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

Learn More →

Reduced Alternating Gaze During Social Interaction in Infancy is Associated with Elevated Symptoms of Autism in Toddlerhood

Reduced Alternating Gaze During Social Interaction in Infancy is Associated with Elevated... In typical development, infants often alternate their gaze between their interaction partners and interesting stimuli, increasing the probability of joint attention toward surrounding objects and creating opportunities for communication and learning. Children with Autism Spectrum Disorder (ASD) have been found to engage less in behaviors that can initiate joint attention compared to typically developing children, but the role of such atypicalities in the development of ASD during infancy is not fully understood. Here, using eye tracking technology in a live setting, we show that 10- month-olds at high familial risk for ASD engage less in alternating gaze during interaction with an adult compared to low risk infants. These differences could not be explained by low general social preference or slow visual disengagement, as the groups performed similarly in these respects. We also found that less alternating gaze at 10 months was associated with more social ASD symptoms and less showing and pointing at 18 months. These relations were similar in both the high risk and the low risk groups, and remained when controlling for general social preference and disengagement latencies. This study shows that atypicalities in alternating gaze in infants at high risk for ASD emerge already during the first 10 months of life - a finding with theoretical as well as potential practical implications. . . . . . Keywords Autism Joint attention Communication Eye tracking Motivation Infant siblings Joint attention (JA) – whichcan be definedasthe sharingof such as pointing (Beuker et al. 2013; Carpenter et al. 1998). In attention between individuals and an object (Bruner 1975; other words, the infants no longer merely respond to joint at- Scaife and Bruner 1975) - plays an important role for learning tention; they initiate it as well. Although initiating joint attention and social interaction early in life. Young infants respond to (IJA) and responding to joint attention (RJA) may seem like others’ joint attention bids by following their gaze and pointing two sides of the same coin, evidence suggests that they repre- gestures (Schmitow et al. 2016; Corkum and Moore 1998). sent partially different processes. For example, longitudinal cor- This allows them to experience what others see, and facilitates relations within each type of JA are higher than correlations learning about objects and events in the world. After a while, between the two types of JA at a given time point (Mundy infants start directing others’ attention in order to share their et al. 2007). IJA and RJA also follow different developmental own experiences, first by alternating their gaze between an trajectories. Whereas RJA begins to emerge between 3 and interaction partner and an object, and later by the use of gestures 4 months of age (Gredebäck et al. 2010; D'Entremont et al. 1997; Perra and Gattis 2010), IJA behaviors typically emerge between 8 and 13 months (Beuker et al. 2013). Also at the * Emilia Thorup neural level, there is some evidence that the two types of JA emilia.thorup@psyk.uu.se are supported by partially different systems. It has been sug- gested that RJA relies primarily on a posterior attention system Uppsala Child and Baby Lab, Department of Psychology, Uppsala that regulates involuntary attention, whereas IJA seems to be University, Uppsala, Sweden supported by a later developing anterior network, controlling Karolinska Institutet Center of Neurodevelopmental Disorders volitional attention (Mundy et al. 2000, 2009;Van Heckeetal. (KIND), Pediatric Neuropsychiatry Unit, Department of Women’s& 2007). Further, IJA but not RJA is associated with activity in Children’s Health, Karolinska Institutet, Stockholm, Sweden reward-related brain areas (Schilbach et al. 2010;Gordonet al. BUP Stockholm, Center for Psychiatry Research, Stockholm County 2013) suggesting a link between IJA and motivation. Council, Stockholm, Sweden 1548 J Abnorm Child Psychol (2018) 46:1547–1561 It is well known that children with autism spectrum disor- gestural behaviors such as pointing and showing have long der (ASD) engage less in JA. Parental report of JA behaviors been part of most children’s repertoires. It is thus possible at 8 months has been shown to predict diagnostic status at that these behaviors as well as language had in part re- 7 years (Veness et al. 2014), thus suggesting that JA may be placed alternating gaze as a means of initiating JA. one of the earliest domains where infants with later ASD differ from other infants. Deficits in IJA are considered more Prospective Studies of IJA impairing and tend to persist longer in the development of autistic children than deficits in RJA (Mundy et al. 1994; ASD is rarely diagnosed before the age of 24–36 months, Sigman and Ruskin 1999; Gotham et al. 2007; Lord et al. which implies that assessing IJA in diagnosed children does 2000). The social motivation account (Chevallier et al. not capture possible differences in terms of early develop- 2012) proposes early diminished social motivation as a causal ment. Because 7–20% of younger siblings of children with factor behind the later occurring social-cognitive impairment ASD are expected to receive an ASD diagnosis themselves observed in autistic individuals. If infants who will later be (Messinger et al. 2015; Ozonoff et al. 2011;Gronborget al. diagnosed with ASD find sharing their experiences with 2013), studying younger siblings longitudinally has proven an others less rewarding, they can be expected to initiate JA to important means of gaining insight in the early development a lesser extent than other infants. Considering the important of ASD. role JA plays in social learning and interaction, less initiation Several studies have indicated that receiving an ASD diag- of JA would likely have an attenuating effect on several as- nosis is associated with lower levels or lower growth rates of pects of development. It is therefore imperative that we learn IJA during the second year of life (Macari et al. 2012;Landa more about the early development of IJA in ASD. and Garrett-Mayer 2007; Rozga et al. 2011; Yoder et al. 2009). Also, high risk (HR) infants, i.e. those with an older Measuring Different Types of IJA Behaviors and Their sibling with ASD, have been found to engage in fewer IJA Relation to Each Other and to ASD behaviors than low risk (LR) infants during this time period (Goldberg et al. 2005; Casseletal. 2007, but see Yirmiya et al. Young children may attempt to direct others’ attention by 2006). Only two of these studies (Macari et al. 2012;Landa the use of a variety of behaviors, and studies of IJA typi- and Garrett-Mayer 2007) measured alternating gaze specifi- cally measure gestural behaviors such as pointing and cally. The remaining studies measured IJA as a broader cate- showing as well as eye contact behaviors. Some (e.g. gory, including other communicative behaviors as well. It is Stone et al. 2000; Wetherby and Prizant 2002) include vo- therefore not possible to conclude whether the results were cal behaviors as well. One of the earliest means by which driven by differences in for example gestural behaviors, or infants initiate JA is by looking back and forth between an whether alternating gaze specifically differs between groups. interaction partner and an object or event. Alternating gaze To our best knowledge, only one study has prospectively with the purpose of attention sharing typically emerges investigated IJA before the first birthday (Ibanez et al. 2013). around 8–9 months of age (Beuker et al. 2013), while ges- This study showed that the level of IJA at 8 months predicted tural IJA behaviors such as pointing and showing typically diagnostic outcome (ASD or no diagnosis) at 30 months in a first emerge between 10 and 13 months (Beuker et al. HR sample. As with the majority of the studies mentioned 2013; Carpenter et al. 1998). Focusing on alternating gaze above however, this study did not focus specifically on alter- therefore entails a possibility to detect differences earlier nating gaze, but measured IJA in a broader sense. There is than what is possible when gestural IJA behaviors are mea- however one retrospective study (Clifford and Dissanayake sured. However, some studies have indicated that it is pre- 2009) that has reported a relation between alternating gaze dominantly in terms of the gestural behaviors that the per- as assessed using home videos recorded during the children’s formance of autistic children differs from that of typically first 11 months, and a measure of social responsiveness at developing children (Chiang et al. 2008; Mundy et al. 4years. 1994). In a study of children with ASD, Pickard and One issue that prospective studies can help clarifying is Ingersoll (2015) assessed eye contact based and gestural the timing of the onset of difficulties in alternating gaze in IJA separately. It was discovered that the two types of infants who later develop autism. This question is theoret- IJA were unrelated to each other, and that only gestural ically important because, as pointed out by Jones et al. IJA was associated with language, RJA and imitation. (2014), current research cannot tell us whether atypical This led the authors to suggest that eye contact based IJA alternating gaze could be considered an early risk factor (such as alternating gaze) may be a less optimal measure in itself, or a later consequence of other risk factors. ASD research. It is important to note though, that the chil- Considering that most autistic symptoms emerge during dren in the study by Pickard and Ingersoll (2015)were2– the second year of life (Jones et al. 2014), more studies of alternating gaze before the first birthday are needed. 7 years old, and therefore had reached an age when J Abnorm Child Psychol (2018) 46:1547–1561 1549 Variables that may Influence Alternating Gaze Aims of the Study The majority of the studies assessing IJA have used instru- In the current study, eye tracking technology was used in a live ments where interaction between the child and an experiment- setting, allowing infants’ gaze to be recorded while they were er is coded by a human observer based on a predefined proto- interacting with another (adult) person in an IJA eliciting task. col ((e.g. the Early Social Communication Scales (ESCS; When they reached 18 months, an age where autistic symp- Mundy et al. 1996) or the Screening Tool for Autism in toms are detectable in several domains (Jones et al. 2014), the Two-year-olds (STAT; Stone et al. 2000)). Although this is Autism Diagnostic Observation Schedule-Toddler Module well suited for measuring gestural IJA such as pointing, anal- (ADOS-T; Lord et al. 2012) was administered in order to yses of gaze alternations are likely to benefit from the use of obtain measures of ASD symptom level, as well as gestural more precise methods. Eye tracking has the advantage of ren- IJA. dering exact measures of gaze behavior that allow for various First, we tested the hypothesis that the HR group would analyses of visual attention to be conducted in order to test display less alternating gaze behavior than the LR group at different hypotheses (Falck-Ytter et al. 2013). It can therefore 10 months of age. Second, the relation between alternating be used to generate measures of factors that may potentially gaze at ten months and later ASD symptoms was assessed. have an impact on IJA behaviors. As previously mentioned, Here, the hypothesis was that the extent to which infants en- less IJA engagement could be a consequence of a decreased gaged in alternating gaze at 10 months would be negatively preference for social stimuli in general, which could be associated with ASD symptomatology at 18 months. Next, in assessed via looking times at social and non-social stimuli. It order to explore the potential mechanisms behind alternating has also been shown that infants with later ASD may have gaze and how they may differ between groups, measures of difficulties disengaging their attention from one stimulus to general social preference as well as disengagement ability another (Elsabbagh et al. 2013;Elisonetal. 2013; were included. No a priori hypotheses were formed regarding Zwaigenbaum et al. 2005), which could result in less alternat- these measures. Finally, we wanted to assess the validity of ing gaze. The possibility therefore remains that the lower fre- alternating gaze as a measure of IJA, by investigating its rela- quency of alternating gaze that has been found in infants with tion to later occurring gestural IJA behaviors (showing and later ASD in some studies (Ibanez et al. 2013; Landa and pointing). It was predicted that alternating gaze at 10 months Garrett-Mayer 2007; Macari et al. 2012) is in fact a direct would be positively associated with showing and pointing at consequence of disengagement difficulties. This could also 18 months. potentially explain the lack of a concurrent correlation be- tween different types of IJA detected in autistic preschoolers (Pickard and Ingersoll 2015), since disengagement difficulties Methods would be expected to influence gaze behavior more than ges- tural behaviors. To our knowledge, only one study has inves- Participants tigated the relation between IJA and the aforementioned fac- tors. In a cross-sectional study of three-year-olds with ASD, A total of 67 infants participated in the study (final sample Schietecatte et al. (2012) reported a negative relation between after exclusion; for participant characteristics, see Table 1). disengagement latencies and eye contact based IJA (including Fifty-one infants (29 girls, 22 boys) were HR infants, having alternating gaze), whereas a non-significant trend was found at least one older full sibling with ASD. Sixteen infants (6 between the former and social preference. girls, 10 boys) were LR infants, having at least one older Despite the previously mentioned advantages of eye track- typically developing full sibling and no family history of ing, the most frequently used application of the method has a ASD. Data from an additional 9 infants (6 HR and 3 LR) clear disadvantage in that it uses pre-recorded stimuli (images was collected but excluded due to poor data quality or the or videos). Humans look at others differently when watching infants not contributing enough data (see Analysis section). them on video as compared to when watching them in real life All infants were part of an ongoing longitudinal project (Early (Foulsham et al. 2011; Laidlaw et al. 2011), and it has been Autism Sweden; EASE; www.smasyskon.se)following HR shown that measuring live interaction evokes different re- and LR infants from 5 to 72 months. The HR group was sponses than displaying pre-recorded stimuli, both in terms recruited through advertisements, the project’s webpage and of brain activation (Redcay et al. 2010; Schilbach et al. clinical units. The LR group was recruited from a database of 2010; Shimada and Hiraki 2006) and behavior (Risko et al. families who had indicated interest in participating in research 2012;Freethet al. 2013). Video stimuli can therefore not be with their infants. Both groups consisted primarily of infants considered optimal when the focus is on social behaviors such from the larger Stockholm, Sweden, area. Exclusion criteria as IJA. Instead, eye tracking with live stimuli should be were pre-term birth (< 36 weeks) and confirmed or suspected considered. medical problems, including visual/auditory impairment. 1550 J Abnorm Child Psychol (2018) 46:1547–1561 Table 1 Participant characteristics by group, final samples (M/SD) Measure HR group N = 51 (29 girls) LR group N = 16 (6 girls) Pairwise comparison p-value Age at 10 month assessment 10.45/0.43 10.37/0.55 > 0.25 MSEL Early Learning Composite 10 months 99.92/13.50 101.44/11.64 > 0.25 Age at 18 month assessment 18.51/0.76 18.51/0.94 > 0.25 MSEL Early Learning Composite 18 months 97.29/15.67 97.75/13.54 > 0.25 b a SES - 0.02/0.84 0.08/084 > 0.25 ADOS-T Total CSS 18 months 3.18/1.74 2.06/1.53 0.009 ADOS-T SA CSS 3.24/1.91 1.94/1.73 0.004 ADOS-T RRB CSS 3.98/2.09 3.31/2.12 0.204 ADOS-T Showing 18 months 0.92/0.91 0.44/0.81 0.03 ADOS-T Pointing 18 months 0.71/0.81 0.38/0.50 0.157 Independent samples t-test Socioeconomic status calculated on the basis of parental education and income (equal weighting), expressed as a z-score (for this measure, N =50 inthe HR group and 15 in the LR group since two families did not disclose this information) Mann-Whitney U Test There was no difference between HR and LR groups in terms experimenter moved a squeaky toy across predefined calibra- of socioeconomic status based on family income and educa- tion points, making the toy emit a sound at each point to attract tion level. The developmental level of the infants was assessed the attention of the infant. The procedure was repeated if nec- using the Mullen Scales of Early Learning (MSEL; Mullen essary until calibration was satisfactory (determined via in- 1995) at each visit, and did not differ between groups. The spection of online gaze replay on a monitor in the background diagnosis of the older sibling was confirmed through inspec- of the room). The live eye tracking session comprised a vari- tion of obtained child psychiatric or pediatric records (more ety of different tasks and lasted approximately 10–15 min in than 70% of all assessments included the ADOS (Lord et al. total. Data from the IJA task and the disengagement task (de- 2000) and/or the Autism Diagnostic Interview-Revised (ADI- scribed below) will be reported in the present paper, as well as R; Rutter et al. 2003)). The study was approved by the a measure of general social preference that was obtained dur- Regional Ethical Review Board in Stockholm and conducted ing all other parts of the session. The IJA task always took in accordance with the standards specified in the 1964 place at the end of the session, and the disengagement task Declaration of Helsinki. All parents provided written in- was conducted in the middle of the session. formed consent. Stimuli and Procedure 10-Month Visit The participants took part in a comprehensive assessment, typically spending 4–5 h in the lab. This study includes data from the eye tracking experiment and the MSEL (Mullen 1995). The eye tracking