few degrees of freedom (Kenny et al., 2015), but chi-square was significant, this measure of fit is often unre- other fit indices for this academic functioning latent vari- liable (Vandenberg, 2006), so we prioritized other fit indi- able were acceptable, a latent variable was implemented for ces. For this alternative structural model, CFI, RMSEA, academic functioning. In contrast, the social problems latent SRMR all demonstrated good fit where ADHD symp- variable showed poor model fit [χ (2) = 43.51, p < 0.001, toms (β = 0.23, SE = 0.10, p = 0.01), but not EF (β = 0.02, CFI = .71, RMSEA = 0.33 and SRMR = 0.09]. Therefore, SE = 0.13, p = 0.87), positively predicted depression. Model we created two composite variables to estimate parent- and 2 accounted for 5.3% of the variance in Wave 3 depression. teacher-rated social problems, respectively (see Table  3) Model 3. To improve specificity, we next tested EF by z-scoring the CBCL Social Problems raw score and the and inattention as predictors of Wave 3 depression with Dishion Negative Social Preference rating followed by cal- identical covariates. The model fit the data well (Table  5) culating the average z-score; the same approach yielded a [χ (59) = 82.79, p < 0.02, CFI = 0.96, RMSEA = 0.04, teacher-rated social functioning composite. Higher scores SRMR = 0.07] where inattention positively predicted Wave 3 reflected worse social functioning. early adolescent depression (β = 0.33, SE = 0.09, p < 0.001), but EF did not (β = 0.06, SE = 0.12, p = 0.60). In addition Prediction of Early Adolescent Depression to demonstrating good fit to the data, Model 3 accounted for 9.0% of the variance in the Wave 3 latent depression To review, controlling for baseline depression, SES, and variable. pubertal status, we tested childhood inattention, hyper- Model 4. Hyperactivity-impulsivity and EF were also activity-impulsivity, and EF as independent predictors of tested as predictors of depression. Although the model early adolescent depression; we also controlled for baseline showed good fit (Table  5) [χ (59) = 72.43 p < 0.11, age on the EF factor (depression outcomes were adjusted C F I = 0 . 9 8 , R M SE A = 0 . 0 3 , S RM R = 0 . 0 7 ], it o n ly for age/grade and sex). When inattention significantly pre- accounted for 2.0% of the variance in Wave 3 depression. dicted depression, we also controlled for Wave 1 anxiety on Additionally, neither hyperactivity-impulsivity (β = 0.10, inattention. SE = 0.11, p = 0.36) nor EF (β = -0.04, SE = 0.13 p = .77) Model 1. We regressed a latent depression variable from predicted depression, suggesting overall that inattention is Wave 3 on Wave 1 inattention, hyperactivity-impulsivity, the primary risk factor for later depression. and EF latent variables. Key indices suggested model Model 5. We reproduced Model 3 (i.e., inattention and misspecification [χ (93) = 221.02, p < 0.001, CFI = 0.90, EF predicting depression) but conservatively added base- RMSEA = 0.08, SRMR = 0.08; Table 5]. Additionally, the line anxiety (i.e., CBCL Anxiety Problems). Even with latent variable covariance matrix was not positive definite, control of anxiety on the latent inattention variable, the reflecting the high correlation between the inattention and model showed good fit (Table  5) [χ (72) = 99.09, p < 0.01, hyperactivity-impulsivity latent variables. Wave 1 inat- CFI = 0.96, RMSEA = 0.04, SRMR = 0.08] and accounted tention positively predicted Wave 3 depression (β = 0.49, for 10.8% of the variance in Wave 3 depression. Base- SE = 0.15, p = 0.001), but hyper activity-im pulsivity line inattention continued to predict depression (β = 0.35, (β = -0.24, SE = 0.17, p = 0.14) and EF (β = 0.06, SE = 0.12, SE = 0.09, p < 0.001; Fig. 2) whereas EF did not (β = 0.10, p = 0.60) did not. To further examine the initial aim of testing SE = 0.12, p = 0.40). 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 761 DISC DBD-P DBD-T Inattention Inattention Inattention Parent Parent Wave 1 CDITotal Wave 3 *** .92 (.03) *** Income Education T-Score Puberty .89 (.03) *** .43 (.07) *** CBCL Anxiety .36 (.06) T-Score Wave 1 Inattention .03 (.10) -.04 (.11)-.01 (.12) -.01 (.09) *** .35 (.09) CDITotal T-Score *** .66 (.09) *** Wave 3 RCADS .98 (.10) Depression *** Depression -.56 (.08) T-Score *** .76 (.11) YSR Affective T-Score .10 (.12) Wave 1 Executive *** .63 (.05) Functioning Wave 1 Age *** .51 (.06) *** .73 (.05) *** .58 (.06) TMT Stroop WISC-IVDS B CW Backward Fig. 2 Represents Model 5 with standardized estimates. Standard- with additional control of anxiety on inattention. The model accounts *** ized errors are in parentheses. Solid lines indicate significant relation- for 10.8% of the variance in Wave 3 depression. p ≤ 0.001. ** * ships among variables. Inattention positively predicts early adolescent p ≤ 0.01. p ≤ 0.05 depression with control of SES, baseline depression, and puberty Multi-group Analysis. We conducted multi-group analy- A model consisting of Wave 2 academic function- ing and social problems as mediators of predictions from ses based on Wave 1 ADHD diagnostic status to determine whether predictions of youth depression from baseline EF Wave 1 latent inattention, hyperactivity-impulsivity, and EF showed poor fit [χ (194) = 255.03 p < 0.01, CFI = 0.93, differed according to group. However, this model had poor fit (SRMR = 0.22), suggesting that EF predicted depression RMSEA = 0.04, SRMR = 0.09]. Examination of model pathways revealed no mediated effects. Wave 1 inattention similarly among youth with and without ADHD, though concerns related to power for multigroup models (Kline, inversely predicted Wave 2 academic functioning (β = -0.83, SE = 0.34, p = 0.01) and positively predicted Wave 2 par- 2015) limit conclusions. ent-rated social problems (β = 0.35, SE = 0.17 p = 0.03). However, neither Wave 2 academic achievement (β = 0.20, Early Adolescent Depression: Mediation by Academic Functioning and Social Problems SE = 0.23, p = 0.37) nor Wave 2 parent-rated social problems (β = 0.05, SE = 0.13, p = 0.71) predicted Wave 3 depression. Using SEM, we tested academic functioning and social Baseline hyperactivity-impulsivity and EF did not predict any mediators. Finally, none of the mediators (i.e., academic problems as temporally-ordered mediators of predictions of Wave 3 youth self-reported depression from baseline inatten- functioning, parent-rated social problems, teacher-rated social problems) predicted Wave 3 depression. There was tion, hyperactivity-impulsivity, and EF. Specifically, a Wave 2 latent academic functioning variable and two composite a significant direct effect from Wave 1 inattention to Wave 3 depression (β = 0.68, SE = 0.30, p = 0.02), but no signifi - social problems variables (i.e., parent-report, teacher-report) were entered as mediators, controlling for the same covari- cant direct effects from hyperactivity-impulsivity (β = -0.44, SE = 0.24, p = 0.06) or EF (β = 0.00, SE = 0.15, p = 0.37) to ates previously described. 1 3 762 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Table 1 Descriptive Statistics Variable M (SD) or % of Sample Range n of Demographics and Key Study Variables Wave 1 Age 7.39 (1.07) 6-9 216 Wave 2 Age 9.68 (1.27) 7-13 193 Wave 3 Age 12.07 (1.30) 9-15 172 Sex (% Male) 66.67 - 216 Race-Ethnicity (% Caucasian) 50.93 - 216 SES Income (% $75,001 or more) 65.99 - 197 Parent Education 4.92 (0.91) 1.5-6 202 Wave 1 Depression (CDI T-Score) 46.84 (7.47) 35-71 144 Wave 3 Puberty 2.27 (0.73) 1-3.8 143 Wave 1 Anxiety (CBCL Anxiety Problems T-Score) 56.21 (7.49) (50-75) 214 FSIQ 107.29 (14.24) 73-144 216 Wave 1 Inattention Symptoms Inattention Symptoms (DSIC-IV) 4.54 (3.16) 0-9 216 Inattention Symptoms (DBD Parent) 11.17 (7.63) 0-27 210 Inattention Symptoms (DBD Teacher) 9.35 (8.33) 0-27 150 Wave 1 Hyperactivity-Impulsivity Symptoms Hyperactivity-Impulsivity Symptoms (DSIC-IV) 3.34 (3.08) 0-9 216 Hyperactivity-Impulsivity Symptoms (DBD Parent) 9.32 (7.33) 0-27 209 Hyperactivity-Impulsivity Symptoms (DBD Teacher) 7.79 (8.38) 0-27 150 Wave 1 Executive Functioning TMT-B (min) -1.22 (0.81) -5.02 - -0.27 212 WISC-IV DSB Raw 5.86 (1.58) 0-10 210 Stroop C-W 21.42 (6.43) 4-41 201 Wave 2 Academic Functioning WIAT-II Word Reading Standard Score 107.58 (14.27) 53-141 184 WIAT-II Math Reasoning Standard Score 111.86 (16.82) 61-160 183 School Competence T-Score (CBCL) 44.79 (9.46) 24-55 182 Academic Performance T-Score (TRF) 49.27 (10.12) 35-65 95 Wave 2 Social Problems Negative Social Preference (Parent Dishion) -2.60 (1.85) -4-4 172 Negative Social Preference (Teacher Dishion) -2.18 (2.13) -4-4 92 Social Problems Raw (CBCL) 3.27 (3.48) 0-16 188 Social Problems Raw (TRF) 2.28 (2.67) 0-14 91 Wave 3 Depression CDI T-Score 44.14 (7.74) 34-75 149 RCADS T-Score 43.91 (10.55) 30-78 148 Affective Problems T-Score (YSR) 54.66 (7.49) 50-80 61 SES  =  Socioeconomic Status, CDI  =  Children's Depression Inventory, CBCL  =  Child Behavior Check- list, DISC-IV = Diagnostic Interview Schedule for Children Fourth Edition, DBD  =  Disruptive Behav- ior Disorder Rating Scale, TMT-B  =  Trail Making Test Part B, WISC-IV  =  Wechsler Intelligence Scale for Children-Fourth Edition, DSB  =  Digit Span Backwards, Stroop C-W = Stroop Color-Word Condi- tion, WIAT-II  =  Wechsler Individual Achievement Scale Second Edition, TRF  =  Teacher Report Form, RCADS = Revised Children's Anxiety and Depression Scale, YSR = Youth Self-Report Wave 3 depression emerged. Because no model pathways Discussion suggested significant mediation, respecified models were not investigated. We also tested a multigroup model with ADHD Identifying childhood predictors and mediators of depres- diagnostic status as a moderator of mediation by academic sion, prior to the acute vulnerability in adolescence, will and social functioning, but the model’s failure to converge accelerate innovations in prevention and intervention. prevented strong inferences. In a study of 216 youth ages 6–9, baseline inattention, 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 763 Table 2 Auxiliary Variables Demographic Wave 1 Wave 2 Wave 3 Child Age (Waves WIAT-II Word Reading SS DISC-IV Inattention Symptoms CDI Total Raw Score 2 and 3) Child Sex WIAT-II Math Reasoning SS DISC-IV HI Symptoms RCADS Depression Raw Score FSIQ CBCL School Competence T-score DBD-Parent Inattention Symptoms YSR Affective Problems Raw Score CBCL Social Problems T-score DBD-Parent HI Symptoms TRF Academic Performance T-score DBD-Teacher Inattention Symptoms TRF Social Problems T-score DBD-Teacher HI Symptoms TMT-B Stroop C-W Condition WISC-IV Digit Span Backwards RCADS Depression T-score FSIQ  =  Full Scale IQ, WIAT-II   =  Wechsler Individual Achievement Scale Second Edition; SS  =  Standard Score, CBCL  =  Child Behavior Checklist, TRF = Teacher Report Form, DISC-IV = Diagnostic Interview Schedule for Children Fourth Edition, HI = Hyperactivity-Impulsivity, DBD = Disruptive Behavior Disorder Rating Scale; TMT-B = Trail Making Test Part B, Stroop C-W =  Stroop Color-Word Condition, WISC- IV = Wechsler Intelligence Scale for Children-Fourth Edition, RCADS = Revised Children's Anxiety and Depression Scale, CDI  =  Children's Depression Inventory, YSR = Youth Self-Report hyperactivity-impulsivity, and EF were tested as simulta- competency-based model (Cole, 1991), self-esteem may be neous predictors of youth self-reported depression approx- primary given its proximity to depression. Therefore, inter- imately four years later. Controlling for SES, baseline pretations or appraisals of academic and social functioning depression, and puberty as well as age (on the EF latent may be more salient. For example, negative self-schema acti- factor), inattention positively predicted early adolescent vated by stressors and/or low mood potentiates depression, depression, even with added control of baseline anxiety in including through maladaptive attributional styles (Jacobs a model with inattention and EF as latent predictors. In et al., 2008). Given the centrality of multifinality to ADHD, contrast, academic functioning and social problems did not across multiple settings and functional contexts, we await mediate predictions of depression from baseline inattention, strong tests of cognitive factors (e.g., attributional style) and hyperactivity-impulsivity, and EF. competency-based constructs, as mediators of depression. The unique, robust association of early ADHD (inatten- Further, because adolescence confers heighted sensitivity to tion in particular) with early adolescent depression observed social relationships, academic and social functioning meas- in this study is well-aligned with previous evidence, includ- ured at an older age may mediate emerging depression later in ing predictions through young adulthood (Chronis-Tuscano development from childhood ADHD (Powell et al., 2020). In et al., 2010; Meinzer et al., 2016). The centrality of inatten- the current study, social measures were not derived from self- tion to depression was also reported previously, across mul- report; instead, social functioning was estimated from broader tiple informants, in a cross-sectional study (Fenesy & Lee, constructs (i.e., social acceptance, social problems) according 2019). Although inattention is transdiagnostic in nature, to informant report that may lack the precision necessary to inattention assessed via measures of ADHD across inform- capture successful peer interactions. For example, a multi- ants (i.e., parent, teacher) prospectively predicted depression ple mediation model testing whether individual social skills independent of co-occurring EF, anxiety, and baseline depres- (e.g., cooperation, empathy) collectively and uniquely medi- sion. Whereas youth with ADHD struggle with selective and ate predictions of depression from inattention, hyperactivity- sustained attention, anxiety is characterized by attentional impulsivity, and EF could identify particular aspects of social biases secondary to threat (Weissman et al., 2012). Further, behavior to target for intervention and reduce depression in directed attention is generally a top-down cognitive process, youth. Social skills training improves social functioning and but responses to threat are primarily mediated subcortically reduces the risk of depression in youth with autism spectrum (Nigg, 2017). Thus, inattention secondary to ADHD is evident disorder (Hotton & Coles, 2016), suggesting that continued across settings and contexts and contributes to impairments research examining social functioning as an underlying mech- across multiple domains, independently conferring risk for anism of depression in youth with neurodevelopmental disor- depression later in development. The hypothesis that symp- ders is worthwhile and will facilitate innovations in interven- toms of inattention and hyperactivity-impulsivity as well tion development. Overall, appropriate interventions targeting as EF deficits would predict multiple academic and social either a behavioral deficit or maladaptive cognitions could “failures” was not supported in this study. According to a subsequently be applied to mitigate the risk of depression. 