was administered early during the visit. The infant was seated on the lap of the parent, at a distance of 200 cm from the experimenter, who was seated at a low table (see Fig. 1). At each side of the table was an oblong transpar- ent lamp. The experimenter controlled the lamps with two remote controls, hidden beneath the table and thus invisible to the child. A Tobii TX300 eye tracker, placed on a table in Fig. 1 Sketch of the experimental setting. The infant was seated at a front of the infant, was used to record the infant’sgaze witha distance of 200 cm from the experimenter. The infant’s gaze was sample rate of 120 Hz. Two video cameras recorded the be- recorded by a Tobii TX300 eye tracker (placed at the table between the havior of the infant and the stimulus area. Before the session, a infant and experimenter). The infant’s behavioraswell asthe scenearea five point calibration procedure was conducted. The was recorded by two video cameras (not visible in the sketch) J Abnorm Child Psychol (2018) 46:1547–1561 1551 The IJA task consisted of four trials and lasted approxi- autism in children younger than 30 months, with low verbal mately 1–1.5 min in total. The experimenter started each trial skills and a developmental age of at least 12 months. The by attracting the infant’s attention. Once that was achieved the ADOS-T was administered by three experienced clinical psy- experimenter turned on the lamp on his/her right side and chologists. One of the clinicians had achieved formal research lights started to flash, changing color approximately every reliability, and was supervising the work of the other two, who second. The switch controlling the lamp was out of sight for both had extensive clinical experience in using the instrument. the infant, who was not aware that the lights were controlled by the experimenter. While the lights were flashing, the ex- Analysis perimenter sat still looking at the infant, intermittently speak- ing in a quiet tone of voice. This was to provide the infant with Data text files were exported from the Tobii Studio Software. an opportunity to initiate joint attention toward the lamp. If the Analyses were conducted in MATLAB r2015b (MathWorks child made an explicit attempt to direct the experimenter’s Inc., Natick, MA), using the TimeStudio project framework, a attention to the lights, e.g. by pointing or vocalizing (e.g. MATLAB-based open access analysis tool for time series data saying Bthere^ while looking toward the lights), the experi- (version 3.15; timestudioproject.com; Nyström et al. 2016) menter responded by turning toward the lights and and SPSS (SPSS Inc., Chicago, IL). commenting on them. The rationale behind this was to not upset the infant or possibly extinguish IJA behaviors by being IJA Task at 10 Months unresponsive. The lights were flashing for a total of 10 s. The lamp was then turned off, and the second trial was initiated. For each individual recording, four areas of interest (AOIs) All four trials were identical, except that the lights on the were defined from which gaze data was extracted. One cov- experimenter’s right side were activated for the first and third ered the face of the experimenter, two covered the lamps and trials, and the lights on the left side were activated for the the fourth AOI covered the entire scene area. The face AOI second and fourth trials. subtended 11 by 11 visual degrees and the lamp AOIs each In order to check for differences in the ability to disengage subtended 14 by 17 visual degrees. A fixation filter (Tobii from one stimulus to another, a modified version of the disen- Fixation Filter) with a velocity threshold of 35 pixels/ gagement task from the Autism Observation Scale for Infants window and a distance threshold of 35 pixels was applied. (AOSI; Bryson et al. 2008) was included in the eye tracking Because manual inspection revealed that on some trials the session. The experimenter shook a rattle approximately 20 cm lights were activated slightly shorter than 10 s, only data from to the right from his/her face, keeping another rattle hidden the first 9 s of light activation time were analyzed for all trials. underneath the table. Once the infant looked at the first rattle, Only trials with at least 50% gaze data during light activation the second was presented and shaken at the left side of the were included in the analysis. Further, we required at least two experimenter’s face. In total six disengagement trials were valid trials (out of four) from each child. This resulted in an conducted, alternating between the left and right sides. average of 3.20 (SD = 0.87) usable trials in the HR group and The experiment was performed by six individual experi- 3.31 (SD = 0.87) usable trials in the LR group (p >0.25). As menters (2 males, 4 females). In order to minimize the influ- the primary measure of alternating gaze behavior, we used the ence of individual experimenters, each experimenter was mean number of gaze alternations between the face of the trained according to a written protocol and a video template. experimenter and the flashing lights (i.e. the sum of gaze al- The MSEL (Mullen 1995) was administered after the eye ternations from lights to face and face to lights) per trial. In tracking experiment. order to control for general differences in looking patterns and durations, the total number of fixations on the scene area and 18-Month Visit the total looking times in the different AOIs were also com- pared across groups. Similar to the 10-month visit, the children spent a total of 4– 5 h in the lab taking part in a variety of tasks and assessments. Disengagement Task at 10 Months This study includes data from the MSEL and the ADOS-T. The MSEL was administered early during the day, and the Two AOIs covering the two rattles were defined, subtending 6 ADOS-T was typically administered midday after a lunch by 18 visual degrees each. The dependent measure was the time break. The ADOS (Lord et al. 1999) is a widely used diag- it took the infant’s gaze to enter the AOI of the second rattle nostic instrument in autism assessment, and recommended as once it became visible to the infant. All trials in which the first choice instrument by several international clinical guide- infants made a predictive gaze shift (i.e. disengaged from one lines (e.g. NICE.org.uk). Considering the young age of the rattle before the other was presented) were excluded from the participants, the Toddler Module (Lord et al. 2012) was used, analysis. Each infant had to contribute at least two valid disen- which has been developed to assess behaviors indicative of gagement trials in order to be included, which resulted in 57 1552 J Abnorm Child Psychol (2018) 46:1547–1561 infants (45 HR and 12 LR) contributing data to this analysis. scale from 0 to 3 with lower values indicating higher incidence The mean number of usable trials was 3.98 (SD = 1.84) in the and quality of the behaviors. The ADOS-T also contains an HR group and 3.80 (SD = 2.21) in the LR group (p >0.25). item where alternating gaze is scored. Unfortunately, our data did not contain enough variation for analyses regarding this Social Preference at 10 Months item to be conducted. In order to assess possible differences in how much relative Statistical Analysis time the infants spent looking at the experimenter, a social preference quotient was calculated during the entire live eye The ADOS-T variables were not normally distributed and tracking session except the IJA task and the disengagement non-parametric tests (Spearman correlations and multiple lin- task. The eye tracking session builds on naturalistic social ear regression analysis with bootstrap for coefficients (1000 interaction between the experimenter and the infant. It in- samples)) were therefore used to assess relations and group cludes a multitude of components such as the experimenter comparisons involving these variables unless otherwise spec- playing with and looking at a variety of toys; the experimenter ified. All other variables were normally distributed with equal talking to and calling the infant’s name etc. Two tasks are variances across groups, and thus analyzed by the use of para- described elsewhere (Thorup et al. 2016; Nyström et al. metric statistics. For distributions of the experimental vari- 2017). The present objects were at all times kept out of the ables, see Fig. 2. The gender distribution did not differ signif- experimenter’s face area, which allowed us to calculate a ratio icantly between groups, X(1) = 2.23, p = 0.136, and no differ- of relative face looking. The total amount of the time the ences between boys and girls were found on any of the out- infants spent looking at the experimenter’s face was divided come variables, all p-values ≥0.216. Gender was therefore not by the total amount of time they spent looking anywhere on used as a covariate. the scene area, including the experimenter and the various present objects. Higher values would thus indicate a higher preference for the experimenter over the non-social elements. Results Relative looking times at social vs. non-social stimuli have previously been used to evaluate social interest in infants High vs Low Risk Comparisons at 10 Months and toddlers at risk for ASD (e.g. Pierce et al. 2016; Chawarska et al. 2013). To be included in the social preference An independent samples t-test on the mean number of gaze analysis the infants were required to look at the scene (con- alternations between experimenter and lights revealed a group tribute valid gaze data) for at least 25% of the entire session, difference, t(65) = 2.66, p = 0.012, d = 0.77, with the LR which resulted in the inclusion of 65 infants (50 HR, 15 LR). group making more such gaze alternations than the HR group (see Fig. 3). The groups did not differ in terms of the duration ADOS-T Measures at 18 Months of time they spent looking at either the face of the experiment- er, t(65) = 0.84, p >0.25, d = 0.24, or the lights, t(65) = 0.17, Given the young age of the children, we did not classify differ- p >0.25, d = 0.05 (for descriptive statistics for these and the ent outcome groups, but rather investigated autism symptoms following eye tracking measures, see Table 2). The mean quantitatively, operationalized as performance on the ADOS-T. number of fixations on the scene area also did not differ be- As a general measure of ASD symptomatology the Total Social tween groups, t(65) = −0.01, p >0.25, d =0.002. In order to Affect and Restricted Repetitive Behavior Score was used. In check for possible effects of experimenter, we compared the order to test whether alternating gaze may be differently related alternating gaze performance of the children who saw the to the social-communicative domain and the restricted and re- experimenter who tested the largest amount of infants (48%) petitive behavior (RRB) domain,we alsousedtheSocial Affect to the performance of the remaining infants (who each saw 1 (SA) and RRB scores separately. For these three measures, of the 5 other experimenters). No difference was found be- rather than using the raw scores we used the calibrated severity tween the groups, t(65) = − 1.20, p =0.233, d =0.29. scores (CSS), as suggested by Esler et al. (2015). CSS have The results from the disengagement task showed that the been shown to result in more uniform distributions across age groups did not differ in terms of the time it took the infants to and language level and to be less influenced by non-ASD spe- disengage their attention from one active stimulus to another, cific child characteristics compared to raw scores (Esler et al. t(55) = 0.62, p >0.25, d = 0.22. A significant negative corre- 2015; Gotham et al. 2009). CSS range from 1 to 10, with higher lation between disengagement and alternating gaze was found values indicating more symptoms and a higher degree of con- in the LR group, r(10) = − 0.77, p = 0.004, suggesting that cern. In order to investigate the relation between alternating longer latencies to disengage were associated with less alter- gaze at 10 months and later gestural IJA, the items showing nating gaze. In the HR group, a marginally significant trend in and pointing were used. Both items are scored on a four point the same direction was detected, r(43) = − 0.28, p =0.062. J Abnorm Child Psychol (2018) 46:1547–1561 1553 Fig. 2 Histograms showing distributions for the experimental variables group; e) social preference, HR group; f) social preference, LR group. per group. a) alternating gaze, HR group; b) alternating gaze, LR group; Note that none of the distributions differs significantly from normality c) disengagement latency, HR group; d) disengagement latency, LR according to the Kolmogorov-Smirnov test 1554 J Abnorm Child Psychol (2018) 46:1547–1561 alternating gaze alone explained 14% of the variance in ADOS-T scores, p =0.002. In order to test how alternating gaze was related to the social and the RRB domains, the regression approach de- scribed above was repeated twice, using the SA as well as RRB scores as dependent variables. The results of the regres- sion analysis with ADOS-T SA as dependent variable mim- icked those of the original regression model, R = 0.18, F(2,64) = 6.92, p = 0.002, whereas the model with ADOS-T RRB as dependent variable was non-significant, R = 0.02, F(2,64) = 0.61, p >0.25 (Table 3 and Fig. 4). The CSS scores for RRB deviated much more from normality than the other ADOS-T variables (this is a consequence of how the CSS are converted from raw scores in combination with the limited range of scores for this subscale). Therefore, we also conduct- Fig. 3 Number of gaze alternations made between the face of the ed non-parametric correlational analyses of the relation be- experimenter and the lights per second. Error bars represent standard errors. * p <0.05 tween alternating gaze and ADOS-T RRB (Spearman). No associations between alternating gaze and ADOS-T RRB No group difference was found in terms of social prefer- were found in either the HR group, r (49) = − 0.11, p >0.25, ence, t(63) = 0.69, p = > 0.25, d = 0.27 (independent samples or the LR group, r (14) = 0.20, p > 0.25. In contrast, and in t-test). Social preference and alternating gaze did not correlate line with the results from the parametric analysis above, sig- in either the HR group, r(48) = 0.23, p = 0.109, or the LR nificant non-parametric associations between alternating gaze group, r(13) = −0.23, p >0.25. and ADOS-T SA scores were found in both the HR group, r (49) = − 0.29, p = 0.036, and the LR group, r (14) = − 0.63, s s Longitudinal Relations Between 10-Month Measures p = 0.008. Taken together, these analyses indicate that the fre- and 18-Month Symptom Level quency of gaze alternations at 10 months is related to the social affect domain of ASD, but not to the RRB domain. In order to test whether alternating gaze could predict later Consequently, ADOS-T SA rather than total ADOS-T scores symptom severity, a hierarchical multiple regression was con- will be used as dependent measure in the analyses in the fol- ducted. Total ADOS-T score was entered as the dependent lowing section. variable, and as independent variables, alternating gaze and group were added in the first block, and the interaction be- tween group and alternating gaze was added in the second Potential Influences of Disengagement and Social block. Bootstrapping was used to account for the skewed dis- Preference on the Relation Between 10-Month tribution of the ADOS-T scores. The first model was signifi- Alternating Gaze and 18-Month ADOS-T Social Affect cant, R = 0.16, F(2,64) = 6.29, p = 0.003, and showed that Scores alternating gaze at 10 months, but not group status, made a significant contribution to 18 month ADOS-T scores. The Although there were no group differences in terms of disen- interaction term had no significant individual contribution to gagement or social preference at 10 months, it is still possible the ADOS-T scores and was thus excluded from the model that performance on those tasks influences the relation be- (Table 3). With no other factors included in the model, tween alternating gaze and later symptom level. To address Table 2 Descriptive statistics by Measure HR group N = 51 (29 girls) LR group N = 16 (6 girls) group for 10-month eye tracking measures (M/SD) Looking time (s) at model’s face 2.30/1.45 2.65/1.48 Looking time (s) at lights 4.41/1.55 4.49/1.45 N fixations on scene area 13.32/6.05 13.31/4.91 Disengagement latency (s) 0.74/0.25 0.79/0.20 b, c Social preference 0.44/0.12 0.47/0.10 Based on 45 HR and 12 LR infants Based on 50 HR and 15 LR infants Looking time at model divided by total looking time on scene area J Abnorm Child Psychol (2018) 46:1547–1561 1555 Table 3 Regression table displaying the results of analyses with ADOS-T SA scores, we can conclude that neither disengage- alternating gaze and group as independent variables, and Total ADOS-T ment latencies, social preference, nor their interactions with scores, ADOS-T SA scores, and ADOS-T RRB scores as dependent group status influenced the relation between early alternating variables gaze and later symptomatology. See Table 4. Dependent variable B SE B β Total ADOS-T score Constant 3.46 0.67 Longitudinal Relations Between 10-Month Alternating gaze −0.35 0.14 −0.31* Alternating Gaze and 18-Month Gestural IJA Group 0.71 0.49 0.18 ADOS-T SA score Constant 3.55 0.74 In order to investigate whether early alternating gaze was re- Alternating gaze −0.41 0.15 −0.33* lated to later gestural IJA, two bootstrapped hierarchical re- Group 0.84 0.54 0.19 gression analyses with the ADOS-T items showing and ADOS-T RRB score Constant 3.25 0.88 pointing as dependent variables were conducted. As before, alternating gaze and group were entered as independent vari- Alternating gaze 0.02 0.18 0.01 Group 0.69 0.64 0.14 ables in the first block, and the interaction term was entered in the second block. Both models were significant in the first step In neither of the regression models was the interaction between alternat- (for showing: R = 0.17, F(2,64) = 6.53, p =0.003; for ing gaze and group significant (β =0.07, p >0.25 with total ADOS-T as pointing: R = 0.13, F(2,64) = 4.83, p = 0.011) and as none dependent variable; β =0.13, p > 0.25 with ADOS-T SA as dependent of the interaction terms were significant, they were removed variable; β = − 0.29, p > 0.25 with ADOS-T RRB as dependent variable); *= p <0.005 from the models (Table 5). In both analyses, alternating gaze was the only significant predictor. Following up the