1 3 764 Research on Child and Adolescent Psychopathology (2022) 50:753–770 1 3 Table 3 Correlations Among Covariates, Predictor, Mediator, and Outcome Variables Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Wave 1 Age - 2. Sex (% Male) -0.09 - 3. Income (1 = $75,001 or more) -0.05 0.12 - * *** 4. Parent Education -0.04 0.14 0.46 - * *** *** 5. Wave 1 Depression (CDI T-Score) 0.11 -0.17 -0.27 -0.32 - *** *** 6. Wave 3 Puberty 0.49 -0.35 -0.07 -0.02 0.17 - 7. Wave 1 Anxiety (CBCL Anxiety Problems T-Score) 0.04 0.07 -0.03 -0.03 0.06 0.06 - * * *** * 8. FSIQ -0.04 0.13 0.15 0.35 -0.17 -0.07 -0.06 - ** *** *** 9. Wave 1 Inattention Symptoms (DSIC-IV) -0.03 0.11 -0.09 -0.08 0.24 0.01 0.34 -0.24 - * *** *** *** 10. Wave 1 Inattention Symptoms (DBD P) -0.05 0.09 -0.06 -0.09 0.17 0.03 0.31 -0.26 0.82 - *** *** *** 11. Wave 1 Inattention Symptoms (DBD T) -0.10 0.15 -0.11 -0.04 0.14 0.06 0.08 -0.28 0.40 0.41 - ** ** *** *** *** *** 12. Wave 1 Hyperactivity-Impulsivity Symptoms (DSIC-IV) -0.19 0.19 0.10 -0.04 0.10 0.02 0.32 -0.03 0.58 0.60 0.28 - * ** *** *** *** *** *** 13. Wave 1 Hyperactivity-Impulsivity Symptoms (DBD P) -0.16 0.16 0.10 -.04 0.09 0.05 0.34 -0.08 0.61 0.75 0.34 0.87 - *** ** ** *** *** *** *** 14. Wave 1 Hyperactivity-Impulsivity Symptoms (DBD T) -0.31 0.21 0.10 -0.03 0.11 0.14 0.14 -0.14 0.21 0.28 0.64 0.49 0.49 *** * *** *** *** ** ** 15. Wave 1 TMT-B (min) 0.46 0.00 -0.02 0.05 -0.01 0.18 0.06 0.38 -0.27 -0.28 -0.15 -0.19 -0.18 *** *** * * 16. Wave 1 WISC-IV DSB Raw 0.29 0.02 0.08 0.10 -0.11 0.09 0.02 0.38 -0.12 -0.15 -0.12 -0.14 -0.09 *** ** *** *** *** ** * * 17. Wave 1 Stroop C-W 0.38 -0.04 0.09 0.06 -0.06 0.22 -0.04 0.31 -0.23 -0.27 -0.23 -0.15 -0.16 * ** *** *** *** *** 18. Wave 2 WIAT-II Word Reading Standard Score -0.08 0.05 0.14 0.16 -0.21 -0.08 0.04 0.57 -0.24 -0.24 -0.31 -0.06 -0.11 ** *** *** *** *** *** * ** 19. Wave 2 WIAT-II Math Reasoning Standard Score -0.10 0 .11 0.21 0.39 -0.16 -0.06 -0.07 0.76 -0.31 -0.36 -0.31 -0.14 -0.18 *** *** ** ** *** *** *** *** *** *** 20. Wave 2 School Competence T-Score (CBCL) -0.02 0.07 0.27 0.28 -0.21 0.01 -0.21 0.59 -0.48 -0.49 -0.38 -0.24 0.33 ** ** * *** ** ** *** 21. Wave 2 Academic Performance T-Score (TRF) 0.08 0.05 0.28 0.26 -0.26 0.16 -0.02 0.55 -0.31 -0.32 -0.39 -0.04 -0.07 *** * *** *** *** *** *** 22. Wave 2 Negative Social Preference (Parent Dishion) 0.07 0.01 -0.00 -0.07 0.10 0.08 0.26 -0.16 0.35 0.38 0.28 0.32 0.37 ** * * * * 23. Wave 2 Negative Social Preference (Teacher Dishion) 0.08 0.13 -0.13 -0.08 0.12 0.04 0.31 -0.16 0.22 0.20 0.28 0.23 0.23 * *** *** *** *** * *** *** 24. Wave 2 Social Problems Raw (CBCL) -0.01 0.01 -0.02 -0.16 0.10 0.03 0.39 -0.24 0.41 0.46 0.23 0.46 0.49 * * ** *** ** 25. Wave 2 Social Problems Raw (TRF) 0.02 0.13 -0.03 -0.03 0.01 0.00 0.19 -0.13 0.22 0.25 0.07 0.26 0.25 * ** *** *** * * * 26. Wave 3 CDI T-Score 0.17 -0.23 -0.13 -0.15 0.09 0.28 0.27 -0.17 0.11 0.17 0.22 -0.02 -0.00 *** * ** *** 27. Wave 3 RCADS T-Score 0.06 0.03 -0.02 -0.06 0.00 0.03 0.32 -0.17 0.24 0.28 0.12 0.10 0.14 28. Wave 3 Affective Problems T-Score (YSR) 0.04 0.07 0.02 0.09 0.23 0.17 0.14 0.13 0.07 0.14 0.05 0.13 0.11 Variable 14 15 16 17 18 19 20 21 22 23 24 25 26 27 14. Wave 1 Hyperactivity-Impulsivity Symptoms (DBDT) - 15. Wave 1 TMT-B (min) -0.14 - *** 16. Wave 1 WISC-IV DSB Raw -0.13 0.45 - *** *** 17. Wave 1 Stroop C-W -0.11 0.34 0.26 - * *** *** ** 18. Wave 2 WIAT-II Word Reading Standard Score -0.17 0.25 0.36 0.20 - *** *** *** *** 19. Wave 2 WIAT-II Math Reasoning Standard Score -0.11 0.39 0.39 0.26 0.61 - Research on Child and Adolescent Psychopathology (2022) 50:753–770 765 Critically, screening for childhood inattention may iden- tify youth vulnerable to later onset depression. These indi- viduals may benefit from prevention efforts (e.g., learning CBT skills to enhance mood) to mitigate depression onset and associated negative outcomes, in line with experimen- tal evidence (Brent et al., 2015). Consistent with pediat- ric treatment guidelines, our findings emphasize the need for ongoing care and monitoring of children with ADHD, especially as inattentive symptoms are likely to persist into adolescence and adulthood (Wolraich et al., 2019) and may increase the likelihood of co-occurring depression. Effective intervention for childhood ADHD requires treatment by a multidisciplinary team to properly assess and intervene prior to depression onset (e.g., medical providers, psychologists, educators; Barbaresi, 2020). Future studies should test the associations observed in the present study in youth diag- nosed with major depressive disorder or those at elevated risk (e.g., offspring of depressed mothers) as well. Parental depression, for example, is a robust risk factor for adolescent depression. If inattention symptoms predicted clinically sig- nificant depression in other populations, this would substan- tiate the rationale to screen inattention even in the absence of ADHD. Additionally, greater refinement of specific inatten- tion symptoms or combinations of symptoms that increase risk for later depression would enable clinicians to improve screening techniques. Surprisingly, with control of baseline ADHD, childhood EF did not predict early adolescent depression. Consist- ent with evidence that EF impairments remit following a depressive episode in adults and adolescents (Biringer et al., 2005; Maalouf et al., 2011), present findings suggest that childhood EF did not uniquely confer vulnerability for early adolescent depression. However, developmentally-sensitive socio-emotional and neurobiological changes critically con- textualize these findings. For example, most of the current participants had only begun their transition to adolescence, thus still undergoing well-characterized developmental unfolding of limbic systems and prefrontal regions under- lying emotion dysregulation and heightened reward sensi- tivity (Powers & Casey, 2015). Further, poor EF may be more acutely related to adolescent-onset depression because frontal networks supporting cognitive