regression this issue, another regression analysis was thus performed, models with zero-order correlations revealed significant neg- including only those infants who had contributed data to both ative correlations between alternating gaze and the ADOS-T the disengagement task and the social preference task (44 HR item showing in both the HR group, r (49) = −0.33, p =0.018, infants and 11 LR infants). Again, a bootstrapped hierarchical and the LR group, r (14) = −0.50, p = 0.047. In other words, multiple regression with ADOS-T SA score as dependent var- infants who engaged more in alternating gaze at 10 months iable was conducted, this time entering alternating gaze, dis- also engaged more in showing at 18 months. For pointing engagement, social preference and group in the first block, however, only a trend toward a negative correlation with al- and the three interaction terms between those variables and ternating gaze was found in the HR-group, r (49) = −0.25, p = group in the second block. The initial model was significant, 0.082, indicating that in this group, making fewer gaze alter- R =0.17, F(4,50) = 2.62, p = 0.046. None of the interaction nations at 10 months was associated with less pointing at terms contributed significantly to the ADOS-T SA scores and 18 months. In the LR group, the relation was similar in mag- they were therefore removed from the model. As alternating nitude and direction, but did not reach statistical significance, gaze remained the only variable that uniquely predicted later r (14) = − 0.32, p =0.223. Supplementary Analyses Finally, in order to test the robustness of the results, all anal- yses were repeated excluding individuals who could be Table 4 Regression table with alternating gaze, disengagement and social preference as independent variables and ADOS-T SA score as dependent variable BSEB β Constant 4.31 1.66 Alternating gaze −0.51 0.19 −0.41* Disengagement latency −1.92 1.20 −0.23 Social preference 2.54 2.42 0.15 Group 0.50 0.67 0.11 Fig. 4 Alternating gaze plotted against ADOS-T SA calibrated severity None of the interaction terms were significant (alternating gaze*group: β = − 0.26, p > 0.25; disengagement latency*group: β = − 1.28, p = scores. Figure shows individual data points and separate regression lines 0.215; social preference*group: β =0.96, p >0.25); * p =0.01 for each group 1556 J Abnorm Child Psychol (2018) 46:1547–1561 Table 5 Regression table with ADOS-T items showing and pointing as Discussion dependent variables This study assessed infants’ alternating gaze behavior between Dependent variable B SE B β an adult and an event occurring outside the adult’s field of vision Showing Constant 1.27 0.35 as a potential measure of early IJA. As hypothesized, we found Alternating gaze −0.21 0.07 −0.36* that 10-month-old infants at high familial risk for ASD made Group 0.25 0.25 0.12 fewer gaze alternations than low risk controls. Moreover, a lower Pointing Constant 1.00 0.30 frequency of gaze alternations at 10 months was associated with Alternating gaze −0.16 0.06 −0.33* more core ASD symptoms at 18 months. Thus, alternating gaze Group 0.15 0.22 0.09 behavior can discriminate LR controls from infants at familial risk for ASD, as well as be used to predict later autism symptoms In neither of the regression models was the interaction between alternat- dimensionally. When investigating the relations between alternat- ing gaze and group significant (β =0.28, p > 0.25 with showing as de- ing gaze and the SA and RRB domains of the ADOS-T sepa- pendent variable; β = − 0.20, p > 0.25 with pointing as dependent vari- able); * = p <0.05 rately, they remained significant in both groups in the former case, whereas no relation was detected with the second measure. suspected to have influenced the results due to one of two sets This suggests that alternating gaze is related to the social symp- of special circumstances, namely explicit initiation of JA dur- toms of ASD, but not to restricted and repetitive behaviors. ing the IJA task (pointing toward or commenting on the lights) The majority of the previous studies of early IJA have used or atypically high ADOS-T scores (LR infants only). more naturalistic methods or clinical instruments. Although these In total 5 HR and 1 LR infants made one or more explicit approaches are associated with several advantages - such as high attempts at directing the experimenter’s attention to the ecological validity - they also have some limitations, one being lights by pointing toward them or commenting on them. that it is difficult to control for factors other than IJA that may As noted, in these cases, the experimenter responded by explain the differences one observes. Taking advantage of precise turning toward the lights and verbally acknowledging them. eye tracking technology, one of the aims of the current study was Since this opens up for the possibility that the alternating to investigate the influence of two such potential variables, name- gaze behavior of these infants was affected by the change in ly attention disengagement and general social interest. Problems experimenter behavior, all analyses were repeated exclud- with disengagement of attention have previously been found in ing the 6 infants who pointed or vocalized. All results children with or at risk for ASD and could result in fewer gaze remained the same as before, except for that the group fac- alternations. Latencies to disengage from one stimulus to another tor now contributed significant variance to the total ADOS- were therefore measured. The performance of the groups did not Tscores, β =0.23, p = 0.011,aswellastothe ADOS-T SA differ on this task, and disengagement difficulties can thus be scores, β = 0.24, p = 0.019 (as in the previous analysis, sig- ruled out as a cause behind the fewer gaze alternations of the nificance levels using bootstrapping are reported). HR group. Longer disengagement latencies were however con- All data was checked for outliers, and investigating Cook’s currently associated with less alternating gaze in both groups at distances confirmed that there were no bivariate outliers in any 10 months. It is not surprising that infants who take longer to of the analyses. Considering the ADOS scores however, 2 disengage their attention display fewer gaze alternations, as mak- children in the LR group had (total and SA) scores about 2 ing frequent gaze shifts is dependent on fast disengagement la- SD (but <3 SD) above their group’s M. The ADOS is devel- tencies. Importantly though, disengagement latencies were unre- oped for clinical use, and some variation is expected when lated to later ASD symptoms in the current study. This may be used in a typical sample. Considering that the 2 LR infants surprising considering that previous studies have reported longer had MSEL scores in the typical range, we do not consider disengagement latencies in infants with a later ASD diagnosis these children outliers. Nevertheless, in order to control for (Elison et al. 2013; Elsabbagh et al. 2013; Zwaigenbaum et al. their possible influence on the results, all analyses were re- 2005; Kleberg et al. 2016). This discrepancy may be explained peated removing these 2 children. All significant results by differences in age between the studies, as none of the other remained, but just like in the analysis reported above, group studies measured disengagement at 10 months, but the relation was now a significant predictor of total ADOS-T scores, β = between visual disengagement and ASD clearly warrants further 0.29, p=0.001,aswellasADOS-TSA scores, β =0.31, p = investigation. In terms of social preference, we did not detect 0.002. either a group difference or any relations with alternating gaze Excluding infants either because they pointed/vocalized or ASD symptoms. A higher proportion of time spent looking at during the IJA task, or because they were LR infants with the experimenter (social preference) and more gaze shifts be- atypically high ADOS-T scores, thus enhanced the predic- tween experimenter and lights could both reflect a higher social tive power of the group factor, but did not change the interest or motivation. From that perspective, it may be surprising results otherwise. that our social preference measure was unrelated to both J Abnorm Child Psychol (2018) 46:1547–1561 1557 alternating gaze and later ASD symptoms. The reason for this phenomenon, less engaging in alternating gaze may well impact may be that alternating gaze (and IJA), reflects a more active later development negatively. When the infant initiates joint at- attempt toward sharing experiences with another than does pas- tention with an adult, the adult is likely to respond by providing sively looking at social stimuli. IJA by means of gaze alternation information or social interaction. Alternating gaze therefore ini- has previously been associated with increased activity in the tiates communication and interaction as well as provides oppor- ventral striatum, an area that has been linked to motivation and tunities for learning. Less engagement in IJA early in life could reward seeking, in a sample of TD adults (Schilbach et al. 2010). therefore be expected to have extensive consequences on several It is thus possible that infants who later display more ASD symp- areas of development. It has recently been suggested that patterns toms find attention sharing less rewarding than infants with lower of social looking, which appear to have a substantial genetic degrees of autistic traits, whereas the same is not necessarily the component in young children (Constantino et al. 2017), may case for more passive looking patterns. contribute to so-called evocative gene-environment correlations The study also aimed to clarify the extent to which early whereby the child actively shapes its own social environment alternating gaze is specifically related to established IJA behav- (Kennedy et al. 2017). Measuring alternating gaze is fast (the iors later on. Looking back and forth between a person and an current results are based on an experiment that only lasted 1– object is indeed a less explicit manifestation of joint attention 1.5 min in total) and does not necessarily require eye tracking compared to the later occurring gestural behaviors, such as equipment to be conducted. If the findings are replicated and the pointing and showing. As noted in the introduction, a previous task further improved, the alternating gaze measure could thus study demonstrated that eye contact based IJA, such as alternat- find practical use in the future. Apart from highlighting the role of ing gaze, was concurrently unrelated to gestural IJA in a sample alternating gaze in the early development of infants at risk for of autistic children (Pickard and Ingersoll 2015). In the current ASD, this study also contributes information about IJA in typical study, a higher frequency of alternating gaze at 10 months was development. Most studies on IJA and social functioning have significantly associated with more showing and marginally asso- focused on atypical populations, which entails that our ciated with more pointing at 18 months, suggesting that alternat- knowledge about their relation in typical development is rather ing gaze is longitudinally related to gestural IJA behaviors. As scarce. In a rare study of typical development, Van Hecke et al. suggested in the introduction, the lack of a correlation in the (2007) showed that a higher incidence of IJA at 12 months was studybyPickardandIngersoll (2015) may be related to the related to more optimal social functioning at 30 months. Their age of their participants. The children in that study were 2–7 years and our findings thus both indicate that the relation between old, and language may to some extent have replaced gestural and alternating gaze and later social functioning is not restricted to eye contact based IJA, thus weakening the relation between the clinical samples. This is also compatible with the view that the two. In the present study we measured both alternating gaze and autistic phenotype represents the extreme end of a continuum, rather than a discrete entity (Lundstrom et al. 2012 gestural IJA at ages where language is not yet an effective means ). of initiating JA for most children. The fact that a longitudinal Although the current study is one of the first to use eye relation between alternating gaze and gestural IJA was found in tracking to measure IJA, the technology has previously been the current study strengthens the validity of alternating gaze as an used for the same purpose by Billeci et al. (2016). In their early IJA measure. The measure further has a clear advantage study, 2-year-olds with and without ASD watched a video of compared to gestural measures in that it allows for earlier use. a model seated at a table where a toy car suddenly started to Pointing, for example, develops quite late with respect to when move. In contrast to the current results, it was discovered that investigations of IJA are of interest. Although pointing was not the ASD group made more gaze alternations between the the focus of the current study, all instances when infants pointed model and the car than the control group. When comparing toward the lights were coded. In total, only three infants pointed the two studies some caution must be given considering that during the experiment, rendering inferential statistical testing Billeci et al. (2016) focused on diagnosed children, whereas non-applicable. This confirms that at 10 months, pointing is not the current study assessed younger siblings. Nevertheless, the a useful measure of individual differences in IJA, as most infants discrepancy in results is quite striking, and we can think of at did not display the behavior at all. With the alternating gaze least two possible methodological explanations for this. First, measure, on the other hand, we found substantial individual dif- the car used by Billeci et al. (2016) was moving from the ferences at the same age. periphery toward the center of the screen. It is possible that A group difference in alternating gaze was detected already at the TD children interpreted this as the car entering the model’s 10 months, which is very shortly after the first signs of this field of vision, thus rendering alternating gaze in the function behavior typically emerges (Beuker et al. 2013). This suggests of drawing the model’s attention to the car superfluous. In the that the patterns of IJA behaviors in children with later ASD current study, the flashing lights were outside the model’sfield symptoms may be different from the time these behaviors start of vision at all times. Second, Billeci et al. (2016) used pre- to develop. Although the current results do not allow us to deter- recorded stimuli whereas the current study was conducted live. mine whether reduced alternating gaze is a primary or secondary The 2-year-old participants in the study by Billeci et al. (2016) 1558 J Abnorm Child Psychol (2018) 46:1547–1561 could be expected to have an appreciation for the difference symptoms at 18 months, and that this relation is independent of between a real person and one appearing on a screen. It there- differences in general social interest and disengagement abilities. fore remains a possibility that the TD participants’ low inci- The results suggest that a lower tendency to engage in alternating dence of alternating gaze was a consequence of their awareness gaze may be one of the earliest signs of ASD in the social do- that the person on screen was not affected by their behavior. main. Because these early atypicalities may negatively impact The current study highlights the potential as well as the impor- patterns of social interaction with caregivers, their precise role tance of using a more naturalistic approach when studying in shaping the development of children with ASD should be social behaviors (Falck-Ytter 2015; Falck-Ytter et al. 2015). addressed in future studies. By applying eye tracking technology in a live setting we were able to combine the advantages of more highly controlled par- Acknowledgements We would like to thank all the families participating in adigms with those of more ecologically valid ones. this study, and the members of Uppsala Child and Baby lab for valuable The current study has several limitations that deserve atten- comments on an earlier version of this paper. This research was supported tion. ADOS-T scores were used as an indication of the degree by the Swedish Research Council in partnership with FAS, FORMAS, and of ASD symptomatology. However, although the ADOS is VINNOVA (a cross-disciplinary research program concerning the mental health of children and young people; grant number 259-2012-24). The work considered gold standard in ASD assessments, it is not a stand of ET, PN, TFY, and SB was supported by the ESF COST Action BM1004 alone instrument that in itself gives a perfect indication of a BEnhancing the Scientific Study of Early Autism^ (ESSEA). TFY and SB child’s symptom level. This means, for example, that a high were supported by the EU-MSCA Initial Training Network (Brainview). ADOS-T score does not necessarily mean that diagnostic TFY was supported by The Bank of Sweden Tercentenary Foundation (P12-0270:1; NHS14-1802:1), The Swedish Research Council (2015– criteria are fulfilled. Thus the current approach should not be 03670), and the Niclas Öberg Foundation. SB was supported by the confused with one where diagnostic outcome groups are com- Swedish Research Council (No. 523-2009-7054). GG was supported by the pared, which would require a formal diagnostic assessment. European Research Council (StG: CACTUS 312292). The EASE Team The fact that the two groups differed substantially in size is consisted of Linn Andersson Konke, Maria Axnér, Karin Brocki, Elodie Cauvet, Therese L. Ekberg, Johan Lundin Kleberg, Elisabeth Nilsson Jobs, also a limitation, and even more so is the fact that the LR Sheila Norin and Eric Zander. group consisted of only 16 children (with fewer contributing data to some of the analyses). The conclusions regarding re- lations within the LR group therefore need to be interpreted Compliance with Ethical Standards with some caution, as well as the conclusion that the relation between alternating gaze and our 18-month measures are sim- Conflict of Interest The authors declare that they have no conflict of ilar in both groups. The t-tests that were conducted on the interest to declare. primary and supplementary measures however take the differ- Ethical Approval All procedures performed in studies involving human ences in sample sizes into account. Also, it is relatively com- participants were in accordance with the ethical standards of the institu- mon to include more HR than LR infants in longitudinal stud- tional and/or national research committee and with the 1964 Helsinki ies (Zwaigenbaum et al. 2005), partially because the HR sam- declaration and its later amendments or comparable ethical standards. ple will be split into multiple groups after diagnostic assess- ment. Importantly, the relation between alternating gaze at Informed Consent Informed consent was obtained from all individual participants included in the study. 10 months and the total ADOS-T score (explaining 14% of the variance) is based on a reasonable sample size due to Open Access This article is distributed under the terms of the Creative pooling of the groups. Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appro- Conclusions priate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Alternating gaze is one of the first means a young infant has to express and share his/her interest in various objects and events with another person. Looking back and forth between a parent References and an object may result in opportunities for social interaction (the parent acknowledges the infant’s behavior by attending to Beuker, K. T., Rommelse, N. N. J., Donders, R., & Buitelaar, J. K. (2013). him/her), word learning (the parent labels the object), as well as Development of early communication skills in the first two years of motoric exploration (the parent hands the infant the object). life. Infant Behavior & Development, 36(1), 71–83. https://doi.org/ 10.1016/j.infbeh.2012.11.001. Alternating gaze is thus a powerful behavior that affects later Billeci, L., Narzisi, A., Campatelli, G., Crifaci, G., Calderoni, S., development in multiple ways. The results of this study indicate Gagliano, A., et al. (2016). Disentangling the initiation from the not only that infants at risk for ASD engage less in alternating response in joint attention: an eye-tracking study in toddlers with gaze than low risk infants, but also that the tendency to engage in autism spectrum disorders. Translational Psychiatry, 6. https://doi. alternating gaze at 10 months is negatively related to social ASD org/10.1038/tp.2016.75. J Abnorm Child Psychol (2018) 46:1547–1561 1559 Bruner, J. S. (1975). From communication to language - psychological Falck-Ytter, T., Carlström, C., & Johansson, M. (2015). Eye contact modulates cognitive processing differently in children with autism. perspective. Cognition, 3(3), 255–287. https://doi.org/10.1016/ 0010-0277(74)90012-2. Child Development, 86(1), 37–47. https://doi.org/10.1111/cdev. Bryson,S.E.,Zwaigenbaum, L.,McDermott,C.,Rombough,V.,&Brian,J. (2008). The autism observation scale for infants: scale development and Foulsham, T., Walker, E., & Kingstone, A. (2011). The where, what and reliability data. Journal of Autism and Developmental Disorders, 38(4), when of gaze allocation in the lab and the natural environment. 731–738. https://doi.org/10.1007/s10803-007-0440-y. Vision Research, 51(17), 1920–1931. https://doi.org/10.1016/j. Carpenter, M., Nagell, K., & Tomasello, M. (1998). Social cognition, visres.2011.07.002. joint attention, and communicative competence from 9 to 15 months Freeth, M., Foulsham, T., & Kingstone, A. (2013). What affects social of age. [Review]. Monographs of the Society for Research in Child attention? Social presence, eye contact and autistic traits. PLoS One, Development, 63(4), V-143. 8(1). doi:https://doi.org/10.1371/journal.pone.0053286. Cassel, T. D., Messinger, D. S., Ibanez, L. V., Haltigan, J. D., Acosta, S. I., Goldberg, W. A., Jarvis, K. L., Osann, K., Laulhere, T. M., Straub, C., & Buchman, A. C. (2007). Early social and emotional communica- Thomas, E., et al. (2005). Brief report: early social communication tion in the infant siblings of children with autism spectrum disorders: behaviors in the younger siblings of children with autism. Journal of an examination of the broad phenotype. Journal of Autism and Autism and Developmental Disorders, 35(5), 657–664. https://doi. Developmental Disorders, 37(1), 122–132. https://doi.org/10.1007/ org/10.1007/s10803-005-0009-6. s10803-006-0337-1. Gordon, I., Eilbott, J. A., Feldman, R., Pelphrey, K. A., & Vander Wyk, B. Chawarska, K., Macari, S., & Shic, F. (2013). Decreased spontaneous C. (2013). Social, reward, and attention brain networks are involved attention to social scenes in 6-month-old infants later diagnosed with when online bids for joint attention are met with congruent versus autism spectrum disorders. Biological Psychiatry, 74(3), 195–203. incongruent responses. Social Neuroscience, 8(6), 544–554. https:// https://doi.org/10.1016/j.biopsych.2012.11.022. doi.org/10.1080/17470919.2013.832374. Chevallier, C., Kohls, G., Troiani, V., Brodkin, E. S., & Schultz, R. T. Gotham, K., Risi, S., Pickles, A., & Lord, C. (2007). The autism diag- (2012). The social motivation theory of autism. Trends in Cognitive nostic observation schedule: revised algorithms for improved diag- Sciences, 16(4), 231–238. https://doi.org/10.1016/j.tics.2012.02. nostic validity. Journal of Autism and Developmental Disorders, 37(4), 613–627. https://doi.org/10.1007/s10803-006-0280-1. Chiang, C.-H., Soong, W.-T., Lin, T.-L., & Rogers, S. J. (2008). Gotham, K., Pickles, A., & Lord, C. (2009). Standardizing ADOS scores Nonverbal communication skills in young children with autism. for a measure of severity in autism spectrum disorders. Journal of Journal of Autism and Developmental Disorders, 38(10), 1898– Autism and Developmental Disorders, 39(5), 693–705. https://doi. 1906. https://doi.org/10.1007/s10803-008-0586-2. org/10.1007/s10803-008-0674-3. Clifford, S., & Dissanayake, C. (2009). Dyadic and triadic behaviours in Gredebäck, G., Fikke, L., & Melinder, A. (2010). The development of infancy as precursors to later social responsiveness in young chil- joint visual attention: a longitudinal study of gaze following during dren with autistic disorder. Journal of Autism and Developmental interactions with mothers and strangers. Developmental Science, Disorders, 39(10), 1369–1380. https://doi.org/10.1007/s10803-009- 13(6), 839–848. https://doi.org/10.1111/j.1467-7687.2009.00945.x. 0748-x. Gronborg, T. K., Schendel, D. E., & Parner, E. T. (2013). Recurrence of Constantino, J. N., Kennon-McGill, S., Weichselbaum, C., Marrus, N., autism spectrum disorders in full- and half-siblings and trends over Haider, A., Glowinski, A. L., et al. (2017). Infant viewing of social time a population-based cohort study. JAMA Pediatrics, 167(10), scenes is under genetic control and is atypical in autism. Nature, 947–953. https://doi.org/10.1001/jamapediatrics.2013.2259. 547(7663), 340−+, doi:https://doi.org/10.1038/nature22999. Ibanez, L. V., Grantz, C. J., & Messinger, D. S. (2013). The development Corkum, V., & Moore, C. (1998). The origins of joint visual attention in of referential communication and autism symptomatology in high- infants. Developmental Psychology, 34(1), 28–38. https://doi.org/ risk infants. Infancy, 18(5), 687–707. https://doi.org/10.1111/j.1532- 10.1037/0012-1649.34.1.28. 7078.2012.00142.x. D'Entremont, B., Hains, S. M. J., & Muir, D. W. (1997). A demonstration Jones, E. J. H., Gliga, T., Bedford, R., Charman, T., & Johnson, M. H. of gaze following in 3- to 6-month-olds. Infant Behavior & (2014). Developmental pathways to autism: a review of prospective Development, 20(4), 569–572. https://doi.org/10.1016/s0163- studies of infants at risk. Neuroscience and Biobehavioral Reviews, 6383(97)90048-5. 39,1–33. https://doi.org/10.1016/j.neubiorev.2013.12.001. Elison, J. T., Paterson, S. J., Wolff, J. J., Reznick, J. S., Sasson, N. J., Gu, Kennedy, D. P., D’Onofrio, B. M., Quinn, P. D., Bölte, S., Lichtenstein, H., et al. (2013). White matter microstructure and atypical visual P., & Falck-Ytter, T. (2017). Genetic influence on eye movements to or ienting in 7-month-olds at risk for autism. American Journal of complex scenes at short timescales. Current Biology. https://doi.org/ Psychiatry, 170(8), 899–908. https://doi.org/10.1176/appi.ajp.2012. 10.1016/j.cub.2017.10.007. Kleberg, J. L., Thorup, E., & Falck-Ytter, T. (2016). Reduced visual Elsabbagh, M., Fernandes, J., Webb, S. J., Dawson, G., Charman, T., disengagement but intact phasic alerting in young children with Johnson, M. H., et al. (2013). Disengagement of visual attention in autism. [Article in press]. Autism Research. https://doi.org/10. infancy is associated with emerging autism in toddlerhood. 1002/aur.1675. Biological Psychiatry, 74(3), 189–194. https://doi.org/10.1016/j. Laidlaw, K. E. W., Foulsham, T., Kuhn, G., & Kingstone, A. (2011). biopsych.2012.11.030. Potential social interactions are important to social attention. Esler, A. N., Bal, V. H., Guthrie, W., Wetherby, A., Weismer, S. E., & [Article]. Proceedings of the National Academy of Sciences of the Lord, C. (2015). The autism diagnostic observation schedule, tod- United States of America, 108(14), 5548–5553. https://doi.org/10. dler module: standardized severity scores. JournalofAutismand 1073/pnas.1017022108. Developmental Disorders, 45(9), 2704–2720. https://doi.org/10. 1007/s10803-015-2432-7. Landa, R. J., Holman, K. C., & Garrett-Mayer, E. (2007). Social and Falck-Ytter, T. (2015). Gaze performance during face-to-face communi- communication development in toddlers with early and later diag- cation: a live eye tracking study of typical children and children with nosis of autism spectrum disorders. Archives of General Psychiatry, autism. Research in Autism Spectrum Disorders, 17,78–85. 64(7), 853–864. Falck-Ytter, T., Bölte, S., & Gredebäck, G. (2013). Eye tracking in early Lord, C., Rutter, M., DiLavore, P., & Risi, S. (1999). ADOS. Autism diagnostic observation schedule. Los Angeles: Western autism research. [Review]. Journal of Neurodevelopmental Disorders, 5,13. https://doi.org/10.1186/1866-1955-5-28. Psychological Services. 1560 J Abnorm Child Psychol (2018) 46:1547–1561 Lord, C., Risi, S., Lambrecht, L., Cook Jr., E. H., Leventhal, B. L., Biological Psychiatry, 79(8), 657–666. https://doi.org/10.1016/j. DiLavore, P. C., et al. (2000). The autism diagnostic observation biopsych.2015.03.032. schedule-generic: a standard measure of social and communication Redcay, E., Dodell-Feder, D., Pearrow, M. J., Mavros, P. L., Kleiner, M., deficits associated with the spectrum of autism. Journal of Autism Gabrieli, J. D. E., et al. (2010). Live face-to-face interaction during and Developmental Disorders, 30(3), 205–223. https://doi.org/10. fMRI: a new tool for social cognitive neuroscience. NeuroImage, 1023/a:1005592401947. 50(4), 1639–1647. https://doi.org/10.1016/j.neuroimage.2010.01. Lord, C., Luyster, R., Gotham, K., & Guthrie, W. (2012). Autism diag- Risko, E. F., Laidlaw, K. E. W., Freeth, M., Foulsham, T., & Kingstone, nostic observation schedule - toddler module manual. Los Angeles: A. (2012). Social attention with real versus reel stimuli: toward an Western Psychological Services. empirical approach to concerns about ecological validity. Frontiers Lundstrom, S., Chang, Z., Rastam, M., Gillberg, C., Larsson, H., in Human Neuroscience,1–11. https://doi.org/10.3389/fnhum.2012. Anckarsater, H., et al. (2012). Autism spectrum disorders and autisticlike traits similar etiology in the extreme end and the normal Rozga, A., Hutman, T., Young, G. S., Rogers, S. J., Ozonoff, S., Dapretto, variation. Archives of General Psychiatry, 69(1), 46–52. M., et al. (2011). Behavioral profiles of affected and unaffected Macari, S. L., Campbell, D., Gengoux, G. W., Saulnier, C. A., Klin, A. J., siblings of children with autism: contribution of measures of & Chawarska, K. (2012). Predicting developmental status from 12 mother-infant interaction and nonverbal communication. Journal to 24 months in infants at risk for autism spectrum disorder: a pre- of Autism and Developmental Disorders, 41(3), 287–301. liminary report. Journal of Autism and Developmental Disorders, Rutter, M., LeCouteur, A., & Lord, C. (2003). Autism diagnostic inter- 42(12), 2636–2647. view revised (ADI-R). Los Angeles: Western Psychological Messinger, D. S., Young, G. S., Webb, S. J., Ozonoff, S., Bryson, S. E., Services. Carter, A., et al. (2015). Early sex differences are not autism-specif- Scaife, M., & Bruner, J. S. (1975). Capacity for joint visual attention in ic: a baby siblings research consortium (BSRC) study. Molecular infant. Nature, 253(5489), 265–266. https://doi.org/10.1038/ Autism. https://doi.org/10.1186/s13229-015-0027-y. 253265a0. Mullen, E. M. (1995). Mullen scales of early learning (AGS ed.). Circle Schietecatte, I., Roeyers, H., & Warreyn, P. (2012). Exploring the nature Pines: American Guidance Service Inc.. of joint attention impairments in young children with autism spec- Mundy, P., Sigman, M., & Kasari, C. (1994). Joint attention, develop- trum disorder: associated social and cognitive skills. Journal of mental level, and symptom presentation in autism. Development and Autism and Developmental Disorders, 42(1), 1–12. Psychopathology, 6(3), 389–401. https://doi.org/10.1017/ Schilbach, L., Wilms, M., Eickhoff, S. B., Romanzetti, S., Tepest, R., s0954579400006003. Bente, G., et al. (2010). Minds made for sharing: initiating joint Mundy, P., Hogan, A., & Doehring, P. (1996). A preliminary manual for attention recruits reward-related neurocircuitry. Journal of the abriged early social communication scales (ESCS).Coral Cognitive Neuroscience, 22(12), 2702–2715. https://doi.org/10. Gables: University of Miami. 1162/jocn.2009.21401. Mundy, P., Card, J., & Fox, N. (2000). EEG correlates of the development Schmitow, C., Kochukhova, O., & Nystrom, P. (2016). Social perception: of infant joint attention skills. Developmental Psychobiology, 36(4), how do 6-month-old infants look at pointing gestures? Infant 325–338. Behavior & Development, 42,1 52–156. https://doi.org/10.1016/j. Mundy, P., Block, J., Delgado, C., Pomares, Y., Van Hecke, A. V., & infbeh.2015.12.001. Parlade, M. V. (2007). Individual differences and the development Shimada, S., & Hiraki, K. (2006). Infant’s brain responses to live and of joint attention in infancy. Child Development, 78(3), 938–954. televised action. NeuroImage, 32(2), 930–939. https://doi.org/10. https://doi.org/10.1111/j.1467-8624.2007.01042.x. 1016/j.neuroimage.2006.03.044. Mundy, P., Sullivan, L., & Mastergeorge, A. M. (2009). A parallel and Sigman, M., & Ruskin, E. (1999). Continuity and change in the social distributed-processing model of joint attention, social cognition and competence of children with autism, down syndrome, and develop- autism. Autism Research, 2(1), 2–21. https://doi.org/10.1002/aur.61. mental delays. Monographs of the Society for Research in Child Nyström, P., Bölte, S., Falck-Ytter, T., & Ease Team. (2017). Responding Development, 64(1), 1−+, doi:https://doi.org/10.1111/1540-5834. to other people’s direct gaze: alterations in gaze behavior in infants at risk for autism occur on very short timescales. Journal of Autism Stone, W. L., Coonrod, E. E., & Ousley, O. Y. (2000). Brief report: and Developmental Disorders, 47(11), 3498–3509. screening tool for autism in two-year-olds (STAT): development Nyström, P., Falck-Ytter, T., & Gredeback, G. (2016). The TimeStudio and preliminary data. Journal of Autism and Developmental project: an open source scientific workflow system for the behav- Disorders, 30(6), 607–612. https://doi.org/10.1023/a: ioral and brain sciences. Behavior Research Methods, 48(2), 542– 552. https://doi.org/10.3758/s13428-015-0616-x. Thorup, E., Nystrom, P., Gredeback, G., Bolte, S., Falck-Ytter, T., & Ozonoff, S., Young, G. S., Carter, A., Messinger, D., Yirmiya, N., Team, E. (2016). Altered gaze following during live interaction in Zwaigenbaum, L., et al. (2011). Recurrence risk for autism spectrum infants at risk for autism: an eye tracking study. Molecular Autism, 7. disorders: a baby siblings research consortium study. Pediatrics, doi:https://doi.org/10.1186/s13229-016-0069-9. 128(3), e488–e495. Van Hecke, A. V., Mundy, P. C., Acra, C. F., Block, J. J., Delgado, C. E. Perra, O., & Gattis, M. (2010). The control of social attention from 1 to 4 F., Parlade, M. V., et al. (2007). Infant joint attention, temperament, months. British Journal of Developmental Psychology, 28(4), 891– and social competence in preschool children. Child Development, 908. https://doi.org/10.1348/026151010x487014. 78(1), 53–69. Pickard, K. E., & Ingersoll, B. R. (2015). Brief report: high and low level Veness, C., Prior, M., Eadie, P., Bavin, E., & Reilly, S. (2014). Predicting initiations of joint attention, and response to joint attention: differ- autism diagnosis by 7 years of age using parent report of infant ential relationships with language and imitation. Journal of Autism social communication skills. Journal of Paediatrics and Child and Developmental Disorders, 45(1), 262–268. https://doi.org/10. Health, 50(9), 693–700. https://doi.org/10.1111/jpc.12614. 1007/s10803-014-2193-8. Wetherby, A. M., & Prizant, B. M. (2002). CSBS DP manual: Communication and symbolic behavior scales developmental Pierce, K., Marinero, S., Hazin, R., McKenna, B., Barnes, C. C., & profile. Baltimore: Paul H Brookes Publishing Co.. Malige, A. (2016). Eye tracking reveals abnormal visual preference Yirmiya, N., Gamliel, I., Pilowsky, T., Feldman, R., Baron-Cohen, S., & for geometric images as an early biomarker of an autism spectrum Sigman, M. (2006). The development of siblings of children with disorder subtype associated with increased symptom severity. J Abnorm Child Psychol (2018) 46:1547–1561 1561 autism at 4 and 14 months: social engagement, communication, and Disorders, 39(10), 1381–1391. https://doi.org/10.1007/s10803- 009-0753-0. cognition. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 47(5), 511–523. https://doi.org/10.1111/j.1469-7610. Zwaigenbaum, L., Bryson, S., Rogers, T., Roberts, W., Brian, J., & 2005.01528.x. Szatmari, P. (2005). Behavioral manifestations of autism in the first Yoder, P., Stone, W. L., Walden, T., & Malesa, E. (2009). Predicting social year of life. International Journal of Developmental Neuroscience, impairment and ASD diagnosis in younger siblings of children with 23(2-3 SPEC. ISS), 143–152. https://doi.org/10.1016/j.ijdevneu. autism spectrum disorder. Journal of Autism and Developmental 2004.05.001. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Abnormal Child Psychology Springer Journals