down-regulation of negative emotionality are still emerging. EF measures con- tinue to suffer from poor ecological validity whereas rating scales may prove more useful (Barkley & Murphy, 2011). Similarly, computerized EF measures, which include more accurate assessments of subtle variations in response time (e.g., NIH ToolBox; Zelazo et al., 2014), may show better predictive properties. Finally, we used a latent EF variable, derived from multiple EF domains, to reduce measurement error. However, differentiated measures of EF may yield more specific patterns of association. For example, better inhibition in adults supported reappraisal of negative stimuli 1 3 Table 3 (continued) Variable 14 15 16 17 18 19 20 21 22 23 24 25 26 27 ** *** *** *** *** 20. Wave 2 School Competence T-Score (CBCL) -0.22 0.33 0.25 0.26 0.56*** 0.60 - ** *** ** *** *** *** 21. Wave 2 Academic Performance T-Score (TRF) -0.07 0.30 0.38 0.28 0.50 0.57 0.68 - ** ** * *** 22. Wave 2 Negative Social Preference (Parent Dishion) 0.25 -0.10 -0.03 -0.19 -0.12 -0.15 -0.38 -0.18 - * ** *** ** *** 23. Wave 2 Negative Social Preference (Teacher Dishion) 0.26 -0.01 0.02 0.04 -0.04 -0.25 -0.34 -0.28 0.54*** - ** ** ** * *** *** * *** *** 24. Wave 2 Social Problems Raw (CBCL) 0.23 -0.18 -0.13 -0.18 -0.17 -0.31 -0.45 -0.21 0.67 0.37 - * ** *** ** *** *** *** 25. Wave 2 Social Problems Raw (TRF) 0.21 -0.23 0.03 -0.04 -0.12 -0.25 -0.33 -0.25 0.45 0.71 0.38 - ** * * 26. Wave 3 CDI T-Score -0.01 0.01 -0.00 0.05 -0.10 -0.15 -0.21 -0.15 0.18 0.16 0.18 0.06 - ** *** 27. Wave 3 RCADS T-Score 0.00 -0.10 -0.02 -0.09 -0.03 -0.12 -0.13 -0.07 0.23 0.20 0.07 0.16 0.64 - *** *** 28. Wave 3 Affective Problems T-Score (YSR) 0.06 0.13 0.22 0.17 0.21 0.11 -0.02 0.13 0.06 0.12 -0.09 -0.07 0.57 0.76 CDI  =  Children's Depression Inventory, CBCL  =  Child Behavior Checklist, DISC-IV  =  Diagnostic Interview Schedule for Children Fourth Edition, DBD P  =  Disruptive Behavior Disorder Rating Scale Parent Report, DBD T = Disruptive Behavior Disorder Rating Scale Teacher Report, TMT-B = Trail Making Test Part B, WISC-IV = Wechsler Intelligence Scale for Children- Fourth Edition, DSB = Digit Span Backwards; Stroop C-W = Stroop Color-Word Condition, WIAT-II = Wechsler Individual Achievement Scale Second Edition, TRF = Teacher Report Form; RCADS = Revised Children's Anxiety and Depression Scale, YSR = Youth Self-Report ***p ≤ 0.001; **p ≤ 0.01; *p ≤ 0.05 766 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Table 4 Factor Loadings for Factor Factor Loading SE z p Confirmatory Factor Analyses Wave 1 Inattention* Inattention Symptoms (DSIC-IV) 0.85 0.06 13.29 <0.001 Inattention Symptoms (DBD Parent) 0.96 0.06 15.40 <0.001 Inattention Symptoms (DBD Teacher) 0.44 0.07 6.07 <0.001 Wave 1 Hyperactivity-Impulsivity* Hyperactivity-Impulsivity Symptoms (DSIC-IV) 0.93 0.04 21.92 <0.001 Hyperactivity-Impulsivity Symptoms (DBD Parent) 0.93 0.04 23.78 <0.001 Hyperactivity-Impulsivity Symptoms (DBD Teacher) 0.53 0.07 14.95 <0.001 Wave 1 Executive Functioning* TMT-B (min) 0.76 0.08 9.39 <0.001 WISC-IV DSB Raw 0.58 0.07 7.36 <0.001 Stroop C-W 0.48 0.07 6.48 <0.001 Wave 2 Academic Functioning* WIAT-II Word Reading Standard Score 0.73 0.05 16.06 <0.001 WIAT-II Math Reasoning Standard Score 0.77 0.05 16.21 <0.001 School Competence T-Score (CBCL) 0.77 0.05 15.66 <0.001 Academic Performance T-Score (TRF) 0.81 0.06 14.61 <0.001 Wave 2 Social Problems Negative Social Preference (Parent Dishion) 0.85 0.07 12.26 <0.001 Negative Social Preference (Teacher Dishion) 0.67 0.12 5.67 <0.001 Social Problems Raw (CBCL) 0.74 0.06 11.82 <0.001 Social Problems Raw (TRF) 0.64 0.11 5.62 <0.001 Wave 3 Depression* CDI T-Score 0.70 0.05 13.39 <0.001 RCADS T-Score 0.92 0.06 14.90 <0.001 Affective Problems T-Score (YSR) 0.79 0.07 10.68 <0.001 DISC-IV = Diagnostic Interview Schedule for Children Fourth Edition; DBD = Disruptive Behavior Dis- order Rating Scale, TMT-B = Trail Making Test Part B, WISC-IV = Wechsler Intelligence Scale for Chil- dren-Fourth Edition, DSB  =  Digit Span Backwards, Stroop C-W  =  Stroop Color-Word Condition, WIAT- II  =  Wechsler Individual Achievement Scale Second Edition, CBCL  =  Child Behavior Checklist, TRF =    Teacher Report Form, CDI  =  Children's Depression Inventory, RCADS  =  Revised Children's Anxiety and Depression Scale, YSR = Youth Self-Report *Indicates Latent Factors Used in Analyses and effective emotion suppression that mitigated depressive limitations include a sample recruited for ADHD at baseline, symptoms (Joormann & Gotlib, 2010). but one not specifically designed to capture youth depression. The present study had several key strengths and limitations. Thus, few participants demonstrated clinically significant First, the SEM approach reduced measurement error and Type depression in early adolescence. At Wave 3, only one par- I error, which is particularly important for EF given its mul- ticipant had completed pubertal development, a key limiting tidimensionality. We also utilized temporally-ordered, multi- factor given the dramatic increase in psychopathology sec- method/informant data to test causal mediation. Important ondary to pubertal timing. These aspects of the study design, Table 5 Predictive Model Fit Model χ2 CFI RMSEA SRMR Indices Model 1 (Inattention, Hyperactivity-Impulsivity, EF) χ (93) = 221.02, p< 0.001 0.90 0.08 0.08 Model 2 (ADHD, EF) χ (59) = 78.70, p= 0.04 0.97 0.04 0.07 Model 3 (Inattention, EF) χ (59) = 82.79, p< 0.02 0.96 0.04 0.07 Model 4 (Hyperactivity-Impulsivity, EF) χ2(59) = 72.43 p< 0.11 0.98 0.03 0.07 Model 5 (Inattention with control of anxiety, EF) χ (72) = 99.09, p< 0.01 0.96 0.04 0.08 Latent predictor variables in parentheses. Bold text indicates good model fit based on the index 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 767 Informed Consent Prior to each wave of data collection, written con- in combination with the prospective longitudinal framework, sent was collected from parents/guardians and participating youth pro- likely contributed to the somewhat low variance accounted for vided written assent. the Wave 3 latent depression variable. Variance accounted for would likely increase if replicated in samples studying post- Conflicts of Interest The authors declare that they have no conflicts pubertal adolescents and if other risk factors highly predic- of interest. tive of youth depression (e.g., maternal depression; Hammen & Brennan, 2003) were included. Nevertheless, the fact that Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- childhood inattention significantly predicted early adolescent tion, distribution and reproduction in any medium or format, as long depression approximately four years later remains an important as you give appropriate credit to the original author(s) and the source, finding and may be particularly critical to study in a sample of provide a link to the Creative Commons licence, and indicate if changes youth who all meet diagnostic criteria for ADHD. With respect were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated to multigroup models, the