Reduced Alternating Gaze During Social Interaction in Infancy is Associated with Elevated Symptoms of Autism in Toddlerhood

Download PDF
15 pages

Loading next page...
 
/lp/springer-journals/reduced-alternating-gaze-during-social-interaction-in-infancy-is-ppMvpOtM1Y

References (83)

Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Psychology; Child and School Psychology; Neurosciences; Public Health
ISSN
0091-0627
eISSN
1573-2835
DOI
10.1007/s10802-017-0388-0
Publisher site
See Article on Publisher Site

Abstract

In typical development, infants often alternate their gaze between their interaction partners and interesting stimuli, increasing the probability of joint attention toward surrounding objects and creating opportunities for communication and learning. Children with Autism Spectrum Disorder (ASD) have been found to engage less in behaviors that can initiate joint attention compared to typically developing children, but the role of such atypicalities in the development of ASD during infancy is not fully understood. Here, using eye tracking technology in a live setting, we show that 10- month-olds at high familial risk for ASD engage less in alternating gaze during interaction with an adult compared to low risk infants. These differences could not be explained by low general social preference or slow visual disengagement, as the groups performed similarly in these respects. We also found that less alternating gaze at 10 months was associated with more social ASD symptoms and less showing and pointing at 18 months. These relations were similar in both the high risk and the low risk groups, and remained when controlling for general social preference and disengagement latencies. This study shows that atypicalities in alternating gaze in infants at high risk for ASD emerge already during the first 10 months of life - a finding with theoretical as well as potential practical implications. . . . . . Keywords Autism Joint attention Communication Eye tracking Motivation Infant siblings Joint attention (JA) – whichcan be definedasthe sharingof such as pointing (Beuker et al. 2013; Carpenter et al. 1998). In attention between individuals and an object (Bruner 1975; other words, the infants no longer merely respond to joint at- Scaife and Bruner 1975) - plays an important role for learning tention; they initiate it as well. Although initiating joint attention and social interaction early in life. Young infants respond to (IJA) and responding to joint attention (RJA) may seem like others’ joint attention bids by following their gaze and pointing two sides of the same coin, evidence suggests that they repre- gestures (Schmitow et al. 2016; Corkum and Moore 1998). sent partially different processes. For example, longitudinal cor- This allows them to experience what others see, and facilitates relations within each type of JA are higher than correlations learning about objects and events in the world. After a while, between the two types of JA at a given time point (Mundy infants start directing others’ attention in order to share their et al. 2007). IJA and RJA also follow different developmental own experiences, first by alternating their gaze between an trajectories. Whereas RJA begins to emerge between 3 and interaction partner and an object, and later by the use of gestures 4 months of age (Gredebäck et al. 2010; D'Entremont et al. 1997; Perra and Gattis 2010), IJA behaviors typically emerge between 8 and 13 months (Beuker et al. 2013). Also at the * Emilia Thorup neural level, there is some evidence that the two types of JA emilia.thorup@psyk.uu.se are supported by partially different systems. It has been sug- gested that RJA relies primarily on a posterior attention system Uppsala Child and Baby Lab, Department of Psychology, Uppsala that regulates involuntary attention, whereas IJA seems to be University, Uppsala, Sweden supported by a later developing anterior network, controlling Karolinska Institutet Center of Neurodevelopmental Disorders volitional attention (Mundy et al. 2000, 2009;Van Heckeetal. (KIND), Pediatric Neuropsychiatry Unit, Department of Women’s& 2007). Further, IJA but not RJA is associated with activity in Children’s Health, Karolinska Institutet, Stockholm, Sweden reward-related brain areas (Schilbach et al. 2010;Gordonet al. BUP Stockholm, Center for Psychiatry Research, Stockholm County 2013) suggesting a link between IJA and motivation. Council, Stockholm, Sweden 1548 J Abnorm Child Psychol (2018) 46:1547–1561 It is well known that children with autism spectrum disor- gestural behaviors such as pointing and showing have long der (ASD) engage less in JA. Parental report of JA behaviors been part of most children’s repertoires. It is thus possible at 8 months has been shown to predict diagnostic status at that these behaviors as well as language had in part re- 7 years (Veness et al. 2014), thus suggesting that JA may be placed alternating gaze as a means of initiating JA. one of the earliest domains where infants with later ASD differ from other infants. Deficits in IJA are considered more Prospective Studies of IJA impairing and tend to persist longer in the development of autistic children than deficits in RJA (Mundy et al. 1994; ASD is rarely diagnosed before the age of 24–36 months, Sigman and Ruskin 1999; Gotham et al. 2007; Lord et al. which implies that assessing IJA in diagnosed children does 2000). The social motivation account (Chevallier et al. not capture possible differences in terms of early develop- 2012) proposes early diminished social motivation as a causal ment. Because 7–20% of younger siblings of children with factor behind the later occurring social-cognitive impairment ASD are expected to receive an ASD diagnosis themselves observed in autistic individuals. If infants who will later be (Messinger et al. 2015; Ozonoff et al. 2011;Gronborget al. diagnosed with ASD find sharing their experiences with 2013), studying younger siblings longitudinally has proven an others less rewarding, they can be expected to initiate JA to important means of gaining insight in the early development a lesser extent than other infants. Considering the important of ASD. role JA plays in social learning and interaction, less initiation Several studies have indicated that receiving an ASD diag- of JA would likely have an attenuating effect on several as- nosis is associated with lower levels or lower growth rates of pects of development. It is therefore imperative that we learn IJA during the second year of life (Macari et al. 2012;Landa more about the early development of IJA in ASD. and Garrett-Mayer 2007; Rozga et al. 2011; Yoder et al. 2009). Also, high risk (HR) infants, i.e. those with an older Measuring Different Types of IJA Behaviors and Their sibling with ASD, have been found to engage in fewer IJA Relation to Each Other and to ASD behaviors than low risk (LR) infants during this time period (Goldberg et al. 2005; Casseletal. 2007, but see Yirmiya et al. Young children may attempt to direct others’ attention by 2006). Only two of these studies (Macari et al. 2012;Landa the use of a variety of behaviors, and studies of IJA typi- and Garrett-Mayer 2007) measured alternating gaze specifi- cally measure gestural behaviors such as pointing and cally. The remaining studies measured IJA as a broader cate- showing as well as eye contact behaviors. Some (e.g. gory, including other communicative behaviors as well. It is Stone et al. 2000; Wetherby and Prizant 2002) include vo- therefore not possible to conclude whether the results were cal behaviors as well. One of the earliest means by which driven by differences in for example gestural behaviors, or infants initiate JA is by looking back and forth between an whether alternating gaze specifically differs between groups. interaction partner and an object or event. Alternating gaze To our best knowledge, only one study has prospectively with the purpose of attention sharing typically emerges investigated IJA before the first birthday (Ibanez et al. 2013). around 8–9 months of age (Beuker et al. 2013), while ges- This study showed that the level of IJA at 8 months predicted tural IJA behaviors such as pointing and showing typically diagnostic outcome (ASD or no diagnosis) at 30 months in a first emerge between 10 and 13 months (Beuker et al. HR sample. As with the majority of the studies mentioned 2013; Carpenter et al. 1998). Focusing on alternating gaze above however, this study did not focus specifically on alter- therefore entails a possibility to detect differences earlier nating gaze, but measured IJA in a broader sense. There is than what is possible when gestural IJA behaviors are mea- however one retrospective study (Clifford and Dissanayake sured. However, some studies have indicated that it is pre- 2009) that has reported a relation between alternating gaze dominantly in terms of the gestural behaviors that the per- as assessed using home videos recorded during the children’s formance of autistic children differs from that of typically first 11 months, and a measure of social responsiveness at developing children (Chiang et al. 2008; Mundy et al. 4years. 1994). In a study of children with ASD, Pickard and One issue that prospective studies can help clarifying is Ingersoll (2015) assessed eye contact based and gestural the timing of the onset of difficulties in alternating gaze in IJA separately. It was discovered that the two types of infants who later develop autism. This question is theoret- IJA were unrelated to each other, and that only gestural ically important because, as pointed out by Jones et al. IJA was associated with language, RJA and imitation. (2014), current research cannot tell us whether atypical This led the authors to suggest that eye contact based IJA alternating gaze could be considered an early risk factor (such as alternating gaze) may be a less optimal measure in itself, or a later consequence of other risk factors. ASD research. It is important to note though, that the chil- Considering that most autistic symptoms emerge during dren in the study by Pickard and Ingersoll (2015)were2– the second year of life (Jones et al. 2014), more studies of alternating gaze before the first birthday are needed. 7 years old, and therefore had reached an age when J Abnorm Child Psychol (2018) 46:1547–1561 1549 Variables that may Influence Alternating Gaze Aims of the Study The majority of the studies assessing IJA have used instru- In the current study, eye tracking technology was used in a live ments where interaction between the child and an experiment- setting, allowing infants’ gaze to be recorded while they were er is coded by a human observer based on a predefined proto- interacting with another (adult) person in an IJA eliciting task. col ((e.g. the Early Social Communication Scales (ESCS; When they reached 18 months, an age where autistic symp- Mundy et al. 1996) or the Screening Tool for Autism in toms are detectable in several domains (Jones et al. 2014), the Two-year-olds (STAT; Stone et al. 2000)). Although this is Autism Diagnostic Observation Schedule-Toddler Module well suited for measuring gestural IJA such as pointing, anal- (ADOS-T; Lord et al. 2012) was administered in order to yses of gaze alternations are likely to benefit from the use of obtain measures of ASD symptom level, as well as gestural more precise methods. Eye tracking has the advantage of ren- IJA. dering exact measures of gaze behavior that allow for various First, we tested the hypothesis that the HR group would analyses of visual attention to be conducted in order to test display less alternating gaze behavior than the LR group at different hypotheses (Falck-Ytter et al. 2013). It can therefore 10 months of age. Second, the relation between alternating be used to generate measures of factors that may potentially gaze at ten months and later ASD symptoms was assessed. have an impact on IJA behaviors. As previously mentioned, Here, the hypothesis was that the extent to which infants en- less IJA engagement could be a consequence of a decreased gaged in alternating gaze at 10 months would be negatively preference for social stimuli in general, which could be associated with ASD symptomatology at 18 months. Next, in assessed via looking times at social and non-social stimuli. It order to explore the potential mechanisms behind alternating has also been shown that infants with later ASD may have gaze and how they may differ between groups, measures of difficulties disengaging their attention from one stimulus to general social preference as well as disengagement ability another (Elsabbagh et al. 2013;Elisonetal. 2013; were included. No a priori hypotheses were formed regarding Zwaigenbaum et al. 2005), which could result in less alternat- these measures. Finally, we wanted to assess the validity of ing gaze. The possibility therefore remains that the lower fre- alternating gaze as a measure of IJA, by investigating its rela- quency of alternating gaze that has been found in infants with tion to later occurring gestural IJA behaviors (showing and later ASD in some studies (Ibanez et al. 2013; Landa and pointing). It was predicted that alternating gaze at 10 months Garrett-Mayer 2007; Macari et al. 2012) is in fact a direct would be positively associated with showing and pointing at consequence of disengagement difficulties. This could also 18 months. potentially explain the lack of a concurrent correlation be- tween different types of IJA detected in autistic preschoolers (Pickard and Ingersoll 2015), since disengagement difficulties Methods would be expected to influence gaze behavior more than ges- tural behaviors. To our knowledge, only one study has inves- Participants tigated the relation between IJA and the aforementioned fac- tors. In a cross-sectional study of three-year-olds with ASD, A total of 67 infants participated in the study (final sample Schietecatte et al. (2012) reported a negative relation between after exclusion; for participant characteristics, see Table 1). disengagement latencies and eye contact based IJA (including Fifty-one infants (29 girls, 22 boys) were HR infants, having alternating gaze), whereas a non-significant trend was found at least one older full sibling with ASD. Sixteen infants (6 between the former and social preference. girls, 10 boys) were LR infants, having at least one older Despite the previously mentioned advantages of eye track- typically developing full sibling and no family history of ing, the most frequently used application of the method has a ASD. Data from an additional 9 infants (6 HR and 3 LR) clear disadvantage in that it uses pre-recorded stimuli (images was collected but excluded due to poor data quality or the or videos). Humans look at others differently when watching infants not contributing enough data (see Analysis section). them on video as compared to when watching them in real life All infants were part of an ongoing longitudinal project (Early (Foulsham et al. 2011; Laidlaw et al. 2011), and it has been Autism Sweden; EASE; www.smasyskon.se)following HR shown that measuring live interaction evokes different re- and LR infants from 5 to 72 months. The HR group was sponses than displaying pre-recorded stimuli, both in terms recruited through advertisements, the project’s webpage and of brain activation (Redcay et al. 2010; Schilbach et al. clinical units. The LR group was recruited from a database of 2010; Shimada and Hiraki 2006) and behavior (Risko et al. families who had indicated interest in participating in research 2012;Freethet al. 2013). Video stimuli can therefore not be with their infants. Both groups consisted primarily of infants considered optimal when the focus is on social behaviors such from the larger Stockholm, Sweden, area. Exclusion criteria as IJA. Instead, eye tracking with live stimuli should be were pre-term birth (< 36 weeks) and confirmed or suspected considered. medical problems, including visual/auditory impairment. 1550 J Abnorm Child Psychol (2018) 46:1547–1561 Table 1 Participant characteristics by group, final samples (M/SD) Measure HR group N = 51 (29 girls) LR group N = 16 (6 girls) Pairwise comparison p-value Age at 10 month assessment 10.45/0.43 10.37/0.55 > 0.25 MSEL Early Learning Composite 10 months 99.92/13.50 101.44/11.64 > 0.25 Age at 18 month assessment 18.51/0.76 18.51/0.94 > 0.25 MSEL Early Learning Composite 18 months 97.29/15.67 97.75/13.54 > 0.25 b a SES - 0.02/0.84 0.08/084 > 0.25 ADOS-T Total CSS 18 months 3.18/1.74 2.06/1.53 0.009 ADOS-T SA CSS 3.24/1.91 1.94/1.73 0.004 ADOS-T RRB CSS 3.98/2.09 3.31/2.12 0.204 ADOS-T Showing 18 months 0.92/0.91 0.44/0.81 0.03 ADOS-T Pointing 18 months 0.71/0.81 0.38/0.50 0.157 Independent samples t-test Socioeconomic status calculated on the basis of parental education and income (equal weighting), expressed as a z-score (for this measure, N =50 inthe HR group and 15 in the LR group since two families did not disclose this information) Mann-Whitney U Test There was no difference between HR and LR groups in terms experimenter moved a squeaky toy across predefined calibra- of socioeconomic status based on family income and educa- tion points, making the toy emit a sound at each point to attract tion level. The developmental level of the infants was assessed the attention of the infant. The procedure was repeated if nec- using the Mullen Scales of Early Learning (MSEL; Mullen essary until calibration was satisfactory (determined via in- 1995) at each visit, and did not differ between groups. The spection of online gaze replay on a monitor in the background diagnosis of the older sibling was confirmed through inspec- of the room). The live eye tracking session comprised a vari- tion of obtained child psychiatric or pediatric records (more ety of different tasks and lasted approximately 10–15 min in than 70% of all assessments included the ADOS (Lord et al. total. Data from the IJA task and the disengagement task (de- 2000) and/or the Autism Diagnostic Interview-Revised (ADI- scribed below) will be reported in the present paper, as well as R; Rutter et al. 2003)). The study was approved by the a measure of general social preference that was obtained dur- Regional Ethical Review Board in Stockholm and conducted ing all other parts of the session. The IJA task always took in accordance with the standards specified in the 1964 place at the end of the session, and the disengagement task Declaration of Helsinki. All parents provided written in- was conducted in the middle of the session. formed consent. Stimuli and Procedure 10-Month Visit The participants took part in a comprehensive assessment, typically spending 4–5 h in the lab. This study includes data from the eye tracking experiment and the MSEL (Mullen 1995). The eye tracking was administered early during the visit. The infant was seated on the lap of the parent, at a distance of 200 cm from the experimenter, who was seated at a low table (see Fig. 1). At each side of the table was an oblong transpar- ent lamp. The experimenter controlled the lamps with two remote controls, hidden beneath the table and thus invisible to the child. A Tobii TX300 eye tracker, placed on a table in Fig. 1 Sketch of the experimental setting. The infant was seated at a front of the infant, was used to record the infant’sgaze witha distance of 200 cm from the experimenter. The infant’s gaze was sample rate of 120 Hz. Two video cameras recorded the be- recorded by a Tobii TX300 eye tracker (placed at the table between the havior of the infant and the stimulus area. Before the session, a infant and experimenter). The infant’s behavioraswell asthe scenearea five point calibration procedure was conducted. The was recorded by two video cameras (not visible in the sketch) J Abnorm Child Psychol (2018) 46:1547–1561 1551 The IJA task consisted of four trials and lasted approxi- autism in children younger than 30 months, with low verbal mately 1–1.5 min in total. The experimenter started each trial skills and a developmental age of at least 12 months. The by attracting the infant’s attention. Once that was achieved the ADOS-T was administered by three experienced clinical psy- experimenter turned on the lamp on his/her right side and chologists. One of the clinicians had achieved formal research lights started to flash, changing color approximately every reliability, and was supervising the work of the other two, who second. The switch controlling the lamp was out of sight for both had extensive clinical experience in using the instrument. the infant, who was not aware that the lights were controlled by the experimenter. While the lights were flashing, the ex- Analysis perimenter sat still looking at the infant, intermittently speak- ing in a quiet tone of voice. This was to provide the infant with Data text files were exported from the Tobii Studio Software. an opportunity to initiate joint attention toward the lamp. If the Analyses were conducted in MATLAB r2015b (MathWorks child made an explicit attempt to direct the experimenter’s Inc., Natick, MA), using the TimeStudio project framework, a attention to the lights, e.g. by pointing or vocalizing (e.g. MATLAB-based open access analysis tool for time series data saying Bthere^ while looking toward the lights), the experi- (version 3.15; timestudioproject.com; Nyström et al. 2016) menter responded by turning toward the lights and and SPSS (SPSS Inc., Chicago, IL). commenting on them. The rationale behind this was to not upset the infant or possibly extinguish IJA behaviors by being IJA Task at 10 Months unresponsive. The lights were flashing for a total of 10 s. The lamp was then turned off, and the second trial was initiated. For each individual recording, four areas of interest (AOIs) All four trials were identical, except that the lights on the were defined from which gaze data was extracted. One cov- experimenter’s right side were activated for the first and third ered the face of the experimenter, two covered the lamps and trials, and the lights on the left side were activated for the the fourth AOI covered the entire scene area. The face AOI second and fourth trials. subtended 11 by 11 visual degrees and the lamp AOIs each In order to check for differences in the ability to disengage subtended 14 by 17 visual degrees. A fixation filter (Tobii from one stimulus to another, a modified version of the disen- Fixation Filter) with a velocity threshold of 35 pixels/ gagement task from the Autism Observation Scale for Infants window and a distance threshold of 35 pixels was applied. (AOSI; Bryson et al. 2008) was included in the eye tracking Because manual inspection revealed that on some trials the session. The experimenter shook a rattle approximately 20 cm lights were activated slightly shorter than 10 s, only data from to the right from his/her face, keeping another rattle hidden the first 9 s of light activation time were analyzed for all trials. underneath the table. Once the infant looked at the first rattle, Only trials with at least 50% gaze data during light activation the second was presented and shaken at the left side of the were included in the analysis. Further, we required at least two experimenter’s face. In total six disengagement trials were valid trials (out of four) from each child. This resulted in an conducted, alternating between the left and right sides. average of 3.20 (SD = 0.87) usable trials in the HR group and The experiment was performed by six individual experi- 3.31 (SD = 0.87) usable trials in the LR group (p >0.25). As menters (2 males, 4 females). In order to minimize the influ- the primary measure of alternating gaze behavior, we used the ence of individual experimenters, each experimenter was mean number of gaze alternations between the face of the trained according to a written protocol and a video template. experimenter and the flashing lights (i.e. the sum of gaze al- The MSEL (Mullen 1995) was administered after the eye ternations from lights to face and face to lights) per trial. In tracking experiment. order to control for general differences in looking patterns and durations, the total number of fixations on the scene area and 18-Month Visit the total looking times in the different AOIs were also com- pared across groups. Similar to the 10-month visit, the children spent a total of 4– 5 h in the lab taking part in a variety of tasks and assessments. Disengagement Task at 10 Months This study includes data from the MSEL and the ADOS-T. The MSEL was administered early during the day, and the Two AOIs covering the two rattles were defined, subtending 6 ADOS-T was typically administered midday after a lunch by 18 visual degrees each. The dependent measure was the time break. The ADOS (Lord et al. 1999) is a widely used diag- it took the infant’s gaze to enter the AOI of the second rattle nostic instrument in autism assessment, and recommended as once it became visible to the infant. All trials in which the first choice instrument by several international clinical guide- infants made a predictive gaze shift (i.e. disengaged from one lines (e.g. NICE.org.uk). Considering the young age of the rattle before the other was presented) were excluded from the participants, the Toddler Module (Lord et al. 2012) was used, analysis. Each infant had to contribute at least two valid disen- which has been developed to assess behaviors indicative of gagement trials in order to be included, which resulted in 57 1552 J Abnorm Child Psychol (2018) 46:1547–1561 infants (45 HR and 12 LR) contributing data to this analysis. scale from 0 to 3 with lower values indicating higher incidence The mean number of usable trials was 3.98 (SD = 1.84) in the and quality of the behaviors. The ADOS-T also contains an HR group and 3.80 (SD = 2.21) in the LR group (p >0.25). item where alternating gaze is scored. Unfortunately, our data did not contain enough variation for analyses regarding this Social Preference at 10 Months item to be conducted. In order to assess possible differences in how much relative Statistical Analysis time the infants spent looking at the experimenter, a social preference quotient was calculated during the entire live eye The ADOS-T variables were not normally distributed and tracking session except the IJA task and the disengagement non-parametric tests (Spearman correlations and multiple lin- task. The eye tracking session builds on naturalistic social ear regression analysis with bootstrap for coefficients (1000 interaction between the experimenter and the infant. It in- samples)) were therefore used to assess relations and group cludes a multitude of components such as the experimenter comparisons involving these variables unless otherwise spec- playing with and looking at a variety of toys; the experimenter ified. All other variables were normally distributed with equal talking to and calling the infant’s name etc. Two tasks are variances across groups, and thus analyzed by the use of para- described elsewhere (Thorup et al. 2016; Nyström et al. metric statistics. For distributions of the experimental vari- 2017). The present objects were at all times kept out of the ables, see Fig. 2. The gender distribution did not differ signif- experimenter’s face area, which allowed us to calculate a ratio icantly between groups, X(1) = 2.23, p = 0.136, and no differ- of relative face looking. The total amount of the time the ences between boys and girls were found on any of the out- infants spent looking at the experimenter’s face was divided come variables, all p-values ≥0.216. Gender was therefore not by the total amount of time they spent looking anywhere on used as a covariate. the scene area, including the experimenter and the various present objects. Higher values would thus indicate a higher preference for the experimenter over the non-social elements. Results Relative looking times at social vs. non-social stimuli have previously been used to evaluate social interest in infants High vs Low Risk Comparisons at 10 Months and toddlers at risk for ASD (e.g. Pierce et al. 2016; Chawarska et al. 2013). To be included in the social preference An independent samples t-test on the mean number of gaze analysis the infants were required to look at the scene (con- alternations between experimenter and lights revealed a group tribute valid gaze data) for at least 25% of the entire session, difference, t(65) = 2.66, p = 0.012, d = 0.77, with the LR which resulted in the inclusion of 65 infants (50 HR, 15 LR). group making more such gaze alternations than the HR group (see Fig. 3). The groups did not differ in terms of the duration ADOS-T Measures at 18 Months of time they spent looking at either the face of the experiment- er, t(65) = 0.84, p >0.25, d = 0.24, or the lights, t(65) = 0.17, Given the young age of the children, we did not classify differ- p >0.25, d = 0.05 (for descriptive statistics for these and the ent outcome groups, but rather investigated autism symptoms following eye tracking measures, see Table 2). The mean quantitatively, operationalized as performance on the ADOS-T. number of fixations on the scene area also did not differ be- As a general measure of ASD symptomatology the Total Social tween groups, t(65) = −0.01, p >0.25, d =0.002. In order to Affect and Restricted Repetitive Behavior Score was used. In check for possible effects of experimenter, we compared the order to test whether alternating gaze may be differently related alternating gaze performance of the children who saw the to the social-communicative domain and the restricted and re- experimenter who tested the largest amount of infants (48%) petitive behavior (RRB) domain,we alsousedtheSocial Affect to the performance of the remaining infants (who each saw 1 (SA) and RRB scores separately. For these three measures, of the 5 other experimenters). No difference was found be- rather than using the raw scores we used the calibrated severity tween the groups, t(65) = − 1.20, p =0.233, d =0.29. scores (CSS), as suggested by Esler et al. (2015). CSS have The results from the disengagement task showed that the been shown to result in more uniform distributions across age groups did not differ in terms of the time it took the infants to and language level and to be less influenced by non-ASD spe- disengage their attention from one active stimulus to another, cific child characteristics compared to raw scores (Esler et al. t(55) = 0.62, p >0.25, d = 0.22. A significant negative corre- 2015; Gotham et al. 2009). CSS range from 1 to 10, with higher lation between disengagement and alternating gaze was found values indicating more symptoms and a higher degree of con- in the LR group, r(10) = − 0.77, p = 0.004, suggesting that cern. In order to investigate the relation between alternating longer latencies to disengage were associated with less alter- gaze at 10 months and later gestural IJA, the items showing nating gaze. In the HR group, a marginally significant trend in and pointing were used. Both items are scored on a four point the same direction was detected, r(43) = − 0.28, p =0.062. J Abnorm Child Psychol (2018) 46:1547–1561 1553 Fig. 2 Histograms showing distributions for the experimental variables group; e) social preference, HR group; f) social preference, LR group. per group. a) alternating gaze, HR group; b) alternating gaze, LR group; Note that none of the distributions differs significantly from normality c) disengagement latency, HR group; d) disengagement latency, LR according to the Kolmogorov-Smirnov test 1554 J Abnorm Child Psychol (2018) 46:1547–1561 alternating gaze alone explained 14% of the variance in ADOS-T scores, p =0.002. In order to test how alternating gaze was related to the social and the RRB domains, the regression approach de- scribed above was repeated twice, using the SA as well as RRB scores as dependent variables. The results of the regres- sion analysis with ADOS-T SA as dependent variable mim- icked those of the original regression model, R = 0.18, F(2,64) = 6.92, p = 0.002, whereas the model with ADOS-T RRB as dependent variable was non-significant, R = 0.02, F(2,64) = 0.61, p >0.25 (Table 3 and Fig. 4). The CSS scores for RRB deviated much more from normality than the other ADOS-T variables (this is a consequence of how the CSS are converted from raw scores in combination with the limited range of scores for this subscale). Therefore, we also conduct- Fig. 3 Number of gaze alternations made between the face of the ed non-parametric correlational analyses of the relation be- experimenter and the lights per second. Error bars represent standard errors. * p <0.05 tween alternating gaze and ADOS-T RRB (Spearman). No associations between alternating gaze and ADOS-T RRB No group difference was found in terms of social prefer- were found in either the HR group, r (49) = − 0.11, p >0.25, ence, t(63) = 0.69, p = > 0.25, d = 0.27 (independent samples or the LR group, r (14) = 0.20, p > 0.25. In contrast, and in t-test). Social preference and alternating gaze did not correlate line with the results from the parametric analysis above, sig- in either the HR group, r(48) = 0.23, p = 0.109, or the LR nificant non-parametric associations between alternating gaze group, r(13) = −0.23, p >0.25. and ADOS-T SA scores were found in both the HR group, r (49) = − 0.29, p = 0.036, and the LR group, r (14) = − 0.63, s s Longitudinal Relations Between 10-Month Measures p = 0.008. Taken together, these analyses indicate that the fre- and 18-Month Symptom Level quency of gaze alternations at 10 months is related to the social affect domain of ASD, but not to the RRB domain. In order to test whether alternating gaze could predict later Consequently, ADOS-T SA rather than total ADOS-T scores symptom severity, a hierarchical multiple regression was con- will be used as dependent measure in the analyses in the fol- ducted. Total ADOS-T score was entered as the dependent lowing section. variable, and as independent variables, alternating gaze and group were added in the first block, and the interaction be- tween group and alternating gaze was added in the second Potential Influences of Disengagement and Social block. Bootstrapping was used to account for the skewed dis- Preference on the Relation Between 10-Month tribution of the ADOS-T scores. The first model was signifi- Alternating Gaze and 18-Month ADOS-T Social Affect cant, R = 0.16, F(2,64) = 6.29, p = 0.003, and showed that Scores alternating gaze at 10 months, but not group status, made a significant contribution to 18 month ADOS-T scores. The Although there were no group differences in terms of disen- interaction term had no significant individual contribution to gagement or social preference at 10 months, it is still possible the ADOS-T scores and was thus excluded from the model that performance on those tasks influences the relation be- (Table 3). With no other factors included in the model, tween alternating gaze and later symptom level. To address Table 2 Descriptive statistics by Measure HR group N = 51 (29 girls) LR group N = 16 (6 girls) group for 10-month eye tracking measures (M/SD) Looking time (s) at model’s face 2.30/1.45 2.65/1.48 Looking time (s) at lights 4.41/1.55 4.49/1.45 N fixations on scene area 13.32/6.05 13.31/4.91 Disengagement latency (s) 0.74/0.25 0.79/0.20 b, c Social preference 0.44/0.12 0.47/0.10 Based on 45 HR and 12 LR infants Based on 50 HR and 15 LR infants Looking time at model divided by total looking time on scene area J Abnorm Child Psychol (2018) 46:1547–1561 1555 Table 3 Regression table displaying the results of analyses with ADOS-T SA scores, we can conclude that neither disengage- alternating gaze and group as independent variables, and Total ADOS-T ment latencies, social preference, nor their interactions with scores, ADOS-T SA scores, and ADOS-T RRB scores as dependent group status influenced the relation between early alternating variables gaze and later symptomatology. See Table 4. Dependent variable B SE B β Total ADOS-T score Constant 3.46 0.67 Longitudinal Relations Between 10-Month Alternating gaze −0.35 0.14 −0.31* Alternating Gaze and 18-Month Gestural IJA Group 0.71 0.49 0.18 ADOS-T SA score Constant 3.55 0.74 In order to investigate whether early alternating gaze was re- Alternating gaze −0.41 0.15 −0.33* lated to later gestural IJA, two bootstrapped hierarchical re- Group 0.84 0.54 0.19 gression analyses with the ADOS-T items showing and ADOS-T RRB score Constant 3.25 0.88 pointing as dependent variables were conducted. As before, alternating gaze and group were entered as independent vari- Alternating gaze 0.02 0.18 0.01 Group 0.69 0.64 0.14 ables in the first block, and the interaction term was entered in the second block. Both models were significant in the first step In neither of the regression models was the interaction between alternat- (for showing: R = 0.17, F(2,64) = 6.53, p =0.003; for ing gaze and group significant (β =0.07, p >0.25 with total ADOS-T as pointing: R = 0.13, F(2,64) = 4.83, p = 0.011) and as none dependent variable; β =0.13, p > 0.25 with ADOS-T SA as dependent of the interaction terms were significant, they were removed variable; β = − 0.29, p > 0.25 with ADOS-T RRB as dependent variable); *= p <0.005 from the models (Table 5). In both analyses, alternating gaze was the only significant predictor. Following up the regression this issue, another regression analysis was thus performed, models with zero-order correlations revealed significant neg- including only those infants who had contributed data to both ative correlations between alternating gaze and the ADOS-T the disengagement task and the social preference task (44 HR item showing in both the HR group, r (49) = −0.33, p =0.018, infants and 11 LR infants). Again, a bootstrapped hierarchical and the LR group, r (14) = −0.50, p = 0.047. In other words, multiple regression with ADOS-T SA score as dependent var- infants who engaged more in alternating gaze at 10 months iable was conducted, this time entering alternating gaze, dis- also engaged more in showing at 18 months. For pointing engagement, social preference and group in the first block, however, only a trend toward a negative correlation with al- and the three interaction terms between those variables and ternating gaze was found in the HR-group, r (49) = −0.25, p = group in the second block. The initial model was significant, 0.082, indicating that in this group, making fewer gaze alter- R =0.17, F(4,50) = 2.62, p = 0.046. None of the interaction nations at 10 months was associated with less pointing at terms contributed significantly to the ADOS-T SA scores and 18 months. In the LR group, the relation was similar in mag- they were therefore removed from the model. As alternating nitude and direction, but did not reach statistical significance, gaze remained the only variable that uniquely predicted later r (14) = − 0.32, p =0.223. Supplementary Analyses Finally, in order to test the robustness of the results, all anal- yses were repeated excluding individuals who could be Table 4 Regression table with alternating gaze, disengagement and social preference as independent variables and ADOS-T SA score as dependent variable BSEB β Constant 4.31 1.66 Alternating gaze −0.51 0.19 −0.41* Disengagement latency −1.92 1.20 −0.23 Social preference 2.54 2.42 0.15 Group 0.50 0.67 0.11 Fig. 4 Alternating gaze plotted against ADOS-T SA calibrated severity None of the interaction terms were significant (alternating gaze*group: β = − 0.26, p > 0.25; disengagement latency*group: β = − 1.28, p = scores. Figure shows individual data points and separate regression lines 0.215; social preference*group: β =0.96, p >0.25); * p =0.01 for each group 1556 J Abnorm Child Psychol (2018) 46:1547–1561 Table 5 Regression table with ADOS-T items showing and pointing as Discussion dependent variables This study assessed infants’ alternating gaze behavior between Dependent variable B SE B β an adult and an event occurring outside the adult’s field of vision Showing Constant 1.27 0.35 as a potential measure of early IJA. As hypothesized, we found Alternating gaze −0.21 0.07 −0.36* that 10-month-old infants at high familial risk for ASD made Group 0.25 0.25 0.12 fewer gaze alternations than low risk controls. Moreover, a lower Pointing Constant 1.00 0.30 frequency of gaze alternations at 10 months was associated with Alternating gaze −0.16 0.06 −0.33* more core ASD symptoms at 18 months. Thus, alternating gaze Group 0.15 0.22 0.09 behavior can discriminate LR controls from infants at familial risk for ASD, as well as be used to predict later autism symptoms In neither of the regression models was the interaction between alternat- dimensionally. When investigating the relations between alternat- ing gaze and group significant (β =0.28, p > 0.25 with showing as de- ing gaze and the SA and RRB domains of the ADOS-T sepa- pendent variable; β = − 0.20, p > 0.25 with pointing as dependent vari- able); * = p <0.05 rately, they remained significant in both groups in the former case, whereas no relation was detected with the second measure. suspected to have influenced the results due to one of two sets This suggests that alternating gaze is related to the social symp- of special circumstances, namely explicit initiation of JA dur- toms of ASD, but not to restricted and repetitive behaviors. ing the IJA task (pointing toward or commenting on the lights) The majority of the previous studies of early IJA have used or atypically high ADOS-T scores (LR infants only). more naturalistic methods or clinical instruments. Although these In total 5 HR and 1 LR infants made one or more explicit approaches are associated with several advantages - such as high attempts at directing the experimenter’s attention to the ecological validity - they also have some limitations, one being lights by pointing toward them or commenting on them. that it is difficult to control for factors other than IJA that may As noted, in these cases, the experimenter responded by explain the differences one observes. Taking advantage of precise turning toward the lights and verbally acknowledging them. eye tracking technology, one of the aims of the current study was Since this opens up for the possibility that the alternating to investigate the influence of two such potential variables, name- gaze behavior of these infants was affected by the change in ly attention disengagement and general social interest. Problems experimenter behavior, all analyses were repeated exclud- with disengagement of attention have previously been found in ing the 6 infants who pointed or vocalized. All results children with or at risk for ASD and could result in fewer gaze remained the same as before, except for that the group fac- alternations. Latencies to disengage from one stimulus to another tor now contributed significant variance to the total ADOS- were therefore measured. The performance of the groups did not Tscores, β =0.23, p = 0.011,aswellastothe ADOS-T SA differ on this task, and disengagement difficulties can thus be scores, β = 0.24, p = 0.019 (as in the previous analysis, sig- ruled out as a cause behind the fewer gaze alternations of the nificance levels using bootstrapping are reported). HR group. Longer disengagement latencies were however con- All data was checked for outliers, and investigating Cook’s currently associated with less alternating gaze in both groups at distances confirmed that there were no bivariate outliers in any 10 months. It is not surprising that infants who take longer to of the analyses. Considering the ADOS scores however, 2 disengage their attention display fewer gaze alternations, as mak- children in the LR group had (total and SA) scores about 2 ing frequent gaze shifts is dependent on fast disengagement la- SD (but <3 SD) above their group’s M. The ADOS is devel- tencies. Importantly though, disengagement latencies were unre- oped for clinical use, and some variation is expected when lated to later ASD symptoms in the current study. This may be used in a typical sample. Considering that the 2 LR infants surprising considering that previous studies have reported longer had MSEL scores in the typical range, we do not consider disengagement latencies in infants with a later ASD diagnosis these children outliers. Nevertheless, in order to control for (Elison et al. 2013; Elsabbagh et al. 2013; Zwaigenbaum et al. their possible influence on the results, all analyses were re- 2005; Kleberg et al. 2016). This discrepancy may be explained peated removing these 2 children. All significant results by differences in age between the studies, as none of the other remained, but just like in the analysis reported above, group studies measured disengagement at 10 months, but the relation was now a significant predictor of total ADOS-T scores, β = between visual disengagement and ASD clearly warrants further 0.29, p=0.001,aswellasADOS-TSA scores, β =0.31, p = investigation. In terms of social preference, we did not detect 0.002. either a group difference or any relations with alternating gaze Excluding infants either because they pointed/vocalized or ASD symptoms. A higher proportion of time spent looking at during the IJA task, or because they were LR infants with the experimenter (social preference) and more gaze shifts be- atypically high ADOS-T scores, thus enhanced the predic- tween experimenter and lights could both reflect a higher social tive power of the group factor, but did not change the interest or motivation. From that perspective, it may be surprising results otherwise. that our social preference measure was unrelated to both J Abnorm Child Psychol (2018) 46:1547–1561 1557 alternating gaze and later ASD symptoms. The reason for this phenomenon, less engaging in alternating gaze may well impact may be that alternating gaze (and IJA), reflects a more active later development negatively. When the infant initiates joint at- attempt toward sharing experiences with another than does pas- tention with an adult, the adult is likely to respond by providing sively looking at social stimuli. IJA by means of gaze alternation information or social interaction. Alternating gaze therefore ini- has previously been associated with increased activity in the tiates communication and interaction as well as provides oppor- ventral striatum, an area that has been linked to motivation and tunities for learning. Less engagement in IJA early in life could reward seeking, in a sample of TD adults (Schilbach et al. 2010). therefore be expected to have extensive consequences on several It is thus possible that infants who later display more ASD symp- areas of development. It has recently been suggested that patterns toms find attention sharing less rewarding than infants with lower of social looking, which appear to have a substantial genetic degrees of autistic traits, whereas the same is not necessarily the component in young children (Constantino et al. 2017), may case for more passive looking patterns. contribute to so-called evocative gene-environment correlations The study also aimed to clarify the extent to which early whereby the child actively shapes its own social environment alternating gaze is specifically related to established IJA behav- (Kennedy et al. 2017). Measuring alternating gaze is fast (the iors later on. Looking back and forth between a person and an current results are based on an experiment that only lasted 1– object is indeed a less explicit manifestation of joint attention 1.5 min in total) and does not necessarily require eye tracking compared to the later occurring gestural behaviors, such as equipment to be conducted. If the findings are replicated and the pointing and showing. As noted in the introduction, a previous task further improved, the alternating gaze measure could thus study demonstrated that eye contact based IJA, such as alternat- find practical use in the future. Apart from highlighting the role of ing gaze, was concurrently unrelated to gestural IJA in a sample alternating gaze in the early development of infants at risk for of autistic children (Pickard and Ingersoll 2015). In the current ASD, this study also contributes information about IJA in typical study, a higher frequency of alternating gaze at 10 months was development. Most studies on IJA and social functioning have significantly associated with more showing and marginally asso- focused on atypical populations, which entails that our ciated with more pointing at 18 months, suggesting that alternat- knowledge about their relation in typical development is rather ing gaze is longitudinally related to gestural IJA behaviors. As scarce. In a rare study of typical development, Van Hecke et al. suggested in the introduction, the lack of a correlation in the (2007) showed that a higher incidence of IJA at 12 months was studybyPickardandIngersoll (2015) may be related to the related to more optimal social functioning at 30 months. Their age of their participants. The children in that study were 2–7 years and our findings thus both indicate that the relation between old, and language may to some extent have replaced gestural and alternating gaze and later social functioning is not restricted to eye contact based IJA, thus weakening the relation between the clinical samples. This is also compatible with the view that the two. In the present study we measured both alternating gaze and autistic phenotype represents the extreme end of a continuum, rather than a discrete entity (Lundstrom et al. 2012 gestural IJA at ages where language is not yet an effective means ). of initiating JA for most children. The fact that a longitudinal Although the current study is one of the first to use eye relation between alternating gaze and gestural IJA was found in tracking to measure IJA, the technology has previously been the current study strengthens the validity of alternating gaze as an used for the same purpose by Billeci et al. (2016). In their early IJA measure. The measure further has a clear advantage study, 2-year-olds with and without ASD watched a video of compared to gestural measures in that it allows for earlier use. a model seated at a table where a toy car suddenly started to Pointing, for example, develops quite late with respect to when move. In contrast to the current results, it was discovered that investigations of IJA are of interest. Although pointing was not the ASD group made more gaze alternations between the the focus of the current study, all instances when infants pointed model and the car than the control group. When comparing toward the lights were coded. In total, only three infants pointed the two studies some caution must be given considering that during the experiment, rendering inferential statistical testing Billeci et al. (2016) focused on diagnosed children, whereas non-applicable. This confirms that at 10 months, pointing is not the current study assessed younger siblings. Nevertheless, the a useful measure of individual differences in IJA, as most infants discrepancy in results is quite striking, and we can think of at did not display the behavior at all. With the alternating gaze least two possible methodological explanations for this. First, measure, on the other hand, we found substantial individual dif- the car used by Billeci et al. (2016) was moving from the ferences at the same age. periphery toward the center of the screen. It is possible that A group difference in alternating gaze was detected already at the TD children interpreted this as the car entering the model’s 10 months, which is very shortly after the first signs of this field of vision, thus rendering alternating gaze in the function behavior typically emerges (Beuker et al. 2013). This suggests of drawing the model’s attention to the car superfluous. In the that the patterns of IJA behaviors in children with later ASD current study, the flashing lights were outside the model’sfield symptoms may be different from the time these behaviors start of vision at all times. Second, Billeci et al. (2016) used pre- to develop. Although the current results do not allow us to deter- recorded stimuli whereas the current study was conducted live. mine whether reduced alternating gaze is a primary or secondary The 2-year-old participants in the study by Billeci et al. (2016) 1558 J Abnorm Child Psychol (2018) 46:1547–1561 could be expected to have an appreciation for the difference symptoms at 18 months, and that this relation is independent of between a real person and one appearing on a screen. It there- differences in general social interest and disengagement abilities. fore remains a possibility that the TD participants’ low inci- The results suggest that a lower tendency to engage in alternating dence of alternating gaze was a consequence of their awareness gaze may be one of the earliest signs of ASD in the social do- that the person on screen was not affected by their behavior. main. Because these early atypicalities may negatively impact The current study highlights the potential as well as the impor- patterns of social interaction with caregivers, their precise role tance of using a more naturalistic approach when studying in shaping the development of children with ASD should be social behaviors (Falck-Ytter 2015; Falck-Ytter et al. 2015). addressed in future studies. By applying eye tracking technology in a live setting we were able to combine the advantages of more highly controlled par- Acknowledgements We would like to thank all the families participating in adigms with those of more ecologically valid ones. this study, and the members of Uppsala Child and Baby lab for valuable The current study has several limitations that deserve atten- comments on an earlier version of this paper. This research was supported tion. ADOS-T scores were used as an indication of the degree by the Swedish Research Council in partnership with FAS, FORMAS, and of ASD symptomatology. However, although the ADOS is VINNOVA (a cross-disciplinary research program concerning the mental health of children and young people; grant number 259-2012-24). The work considered gold standard in ASD assessments, it is not a stand of ET, PN, TFY, and SB was supported by the ESF COST Action BM1004 alone instrument that in itself gives a perfect indication of a BEnhancing the Scientific Study of Early Autism^ (ESSEA). TFY and SB child’s symptom level. This means, for example, that a high were supported by the EU-MSCA Initial Training Network (Brainview). ADOS-T score does not necessarily mean that diagnostic TFY was supported by The Bank of Sweden Tercentenary Foundation (P12-0270:1; NHS14-1802:1), The Swedish Research Council (2015– criteria are fulfilled. Thus the current approach should not be 03670), and the Niclas Öberg Foundation. SB was supported by the confused with one where diagnostic outcome groups are com- Swedish Research Council (No. 523-2009-7054). GG was supported by the pared, which would require a formal diagnostic assessment. European Research Council (StG: CACTUS 312292). The EASE Team The fact that the two groups differed substantially in size is consisted of Linn Andersson Konke, Maria Axnér, Karin Brocki, Elodie Cauvet, Therese L. Ekberg, Johan Lundin Kleberg, Elisabeth Nilsson Jobs, also a limitation, and even more so is the fact that the LR Sheila Norin and Eric Zander. group consisted of only 16 children (with fewer contributing data to some of the analyses). The conclusions regarding re- lations within the LR group therefore need to be interpreted Compliance with Ethical Standards with some caution, as well as the conclusion that the relation between alternating gaze and our 18-month measures are sim- Conflict of Interest The authors declare that they have no conflict of ilar in both groups. The t-tests that were conducted on the interest to declare. primary and supplementary measures however take the differ- Ethical Approval All procedures performed in studies involving human ences in sample sizes into account. Also, it is relatively com- participants were in accordance with the ethical standards of the institu- mon to include more HR than LR infants in longitudinal stud- tional and/or national research committee and with the 1964 Helsinki ies (Zwaigenbaum et al. 2005), partially because the HR sam- declaration and its later amendments or comparable ethical standards. ple will be split into multiple groups after diagnostic assess- ment. Importantly, the relation between alternating gaze at Informed Consent Informed consent was obtained from all individual participants included in the study. 