present study may have been under- otherwise in a credit line to the material. If material is not included in powered to detect ee ff cts given 216 participants, missing data, the article's Creative Commons licence and your intended use is not and inclusion of several parameters (Kline, 2015); therefore, permitted by statutory regulation or exceeds the permitted use, you will results related to the multigroup model must be interpreted need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . with caution. Multigroup models examining whether the pro- spective prediction of depression from childhood EF varies according to ADHD diagnostic status should be evaluated in References larger samples to improve confidence in our results. Addi- tionally, consideration of sex and age moderators would be Achenbach, T. M., & Rescorla, L. (2001). ASBEA School-age Forms compelling, especially as the risk for depression increases dur- & Profiles . University of Vermont, Research Center for Children, ing adolescence and the prevalence of depression is higher in Youth, and Families. Arán Filippetti, V., & Richaud, M. C. (2017). A structural equation adolescent girls relative to boys (Thapar et al., 2012). Finally, modeling of executive functions, IQ and mathematical skills in a small number of participants with a diagnosis of ADHD primary students: Differential effects on number production, men- were on stimulant medication on the day of baseline testing tal calculus and arithmetical problems. Child Neuropsychology, (n = 13). Simulant medication improves performance on some 23(7), 864–888. https://doi. or g/10. 1080/ 09297 049. 2016. 11996 65 Avenevoli, S., Swendsen, J., He, J.-P., Burstein, M., & Merikangas, K. EF dimensions in children with ADHD (Barnett et al., 2001; R. (2015). Major Depression in the National Comorbidity Survey- Snyder et al., 2008). Medication use in ADHD is likely most Adolescent Supplement: Prevalence, Correlates, and Treatment. reflective of disorder severity due to the intervention selection Journal of the American Academy of Child & Adolescent Psychia- bias, and because the current study did not use random assign- try, 54(1), 37-44.e32. https:// doi. org/ 10. 1016/j. jaac. 2014. 10. 010 Barbaresi, W. J., Campbell, L., Diekroger, E. A., Froehlich, T. E., Liu, ment we are not positioned to make inferences about treatment Y. H., O’Malley, E., & Chan, E. (2020). Society for developmen- effects (Larzelere et al., 2004). tal and behavioral pediatrics clinical practice guideline for the We tested the prospective association of childhood ADHD assessment and treatment of children and adolescents with com- plex attention-deficit/hyperactivity disorder. Journal of Develop - and EF with depression in early adolescence, including medi- mental & Behavioral Pediatrics, 41, S35–S57. https://doi. or g/10. ation by academic functioning and social problems. Latent 1097/ DBP. 00000 00000 000770 inattention positively predicted youth self-reported depression, Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and but hyperactivity-impulsivity and EF did not. These findings executive functions: Constructing a unifying theory of ADHD. Psychological Bulletin, 121(1), 65–94. https:// doi. org/ 10. 1037/ support the centrality of inattention as a unique risk factor 0033- 2909. 121.1. 65 for depression, demonstrating the importance of screening Barkley, R. A. (2012). Executive functions: What they are, how they for inattention and interventions to improve inattention and work, and why they evolved. Guilford Press. reduce risk. Continued investigation of underlying mecha- Barkley, R. A., & Murphy, K. R. (2011). The Nature of Executive Function (EF) Deficits in Daily Life Activities in Adults with nisms between inattention and early adolescent depression ADHD and Their Relationship to Performance on EF Tests. will provide additional targets for intervention. Journal of Psychopathology and Behavioral Assessment, 33(2), 137–158. https:// doi. org/ 10. 1007/ s10862- 011- 9217-x Funding This work was supported by the National Institutes of Mental Barnett, R., Maruff, P., Vance, A., Luk, E. S., Costin, J., Wood, C., & Health predoctoral fellowship in Biobehavioral Issues in Mental and Pantelis, C. (2001). Abnormal executive function in attention defi- Physical Health (MH 15750 to Michelle C. Fenesy) and the National cit hyperactivity disorder: The effect of stimulant medication and Institutes of Health (R03AA020186-01 to Steve S. Lee). age on spatial working memory. Psychological Medicine, 31(6), 1107–1115. https:// doi. org/ 10. 1017/ S0033 29170 10041 72 Best, J. R., Miller, P. H., & Naglieri, J. A. (2011). Relations between execu- Compliance with Ethical Standards tive function and academic achievement from ages 5 to 17 in a large, representative national sample. Learning and Individual Differences, Ethical Approval Investigators obtained Institutional Review Board 21(4), 327–336. https:// doi. org/ 10. 1016/j. lindif. 2011. 01. 007 (IRB) approval from the University of California, Los Angeles. 1 3 768 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Biringer, E., Lundervold, A., Stordal, K., Mykletun, A., Egeland, J., Fergusson, D. M., & Horwood, L. J. (1995). Predictive Validity of Bottlender, R., & Lund, A. (2005). Executive function improve- Categorically and Dimensionally Scored Measures of Disruptive ment upon remission of recurrent unipolar depression. European Childhood Behaviors. Journal of the American Academy of Child Archives of Psychiatry and Clinical Neuroscience, 255(6), 373– & Adolescent Psychiatry, 34(4), 477–487. https://do i.or g/10. 10 97/ 380. https:// doi. org/ 10. 1007/ s00406- 005- 0577-700004 583- 19950 4000- 00015 Birmaher, B., & Brent, D. (2007). Practice Parameter for the Assess- Gau, S. S. F., & Shang, C. Y. (2010). Executive functions as endophe- ment and Treatment of Children and Adolescents With Depressive notypes in ADHD: Evidence from the Cambridge Neuropsycho- Disorders. Journal of the American Academy of Child & Adoles- logical Test Battery (CANTAB). Journal of Child Psychology and cent Psychiatry, 46(11), 1503–1526. https:// doi. org/ 10. 1097/ chi. Psychiatry, 51(7), 838–849. https://doi. or g/10. 1111/j. 1469- 7610. 0b013 e3181 45ae1c2010. 02215.x Brent, D. A., Brunwasser, S. M., Hollon, S. D., Weersing, V. R., Clarke, Golden, C. J., Freshwater, S. M., & Zarabeth, G. (2003). Stroop Color G. N., Dickerson, J. F., & Lynch, F. L. (2015). Effect of a cogni- and Word Test Children's Version for ages 5-14: A Manual for tive-behavioral prevention program on depression 6 years after Clinical and Experimental Uses. Stoelting. implementation among at-risk adolescents: A randomized clini- Greenberg, P. E., Fournier, A.-A., Sisitsky, T., Pike, C. T., & Kessler, cal trial. JAMA Psychiatry, 72(11), 1110–1118. https://doi. or g/10. R. C. (2015). The economic burden of adults with major depres- 1001/ jamap sychi atry. 2015. 1559 sive disorder in the United States (2005 and 2010). The Journal Brown, T. A. (2014). Confirmatory factor analysis for applied of Clinical Psychiatry, 76(2), 155–162. https:// doi. org/ 10. 4088/ research: Guilford Publications.jcp. 14m09 298 Carskadon, M. A., & Acebo, C. (1993). A self-administered rating Gyurak, A., Goodkind, M. S., Kramer, J. H., Miller, B. L., & Levenson, scale for pubertal development. Journal of Adolescent Health, R. W. (2012). Executive functions and the down-regulation and 14(3), 190–195. https:// doi. org/ 10. 1016/ 1054- 139x(93) 90004-9 up-regulation of emotion. Cognition & Emotion, 26(1), 103–118. Castellanos, F. X., & Tannock, R. (2002). Neuroscience of attention-https:// doi. org/ 10. 1080/ 02699 931. 2011. 557291 deficit/hyperactivity disorder: The search for endophenotypes. Hammen, C., & Brennan, P. A. (2003). Severity, chronicity, and timing Nature Reviews Neuroscience, 3, 617–628. https:// doi. org/ 10. of maternal depression and risk for adolescent offspring diagnoses 1038/ nrn896 in a community sample. Archives of General Psychiatry, 60(3), Center for Disease Control and Prevention (2016). National Vital Sta- 253–258. https:// doi. org/ 10. 1001/ archp syc. 60.3. 253 tistics Reports. Deaths: Leading Causes for 2014. Retrieved Harter, S., & Marold, D. R. (1994). The direction of the link between Chorpita, B. F., Yim, L., Moffitt, C., Umemoto, L. A., & Francis, S. E. self-esteem and affect: Beyond causal modeling. Disorders and (2000). Assessment of symptoms of DSM-IV anxiety and depres- dysfunctions of the self. In D. Cicchetti & S. L. Toth (Eds.), Roch- sion in children: A revised child anxiety and depression scale. ester symposium on developmental psychopathology (Vol. 5, pp. Behaviour Research and Therapy, 38(8), 835–855. https://doi. or g/ 333–370). New York: University of Rochester Press. 10. 1016/ S0005- 7967(99) 00130-8 Hotton, M., & Coles, S. (2016). The effectiveness of social skills train- Chronis-Tuscano, A., Molina, B. G., Pelham, W. E., et al. (2010). Very ing groups for individuals with autism spectrum disorder. Review early predictors of adolescent depression and suicide attempts in Journal of Autism and Developmental Disorders, 3(1), 68–81. children with attention-deficit/hyperactivity disorder. Archives of https:// doi. org/ 10. 1007/ s40489- 015- 0066-5 General Psychiatry, 67(10), 1044–1051. https://doi. or g/10. 1001/ Hu, L., & t., & Bentler, P. M. . (1999). Cutoff criteria for fit indexes in archg enpsy chiat ry. 2010. 127. covariance structure analysis: Conventional criteria versus new Cole, D. A. (1991). Preliminary support for a competency-based model alternatives. Structural Equation Modeling: A Multidisciplinary of depression in children. Journal of Abnormal Psychology, Journal, 6(1), 1–55. https://doi. or g/10. 1080/ 10705 51990 95401 18 100(2), 181–190. https:// doi. org/ 10. 1037/ 0021- 843X. 100.2. 181 Huang-Pollock, C., Mikami, A. Y., Pfiffner, L., & McBurnett, K. De Los Reyes, A., Augenstein, T. M., Wang, M., Thomas, S. A., Drabick, (2009). Can Executive Functions Explain the Relationship D. A. G., Burgers, D. E., & Rabinowitz, J. (2015). The validity Between Attention Deficit Hyperactivity Disorder and Social of the multi-informant approach to assessing child and adolescent Adjustment? Journal of Abnormal Child Psychology, 37(5), mental health. Psychological Bulletin, 141(4), 858–900. https://doi. 679–691. https:// doi. org/ 10. 1007/ s10802- 009- 9302-8 org/ 10. 1037/ a0038 498. Humphreys, K. L., Katz, S. J., Lee, S. S., Hammen, C., Brennan, P. A., Diamantopoulou, S., Rydell, A.-M., Thorell, L. B., & Bohlin, G. & Najman, J. M. (2013). The association of ADHD and depres- (2007). Impact of Executive Functioning and Symptoms of Atten- sion: Mediation by peer problems and parent–child difficulties in tion Deficit Hyperactivity Disorder on Children’s Peer Relations two complementary samples. Journal of Abnormal Psychology, and School Performance. Developmental Neuropsychology, 32(1), 122(3), 854–867. https:// doi. org/ 10. 1037/ a0033 895. 521–542. https:// doi. org/ 10. 1080/ 87565 64070 13609 81 Jacobs, R. H., Reinecke, M. A., Gollan, J. K., & Kane, P. (2008). Dickson, K. S., Ciesla, J. A., & Zelic, K. (2017). The role of executive Empirical evidence of cognitive vulnerability for depression functioning in adolescent rumination and depression. Cognitive among children and adolescents: A cognitive science and develop- Therapy and Research, 41(1), 62–72. https:// doi. org/ 10. 1007/ mental perspective. Clinical Psychology Review, 28(5), 759–782. s10608- 016- 9802-0https:// doi. org/ 10. 1016/j. cpr. 2007. 10. 006 Dishion, T. J. (1990). The peer context of troublesome child and ado- Johnston, C., & Murray, C. (2003). Incremental Validity in the Psycho- lescent behavior. In P. E. Leone (Ed.), Understanding troubled and logical Assessment of Children and Adolescents. Psychological troubling youth (pp. 128–153). Sage Publications. Assessment, 15(4), 496–507. https:// doi. org/ 10. 1037/ 1040- 3590. Enders, C. K. (2010). Applied missing data analysis. Guilford Press.15.4. 496. Favre, T., Hughes, C., Emslie, G., Stavinoha, P., Kennard, B., & Car- Joormann, J., & Gotlib, I. H. (2010). Emotion regulation in depression: mody, T. (2008). Executive functioning in children and adoles- Relation to cognitive inhibition. Cognition and Emotion, 24(2), cents with major depressive disorder. Child Neuropsychology, 281–298. https:// doi. org/ 10. 1080/ 02699 93090 34079 48 15(1), 85–98. https:// doi. org/ 10. 1080/ 09297 04080 25773 11. Kenny, D. A., Kaniskan, B., & McCoach, D. B. (2015). The Perfor- Fenesy, M. C., & Lee, S. S. (2019). Association of ADHD and Execu- mance of RMSEA in Models With Small Degrees of Freedom. tive Functioning With Childhood Depression. Journal of Atten- Sociological Methods & Research, 44(3), 486–507. https:// doi. org/ 10. 1177/ 00491 24114 543236 tion Disorders, 23(11), 1368–1378. https://doi. or g/10. 1177/ 10870 54717 723983. 1 3 Research on Child and Adolescent Psychopathology (2022) 50:753–770 769 Kline, R. B. (2015). Principles and practice of structural equation Clinical Psychology Review, 34(8), 595–607. https:// doi. org/ 10. modeling. Guilford publications.1016/j. cpr. 2014. 10. 002 Kofler, M. J., Irwin, L. N., Soto, E. F., Groves, N. B., Harmon, S. L., Meinzer, M. C., Pettit, J. W., Waxmonsky, J. G., Gnagy, E., Molina, & Sarver, D. E. (2019). Executive functioning heterogeneity in B. S. G., & Pelham, W. E. (2016). Does Childhood Attention- pediatric ADHD. Journal of Abnormal Child Psychology, 47(2), Dec fi it/Hyperactivity Disorder (ADHD) Predict Levels of Depres - 273–286. https:// doi. org/ 10. 1007/ s10802- 018- 0438-2 sive Symptoms during Emerging Adulthood? Journal of Abnor- Kovacs, M. (1992). Children’s depression inventory: Manual. North mal Child Psychology, 44(4), 787–797. https:// doi. org/ 10. 1007/ Tonawanda, NJ: Multi-Health Systems.s10802- 015- 0065-0 Lawson, G. M., & Farah, M. J. (2017). Executive function as a mediator Miller, M., Nevado-Montenegro, A. J., & Hinshaw, S. P. (2012). Child- between SES and academic achievement throughout childhood. hood Executive Function Continues to Predict Outcomes in Young International Journal of Behavioral Development, 41(1), 94–104. Adult Females with and Without Childhood-Diagnosed ADHD. https:// doi. org/ 10. 