10 months and the total ADOS-T score (explaining 14% of the variance) is based on a reasonable sample size due to Open Access This article is distributed under the terms of the Creative pooling of the groups. Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appro- Conclusions priate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Alternating gaze is one of the first means a young infant has to express and share his/her interest in various objects and events with another person. Looking back and forth between a parent References and an object may result in opportunities for social interaction (the parent acknowledges the infant’s behavior by attending to Beuker, K. T., Rommelse, N. N. J., Donders, R., & Buitelaar, J. K. (2013). him/her), word learning (the parent labels the object), as well as Development of early communication skills in the first two years of motoric exploration (the parent hands the infant the object). life. Infant Behavior & Development, 36(1), 71–83. https://doi.org/ 10.1016/j.infbeh.2012.11.001. Alternating gaze is thus a powerful behavior that affects later Billeci, L., Narzisi, A., Campatelli, G., Crifaci, G., Calderoni, S., development in multiple ways. The results of this study indicate Gagliano, A., et al. (2016). Disentangling the initiation from the not only that infants at risk for ASD engage less in alternating response in joint attention: an eye-tracking study in toddlers with gaze than low risk infants, but also that the tendency to engage in autism spectrum disorders. Translational Psychiatry, 6. https://doi. alternating gaze at 10 months is negatively related to social ASD org/10.1038/tp.2016.75. J Abnorm Child Psychol (2018) 46:1547–1561 1559 Bruner, J. S. (1975). From communication to language - psychological Falck-Ytter, T., Carlström, C., & Johansson, M. (2015). Eye contact modulates cognitive processing differently in children with autism. perspective. Cognition, 3(3), 255–287. https://doi.org/10.1016/ 0010-0277(74)90012-2. Child Development, 86(1), 37–47. https://doi.org/10.1111/cdev. Bryson,S.E.,Zwaigenbaum, L.,McDermott,C.,Rombough,V.,&Brian,J. (2008). The autism observation scale for infants: scale development and Foulsham, T., Walker, E., & Kingstone, A. (2011). The where, what and reliability data. Journal of Autism and Developmental Disorders, 38(4), when of gaze allocation in the lab and the natural environment. 731–738. https://doi.org/10.1007/s10803-007-0440-y. Vision Research, 51(17), 1920–1931. https://doi.org/10.1016/j. Carpenter, M., Nagell, K., & Tomasello, M. (1998). Social cognition, visres.2011.07.002. joint attention, and communicative competence from 9 to 15 months Freeth, M., Foulsham, T., & Kingstone, A. (2013). What affects social of age. [Review]. Monographs of the Society for Research in Child attention? Social presence, eye contact and autistic traits. PLoS One, Development, 63(4), V-143. 8(1). doi:https://doi.org/10.1371/journal.pone.0053286. Cassel, T. D., Messinger, D. S., Ibanez, L. V., Haltigan, J. D., Acosta, S. I., Goldberg, W. A., Jarvis, K. L., Osann, K., Laulhere, T. M., Straub, C., & Buchman, A. C. (2007). Early social and emotional communica- Thomas, E., et al. (2005). Brief report: early social communication tion in the infant siblings of children with autism spectrum disorders: behaviors in the younger siblings of children with autism. Journal of an examination of the broad phenotype. Journal of Autism and Autism and Developmental Disorders, 35(5), 657–664. https://doi. Developmental Disorders, 37(1), 122–132. https://doi.org/10.1007/ org/10.1007/s10803-005-0009-6. s10803-006-0337-1. Gordon, I., Eilbott, J. A., Feldman, R., Pelphrey, K. A., & Vander Wyk, B. Chawarska, K., Macari, S., & Shic, F. (2013). Decreased spontaneous C. (2013). Social, reward, and attention brain networks are involved attention to social scenes in 6-month-old infants later diagnosed with when online bids for joint attention are met with congruent versus autism spectrum disorders. Biological Psychiatry, 74(3), 195–203. incongruent responses. Social Neuroscience, 8(6), 544–554. https:// https://doi.org/10.1016/j.biopsych.2012.11.022. doi.org/10.1080/17470919.2013.832374. Chevallier, C., Kohls, G., Troiani, V., Brodkin, E. S., & Schultz, R. T. Gotham, K., Risi, S., Pickles, A., & Lord, C. (2007). The autism diag- (2012). The social motivation theory of autism. Trends in Cognitive nostic observation schedule: revised algorithms for improved diag- Sciences, 16(4), 231–238. https://doi.org/10.1016/j.tics.2012.02. nostic validity. Journal of Autism and Developmental Disorders, 37(4), 613–627. https://doi.org/10.1007/s10803-006-0280-1. Chiang, C.-H., Soong, W.-T., Lin, T.-L., & Rogers, S. J. (2008). Gotham, K., Pickles, A., & Lord, C. (2009). Standardizing ADOS scores Nonverbal communication skills in young children with autism. for a measure of severity in autism spectrum disorders. Journal of Journal of Autism and Developmental Disorders, 38(10), 1898– Autism and Developmental Disorders, 39(5), 693–705. https://doi. 1906. https://doi.org/10.1007/s10803-008-0586-2. org/10.1007/s10803-008-0674-3. Clifford, S., & Dissanayake, C. (2009). Dyadic and triadic behaviours in Gredebäck, G., Fikke, L., & Melinder, A. (2010). The development of infancy as precursors to later social responsiveness in young chil- joint visual attention: a longitudinal study of gaze following during dren with autistic disorder. Journal of Autism and Developmental interactions with mothers and strangers. Developmental Science, Disorders, 39(10), 1369–1380. https://doi.org/10.1007/s10803-009- 13(6), 839–848. https://doi.org/10.1111/j.1467-7687.2009.00945.x. 0748-x. Gronborg, T. K., Schendel, D. E., & Parner, E. T. (2013). Recurrence of Constantino, J. N., Kennon-McGill, S., Weichselbaum, C., Marrus, N., autism spectrum disorders in full- and half-siblings and trends over Haider, A., Glowinski, A. L., et al. (2017). Infant viewing of social time a population-based cohort study. JAMA Pediatrics, 167(10), scenes is under genetic control and is atypical in autism. Nature, 947–953. https://doi.org/10.1001/jamapediatrics.2013.2259. 547(7663), 340−+, doi:https://doi.org/10.1038/nature22999. Ibanez, L. V., Grantz, C. J., & Messinger, D. S. (2013). The development Corkum, V., & Moore, C. (1998). The origins of joint visual attention in of referential communication and autism symptomatology in high- infants. Developmental Psychology, 34(1), 28–38. https://doi.org/ risk infants. Infancy, 18(5), 687–707. https://doi.org/10.1111/j.1532- 10.1037/0012-1649.34.1.28. 7078.2012.00142.x. D'Entremont, B., Hains, S. M. J., & Muir, D. W. (1997). A demonstration Jones, E. J. H., Gliga, T., Bedford, R., Charman, T., & Johnson, M. H. of gaze following in 3- to 6-month-olds. Infant Behavior & (2014). Developmental pathways to autism: a review of prospective Development, 20(4), 569–572. https://doi.org/10.1016/s0163- studies of infants at risk. Neuroscience and Biobehavioral Reviews, 6383(97)90048-5. 39,1–33. https://doi.org/10.1016/j.neubiorev.2013.12.001. Elison, J. T., Paterson, S. J., Wolff, J. J., Reznick, J. S., Sasson, N. J., Gu, Kennedy, D. P., D’Onofrio, B. M., Quinn, P. D., Bölte, S., Lichtenstein, H., et al. (2013). White matter microstructure and atypical visual P., & Falck-Ytter, T. (2017). Genetic influence on eye movements to or ienting in 7-month-olds at risk for autism. American Journal of complex scenes at short timescales. Current Biology. https://doi.org/ Psychiatry, 170(8), 899–908. https://doi.org/10.1176/appi.ajp.2012. 10.1016/j.cub.2017.10.007. Kleberg, J. L., Thorup, E., & Falck-Ytter, T. (2016). Reduced visual Elsabbagh, M., Fernandes, J., Webb, S. J., Dawson, G., Charman, T., disengagement but intact phasic alerting in young children with Johnson, M. H., et al. (2013). Disengagement of visual attention in autism. [Article in press]. Autism Research. https://doi.org/10. infancy is associated with emerging autism in toddlerhood. 1002/aur.1675. Biological Psychiatry, 74(3), 189–194. https://doi.org/10.1016/j. Laidlaw, K. E. W., Foulsham, T., Kuhn, G., & Kingstone, A. (2011). biopsych.2012.11.030. Potential social interactions are important to social attention. Esler, A. N., Bal, V. H., Guthrie, W., Wetherby, A., Weismer, S. E., & [Article]. Proceedings of the National Academy of Sciences of the Lord, C. (2015). The autism diagnostic observation schedule, tod- United States of America, 108(14), 5548–5553. https://doi.org/10. dler module: standardized severity scores. JournalofAutismand 1073/pnas.1017022108. Developmental Disorders, 45(9), 2704–2720. https://doi.org/10. 1007/s10803-015-2432-7. Landa, R. J., Holman, K. C., & Garrett-Mayer, E. (2007). Social and Falck-Ytter, T. (2015). Gaze performance during face-to-face communi- communication development in toddlers with early and later diag- cation: a live eye tracking study of typical children and children with nosis of autism spectrum disorders. Archives of General Psychiatry, autism. Research in Autism Spectrum Disorders, 17,78–85. 64(7), 853–864. Falck-Ytter, T., Bölte, S., & Gredebäck, G. (2013). Eye tracking in early Lord, C., Rutter, M., DiLavore, P., & Risi, S. (1999). ADOS. Autism diagnostic observation schedule. Los Angeles: Western autism research. [Review]. Journal of Neurodevelopmental Disorders, 5,13. https://doi.org/10.1186/1866-1955-5-28. Psychological Services. 1560 J Abnorm Child Psychol (2018) 46:1547–1561 Lord, C., Risi, S., Lambrecht, L., Cook Jr., E. H., Leventhal, B. L., Biological Psychiatry, 79(8), 657–666. https://doi.org/10.1016/j. DiLavore, P. C., et al. (2000). The autism diagnostic observation biopsych.2015.03.032. schedule-generic: a standard measure of social and communication Redcay, E., Dodell-Feder, D., Pearrow, M. J., Mavros, P. L., Kleiner, M., deficits associated with the spectrum of autism. Journal of Autism Gabrieli, J. D. E., et al. (2010). Live face-to-face interaction during and Developmental Disorders, 30(3), 205–223. https://doi.org/10. fMRI: a new tool for social cognitive neuroscience. NeuroImage, 1023/a:1005592401947. 50(4), 1639–1647. https://doi.org/10.1016/j.neuroimage.2010.01. Lord, C., Luyster, R., Gotham, K., & Guthrie, W. (2012). Autism diag- Risko, E. F., Laidlaw, K. E. W., Freeth, M., Foulsham, T., & Kingstone, nostic observation schedule - toddler module manual. Los Angeles: A. (2012). Social attention with real versus reel stimuli: toward an Western Psychological Services. empirical approach to concerns about ecological validity. Frontiers Lundstrom, S., Chang, Z., Rastam, M., Gillberg, C., Larsson, H., in Human Neuroscience,1–11. https://doi.org/10.3389/fnhum.2012. Anckarsater, H., et al. (2012). Autism spectrum disorders and autisticlike traits similar etiology in the extreme end and the normal Rozga, A., Hutman, T., Young, G. S., Rogers, S. J., Ozonoff, S., Dapretto, variation. Archives of General Psychiatry, 69(1), 46–52. M., et al. (2011). Behavioral profiles of affected and unaffected Macari, S. L., Campbell, D., Gengoux, G. W., Saulnier, C. A., Klin, A. J., siblings of children with autism: contribution of measures of & Chawarska, K. (2012). Predicting developmental status from 12 mother-infant interaction and nonverbal communication. Journal to 24 months in infants at risk for autism spectrum disorder: a pre- of Autism and Developmental Disorders, 41(3), 287–301. liminary report. Journal of Autism and Developmental Disorders, Rutter, M., LeCouteur, A., & Lord, C. (2003). Autism diagnostic inter- 42(12), 2636–2647. view revised (ADI-R). Los Angeles: Western Psychological Messinger, D. S., Young, G. S., Webb, S. J., Ozonoff, S., Bryson, S. E., Services. Carter, A., et al. (2015). Early sex differences are not autism-specif- Scaife, M., & Bruner, J. S. (1975). Capacity for joint visual attention in ic: a baby siblings research consortium (BSRC) study. Molecular infant. Nature, 253(5489), 265–266. https://doi.org/10.1038/ Autism. https://doi.org/10.1186/s13229-015-0027-y. 253265a0. Mullen, E. M. (1995). Mullen scales of early learning (AGS ed.). Circle Schietecatte, I., Roeyers, H., & Warreyn, P. (2012). Exploring the nature Pines: American Guidance Service Inc.. of joint attention impairments in young children with autism spec- Mundy, P., Sigman, M., & Kasari, C. (1994). Joint attention, develop- trum disorder: associated social and cognitive skills. Journal of mental level, and symptom presentation in autism. Development and Autism and Developmental Disorders, 42(1), 1–12. Psychopathology, 6(3), 389–401. https://doi.org/10.1017/ Schilbach, L., Wilms, M., Eickhoff, S. B., Romanzetti, S., Tepest, R., s0954579400006003. Bente, G., et al. (2010). Minds made for sharing: initiating joint Mundy, P., Hogan, A., & Doehring, P. (1996). A preliminary manual for attention recruits reward-related neurocircuitry. Journal of the abriged early social communication scales (ESCS).Coral Cognitive Neuroscience, 22(12), 2702–2715. https://doi.org/10. Gables: University of Miami. 1162/jocn.2009.21401. Mundy, P., Card, J., & Fox, N. (2000). EEG correlates of the development Schmitow, C., Kochukhova, O., & Nystrom, P. (2016). Social perception: of infant joint attention skills. Developmental Psychobiology, 36(4), how do 6-month-old infants look at pointing gestures? Infant 325–338. Behavior & Development, 42,1 52–156. https://doi.org/10.1016/j. Mundy, P., Block, J., Delgado, C., Pomares, Y., Van Hecke, A. V., & infbeh.2015.12.001. Parlade, M. V. (2007). Individual differences and the development Shimada, S., & Hiraki, K. (2006). Infant’s brain responses to live and of joint attention in infancy. Child Development, 78(3), 938–954. televised action. NeuroImage, 32(2), 930–939. https://doi.org/10. https://doi.org/10.1111/j.1467-8624.2007.01042.x. 1016/j.neuroimage.2006.03.044. Mundy, P., Sullivan, L., & Mastergeorge, A. M. (2009). A parallel and Sigman, M., & Ruskin, E. (1999). Continuity and change in the social distributed-processing model of joint attention, social cognition and competence of children with autism, down syndrome, and develop- autism. Autism Research, 2(1), 2–21. https://doi.org/10.1002/aur.61. mental delays. Monographs of the Society for Research in Child Nyström, P., Bölte, S., Falck-Ytter, T., & Ease Team. (2017). Responding Development, 64(1), 1−+, doi:https://doi.org/10.1111/1540-5834. to other people’s direct gaze: alterations in gaze behavior in infants at risk for autism occur on very short timescales. Journal of Autism Stone, W. L., Coonrod, E. E., & Ousley, O. Y. (2000). Brief report: and Developmental Disorders, 47(11), 3498–3509. screening tool for autism in two-year-olds (STAT): development Nyström, P., Falck-Ytter, T., & Gredeback, G. (2016). The TimeStudio and preliminary data. Journal of Autism and Developmental project: an open source scientific workflow system for the behav- Disorders, 30(6), 607–612. https://doi.org/10.1023/a: ioral and brain sciences. Behavior Research Methods, 48(2), 542– 552. https://doi.org/10.3758/s13428-015-0616-x. Thorup, E., Nystrom, P., Gredeback, G., Bolte, S., Falck-Ytter, T., & Ozonoff, S., Young, G. S., Carter, A., Messinger, D., Yirmiya, N., Team, E. (2016). Altered gaze following during live interaction in Zwaigenbaum, L., et al. (2011). Recurrence risk for autism spectrum infants at risk for autism: an eye tracking study. Molecular Autism, 7. disorders: a baby siblings research consortium study. Pediatrics, doi:https://doi.org/10.1186/s13229-016-0069-9. 128(3), e488–e495. Van Hecke, A. V., Mundy, P. C., Acra, C. F., Block, J. J., Delgado, C. E. Perra, O., & Gattis, M. (2010). The control of social attention from 1 to 4 F., Parlade, M. V., et al. (2007). Infant joint attention, temperament, months. British Journal of Developmental Psychology, 28(4), 891– and social competence in preschool children. Child Development, 908. https://doi.org/10.1348/026151010x487014. 78(1), 53–69. Pickard, K. E., & Ingersoll, B. R. (2015). Brief report: high and low level Veness, C., Prior, M., Eadie, P., Bavin, E., & Reilly, S. (2014). Predicting initiations of joint attention, and response to joint attention: differ- autism diagnosis by 7 years of age using parent report of infant ential relationships with language and imitation. Journal of Autism social communication skills. Journal of Paediatrics and Child and Developmental Disorders, 45(1), 262–268. https://doi.org/10. Health, 50(9), 693–700. https://doi.org/10.1111/jpc.12614. 1007/s10803-014-2193-8. Wetherby, A. M., & Prizant, B. M. (2002). CSBS DP manual: Communication and symbolic behavior scales developmental Pierce, K., Marinero, S., Hazin, R., McKenna, B., Barnes, C. C., & profile. Baltimore: Paul H Brookes Publishing Co.. Malige, A. (2016). Eye tracking reveals abnormal visual preference Yirmiya, N., Gamliel, I., Pilowsky, T., Feldman, R., Baron-Cohen, S., & for geometric images as an early biomarker of an autism spectrum Sigman, M. (2006). The development of siblings of children with disorder subtype associated with increased symptom severity. J Abnorm Child Psychol (2018) 46:1547–1561 1561 autism at 4 and 14 months: social engagement, communication, and Disorders, 39(10), 1381–1391. https://doi.org/10.1007/s10803- 009-0753-0. cognition. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 47(5), 511–523. https://doi.org/10.1111/j.1469-7610. Zwaigenbaum, L., Bryson, S., Rogers, T., Roberts, W., Brian, J., & 2005.01528.x. Szatmari, P. (2005). Behavioral manifestations of autism in the first Yoder, P., Stone, W. L., Walden, T., & Malesa, E. (2009). Predicting social year of life. International Journal of Developmental Neuroscience, impairment and ASD diagnosis in younger siblings of children with 23(2-3 SPEC. ISS), 143–152. https://doi.org/10.1016/j.ijdevneu. autism spectrum disorder. Journal of Autism and Developmental 2004.05.001.

Journal

Journal of Abnormal Child Psychology – Springer Journals

Published: Mar 12, 2018

There are no references for this article.