1177/ 01650 25415 603489 Journal of Abnormal Child Psychology, 40(5), 657–668. https:// Larzelere, R. E., Kuhn, B. R., & Johnson, B. (2004). The interven-doi. org/ 10. 1007/ s10802- 011- 9599-y tion selection bias: An underrecognized confound in intervention Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, research. Psychological Bulletin, 130(2), 289–303. https://doi. or g/ A., & Wager, T. D. (2000). The unity and diversity of executive 10. 1037/ 0033- 2909. 130.2. 289 functions and their contributions to complex “Frontal Lobe” tasks: Lee, K., Bull, R., & Ho, R. M. H. (2013). Developmental Changes in A latent variable analysis. Cognitive Psychology, 41(1), 49–100. Executive Functioning. Child Development, 84(6), 1933–1953. https:// doi. org/ 10. 1006/ cogp. 1999. 0734 https:// doi. org/ 10. 1111/ cdev. 12096 Muthén, L. K., & Muthén, B. O. (1998 –2010). Mplus user’s guide (6th Lee, S. S., & Hinshaw, S. P. (2006). Predictors of Adolescent Func- ed.). Los Angeles, CA: Author. tioning in Girls With Attention Deficit Hyperactivity Disorder Nelson, T. D., Kidwell, K. M., Nelson, J. M., Tomaso, C. C., Hankey, (ADHD): The Role of Childhood ADHD, Conduct Problems, and M., & Espy, K. A. (2018). Preschool Executive Control and Inter- Peer Status. Journal of Clinical Child & Adolescent Psychology, nalizing Symptoms in Elementary School. Journal of Abnormal 35(3), 356–368. https:// doi. org/ 10. 1207/ s1537 4424j ccp35 03_2. Child Psychology. https:// doi. org/ 10. 1007/ s10802- 017- 0395-1 Lewis, A. J., Bertino, M. D., Bailey, C. M., Skewes, J., Lubman, D. I., Nigg, J. T. (2017). Annual Research Review: On the relations among & Toumbourou, J. W. (2014). Depression and suicidal behavior self-regulation, self-control, executive functioning, effortful con- in adolescents: a multi-informant and multi-methods approach to trol, cognitive control, impulsivity, risk-taking, and inhibition for diagnostic classification. Frontiers in Psychology, 5(766), 1–24. developmental psychopathology. Journal of Child Psychology and https:// doi. org/ 10. 3389/ fpsyg. 2014. 00766. Psychiatry, 58(4), 361–383. https:// doi. org/ 10. 1111/ jcpp. 12675 Maalouf, F. T., Brent, D., Clark, L., Tavitian, L., McHugh, R. M., Nigg, J. T., Blaskey, L. G., Stawicki, J. A., & Sachek, J. (2004). Evaluat- Sahakian, B. J., & Phillips, M. L. (2011). Neurocognitive impair- ing the endophenotype model of ADHD neuropsychological defi- ment in adolescent major depressive disorder: state vs. trait illness cit: Results for parents and siblings of children with ADHD com- markers. Journal of Affective Disorders, 133(3), 625–632. https:// bined and inattentive subtypes. Journal of Abnormal Psychology, doi. org/ 10. 1016/j. jad. 2011. 04. 041. 113(4), 614–625. https:// doi. org/ 10. 1037/ 0021- 843X. 113.4. 614 MacKinnon, D. P., & Fairchild, A. J. (2009). Current directions in Øie, M., Hovik, K. T., Andersen, P. N., Czajkowski, P. N., & Skogli, mediation analysis. Current Directions in Psychological Science, E. W. (2016). Gender Differences in the Relationship Between 18(1), 16–20. https:// doi. org/ 10. 1111/j. 14678 721. 2009. 01598 Changes in ADHD Symptoms, Executive Functions, and Self- Martel, M. M., Markon, K., & Smith, G. T. (2017). Research Review: and Parent-Report Depression Symptoms in Boys and Girls With Multi-informant integration in child and adolescent psychopathol- ADHD: A 2-Year Follow-Up Study. Journal of Attention Disor- ogy diagnosis. Journal of Child Psychology and Psychiatry, 58(2), ders. https:// doi. org/ 10. 1177/ 10870 54716 664407 116–128. https:// doi. org/ 10. 1111/ jcpp. 12611 Patterson, G. R., & Stoolmiller, M. (1991). Replications of a dual fail- Martel, M. M., Nikolas, M., & Nigg, J. T. (2007). Executive Function ure model for boys’ depressed mood. Journal of Consulting and in Adolescents With ADHD. Journal of the American Academy Clinical Psychology, 59(4), 491–498. https:// doi. org/ 10. 1037/ of Child & Adolescent Psychiatry, 46(11), 1437–1444. https://doi. 0022- 006X. 59.4. 491 org/ 10. 1097/ chi. 0b013 e3181 4cf953 Pelham, W. E., Gnagy, E. M., Greenslade, K. E., & Milich, R. (1992). McCarty, C. A., Mason, W. A., Kosterman, R., Hawkins, J. D., Lengua, Teacher Ratings of DSM-III-R Symptoms for the Disruptive L. J., & McCauley, E. (2008). Adolescent School Failure Predicts Behavior Disorders. Journal of the American Academy of Child Later Depression Among Girls. Journal of Adolescent Health, & Adolescent Psychiatry, 31(2), 210–218. https://do i.or g/10 .1097 / 43(2), 180–187. https://doi. or g/10. 1016/j. jadoh ealt h.2008. 01. 023 00004 583- 19920 3000- 00006 McClintock, S. M., Husain, M. M., Greer, T. L., & Cullum, C. M. Petersen, A. C., Crockett, L., Richards, M., & Boxer, A. (1988). A self- (2010). Association between depression severity and neurocogni- report measure of pubertal status: Reliability, validity, and initial tive function in major depressive disorder: A review and synthesis. norms. Journal of Youth and Adolescence, 17(2), 117–133. https:// Neuropsychology, 24(1), 9–34. https://doi. or g/10. 1037/ a0017 336 doi. org/ 10. 1007/ BF015 37962 McLaughlin, K. A., & Nolen-Hoeksema, S. (2011). Rumination as Pliszka, S. R. (2019). ADHD and Anxiety: Clinical Implications. a transdiagnostic factor in depression and anxiety. Behaviour Journal of Attention Disorders, 23(3), 203–205. https:// doi. org/ Research and Therapy, 49(3), 186–193. https://doi. or g/10. 1016/j. 10. 1177/ 10870 54718 817365 brat. 2010. 12. 006 Powell, V., Riglin, L., Hammerton, G., Eyre, O., Martin, J., Anney, R., Meinzer, M. C., & Chronis-Tuscano, A. (2017). ADHD and the devel- Rice, F. (2020). What explains the link between childhood ADHD opment of depression: Commentary on the prevalence, proposed and adolescent depression? Investigating the role of peer relation- mechanisms, and promising interventions. Current Develop- ships and academic attainment. European Child & Adolescent mental Disorders Reports, 4(1), 1–4. https:// doi. or g/ 10. 1007/ Psychiatry 1–11. https:// doi. org/ 10. 1007/ s00787- 019- 01463-w s40474- 017- 0106-1 Powers, A., & Casey, B. (2015). The adolescent brain and the emer- Meinzer, M. C., Pettit, J. W., & Viswesvaran, C. (2014). The co- gence and peak of psychopathology. Journal of Infant, Child, and occurrence of attention-deficit/hyperactivity disorder and unipolar Adolescent Psychotherapy, 14(1), 3–15. https:// doi. org/ 10. 1080/ depression in children and adolescents: A meta-analytic review. 15289 168. 2015. 10048 89 1 3 770 Research on Child and Adolescent Psychopathology (2022) 50:753–770 Pritikin, J. N., Brick, T. R., & Neale, M. C. (2018). Multivariate normal Organizational Research Methods, 9(2), 194–201. https://doi. or g/ maximum likelihood with both ordinal and continuous variables, 10. 1177/ 10944 28105 285506. and data missing at random. Behavior Research Methods. https:// Violato, C., & Hecker, K. G. (2007). How to use structural equation doi. org/ 10. 3758/ s13428- 017- 1011-6 modeling in medical education research: A brief guide. Teaching Reitan, R. M., & Wolfson, D. (1992). Neuropsychological evaluation and Learning in Medicine, 19(4), 362–371. https:// doi. org/ 10. of older chilren. Neuropsychology Press.1080/ 10401 33070 15426 85 Roy, A., Oldehinkel, A. J., & Hartman, C. A. (2017). Cognitive Func- Wagner, S., Müller, C., Helmreich, I., Huss, M., & Tadić, A. (2015). A tioning in Adolescents with Self-Reported ADHD and Depres- meta-analysis of cognitive functions in children and adolescents sion: Results from a Population-Based Study. Journal of Abnor- with major depressive disorder. European Child & Adolescent Psy- mal Child Psychology, 45(1), 69–81. https:// doi. or g/ 10. 1007/ chiatry, 24(1), 5–19. https:// doi. org/ 10. 1007/ s00787- 014- 0559-2 s10802- 016- 0160-x Weber, E. B., Studeny, J., Kavanaugh, B. C., Cook, N. E., Gaudet, C. Schlomer, G. L., Bauman, S., & Card, N. A. (2010). Best practices E., McCurdy, K., & Holler, K. A. (2018). Pediatric Depression for missing data management in counseling psychology. Journal Symptoms, Executive Functioning Weaknesses, and Associated of Counseling Psychology, 57(1), 1–10. https:// doi. org/ 10. 1037/ Neuropsychological and Psychiatric Outcomes. Journal of Child a0018 082 and Family Studies, 27, 1661–1670. htt ps :// d oi. o r g/ 1 0. 10 07/ Seymour, K. E., Chronis-Tuscano, A., Iwamoto, D. K., Kurdziel, G., & s10826- 017- 0999-7 MacPherson, L. (2014). Emotion Regulation Mediates the Associ- Wechsler, D. (2002). Wechsler Individual Achievement Test—2nd ation Between ADHD and Depressive Symptoms in a Community Edition. Examiner’s Manual. San Antonio, TX: Psychological Sample of Youth. Journal of Abnormal Child Psychology, 42(4), Corportation. 611–621. https:// doi. org/ 10. 1007/ s10802- 013- 9799-8 Wechsler, D. (2003). Wechsler Intelligence Scale for Children, Fourth Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab-Stone, Edition: Administration and Scoring Manual. Psychological M. E. (2000). NIMH Diagnostic Interview Schedule for Children Corporation. Version IV (NIMH DISC-IV): Description, Differences From Weissman, A. S., Chu, B. C., Reddy, L. A., & Mohlman, J. (2012). Previous Versions, and Reliability of Some Common Diagnoses. Attention mechanisms in children with anxiety disorders and in Journal of the American Academy of Child & Adolescent Psy- children with attention deficit hyperactivity disorder: Implications chiatry, 39(1), 28–38. https:// doi. org/ 10. 1097/ 00004 583- 20000 for research and practice. Journal of Clinical Child & Adolescent 1000- 00014 Psychology, 41(2), 117–126. https:// doi. org/ 10. 1080/ 15374 416. Snyder, H. R., Friedman, N. P., & Hankin, B. L. (2019). Transdi-2012. 651993 agnostic mechanisms of psychopathology in youth: Executive Weston, R., & Gore, P. A., Jr. (2006). A brief guide to structural equa- functions, dependent stress, and rumination. Cognitive Ther- tion modeling. The Counseling Psychologist, 34(5), 719–751. apy and Research, 43(5), 834–851. https:// doi. or g/ 10. 1007/ https:// doi. org/ 10. 1177/ 00110 00006 286345 s10608- 019- 10016-z Wiebe, S. A., Sheffield, T., Nelson, J. M., Clark, C. A. C., Chevalier, Snyder, H. R., & Hankin, B. L. (2016). Spiraling out of control: Stress N., & Espy, K. A. (2011). The structure of executive function in generation and subsequent rumination mediate the link between 3-year-olds. Journal of Experimental Child Psychology, 108(3), poorer cognitive control and internalizing psychopathology. Clini- 436–452. https:// doi. org/ 10. 1016/j. jecp. 2010. 08. 008 cal Psychological Science, 4(6), 1047–1064. https:// doi. org/ 10. Willcutt, E. G., Doyle, A. E., Nigg, J. T., Faraone, S. V., & Pennington, 1177/ 21677 02616 633157. B. F. (2005). Validity of the Executive Function Theory of Atten- Snyder, A. M., Maruff, P., Pietrzak, R. H., Cromer, J. R., & Snyder, tion-Deficit/Hyperactivity Disorder: A Meta-Analytic Review. P. J. (2008). Effect of treatment with stimulant medication on Biological Psychiatry, 57(11), 1336–1346. https:// doi. or g/ 10. nonverbal executive function and visuomotor speed in children 1016/j. biops ych. 2005. 02. 006 with attention deficit/hyperactivity disorder (ADHD). Child Wolraich, M. L., Hagan, J. F., Allan, C., Chan, E., Davison, D., Earls, Neuropsychology, 14(3), 211–226. htt ps:// d oi. or g / 10. 10 80/ M., & Zurhellen, W. (2019). Clinical practice guideline for the 09297 04070 12200 05 diagnosis, evaluation, and treatment of attention-deficit/hyper - Snyder, H. R., Miyake, A., & Hankin, B. L. (2015). Advancing under- activity disorder in children and adolescents. Pediatrics, 144(4). standing of executive function impairments and psychopathology: https:// doi. org/ 10. 1542/ peds. 2019- 2528 Bridging the gap between clinical and cognitive approaches. Fron- Zelazo, P. D., Anderson, J. E., Richler, J., Wallner-Allen, K., Beaumont, tiers in Psychology, 6, 328. https:// doi. org/ 10. 3389/ fpsyg. 2015. J. L., Conway, K. P., & Weintraub, S. (2014). NIH Toolbox Cogni- 00328 tion Battery (CB): Validation of Executive Function Measures in Thapar, A., Collishaw, S., Pine, D. S., & Thapar, A. K. (2012). Depres- Adults. Journal of the International Neuropsychological Society, sion in adolescence. The Lancet, 379(9820), 1056–1067. https:// 20(6), 620–629. https:// doi. org/ 10. 1017/ S1355 61771 40004 72 doi. org/ 10. 1016/ S0140- 6736(11) 60871-4 Zelazo, P. D., & Cunningham, W. A. (2007). Executive function: mech- Timbremont, B., Braet, C., & Dreessen, L. (2004). Assessing Depres- anisms underlying emotion regulation. In J. J. Gross (Ed.), Hand- sion in Youth: Relation Between the Children’s Depression Inven- book of emotion regulation (p. 135–158). The Guilford Press. tory and a Structured Interview. Journal of Clinical Child & Zhou, Q., Chen, S. H., & Main, A. (2012). Commonalities and Differ - Adolescent Psychology, 33(1), 149–157. https:// doi. org/ 10. 1207/ ences in the Research on Children’s Effortful Control and Execu- S1537 4424J CCP33 01_ 14 tive Function: A Call for an Integrated Model of Self-Regulation. Twenge, J. M. J., Thomas, E., Rogers, M. L., & Martin, G. N. (2017). Child Development Perspectives, 6(2), 112–121. https:// doi. org/ Increases in Depressive Symptoms, Suicide-Related Outcomes, 10. 1111/j. 1750- 8606. 2011. 00176.x and Suicide Rates Among U.S. Adolescents After 2010 and Links to Increased New Media Screen Time. Clinical Psychological Sci- Publisher's Note Springer Nature remains neutral with regard to ence, 6(1), 3–17. https:// doi. org/ 10. 1177/ 21677 02617 723376. jurisdictional claims in published maps and institutional affiliations. Vandenberg, R. J. (2006). Introduction: statistical and methodological myths and urban legends: Where, pray tell, did they get this idea? 1 3

Journal

Research on Child and Adolescent PsychopathologySpringer Journals

Published: Jun 1, 2022

Keywords: ADHD; Executive functioning; Depression; SEM

There are no references for this article.