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

Learn More →

The role of prefrontal–subcortical circuitry in negative bias in anxiety: Translational, developmental and treatment perspectives:

The role of prefrontal–subcortical circuitry in negative bias in anxiety: Translational,... Anxiety disorders are the most common cause of mental ill health in the developed world, but our understanding of symptoms and treatments is not presently grounded in knowledge of the underlying neurobiological mechanisms. In this review, we discuss accumulating work that points to a role for prefrontal–subcortical brain circuitry in driving a core psychological symptom of anxiety disorders – negative affective bias. Specifically, we point to converging work across humans and animal models, suggesting a reciprocal relationship between dorsal and ventral prefrontal–amygdala circuits in promoting and inhibiting negative bias, respectively. We discuss how the developmental trajectory of these circuits may lead to the onset of anxiety during adolescence and, moreover, how effective pharmacological and psychological treatments may serve to shift the balance of activity within this circuitry to ameliorate negative bias symptoms. Together, these findings may bring us closer to a mechanistic, neurobiological understanding of anxiety disorders and their treatment. Keywords Anxiety, circuit, negative bias, prefrontal cortex Received: 19 December 2017; accepted: 9 April 2018 Introduction Anxiety disorders are the most common cause of mental illness in consist of shared symptoms which are similar manifestations of the developed world, with large social, economic and psychologi- relatively few underlying dimensions (Caspi et al., 2014; Clark cal impacts (DiLuca and Olesen, 2014; Shin and Liberzon, 2009; et al., 2017; Kaczkurkin et al., 2017; Kotov et al., 2017; Krueger Vos et al., 2016). A propensity towards the development of anxi- and Eaton, 2015; Lahey et al., 2012, 2017). ety disorders is heritable (Hettema et al., 2005), often begins in Recent efforts such as the Research Domain Criteria (RDoC; childhood or adolescence (Beesdo et al., 2009; Pine, 2007) and Insel et al., 2010), therefore, attempt to re-frame the investigation persists into adulthood (Copeland et al., 2014; Craske et al., of psychiatric disorders by advocating a trans-diagnostic approach 2017). It is estimated that close to one in four people will suffer focusing on the neurobiological mechanisms underpinning symp- from an anxiety disorder – including generalised anxiety (GAD), toms that cut across traditional categorical diagnoses. In particu- post-traumatic stress disorder (PTSD), social anxiety or phobias lar, one domain within the RDoC, Negative Valence Systems, – in their lifetime (Kessler et al., 2005, 2012), but currently avail- includes responses to aversive situations such as fear, anxiety, able psychological and pharmacological treatments are effective sustained threat and loss (/reward omission) that overlap with a for less than half of these individuals (Roy-Byrne, 2015; key concept from the clinical psychology literature – negative Community and Mental Health team, 2014) and progress in the affective bias. Negative biases in cognition are thought to promote discovery of anxiolytic drugs has been slow (Griebel and Holmes, and uphold key symptoms of many psychiatric conditions but are 2013). One reason for this treatment gap is that we have a limited understanding of the biological mechanisms by which anxiety Division of Psychology and Language Sciences, University College symptoms emerge or how these mechanisms are modulated by London, London, UK our current interventions. As such, we struggle to develop new Institute of Cognitive Neuroscience, University College London, treatments that can modulate known biological targets. Moreover, London, UK it is increasingly clear that our current diagnoses, based largely on self-reported symptoms, do not map clearly onto underlying biol- Corresponding author: ogy or indeed onto the latent structure of the self-reported symp- Oliver J. Robinson, Institute of Cognitive Neuroscience, University College toms themselves (Cuthbert and Insel, 2013; Kotov et al., 2017). London, Alexandra House, 17-19 Queen Square, London WC1N 3AZ, UK. Indeed, factor analyses suggest that many categorical disorders Email: o.robinson@ucl.ac.uk Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). 2 Brain and Neuroscience Advances especially prominent in anxiety disorders, perpetuated by antici- development and treatment (Aue and Okon-Singer, 2015; Cisler pation of – and uncertainty about – future events (Grupe and and Koster, 2010; Dodd et al., 2017). Nitschke, 2011, 2013). Although none of our current treatments or diagnoses are based on a mechanistic neurobiological understand- Animal models of negative bias ing of negative bias, recent work has begun to delineate the role that interactions between the prefrontal cortex (PFC) and subcor- Research falling within the Negative Valence Systems domain of tical regions such as the amygdala play in the manifestation of RDoC in animal models often makes a distinction between fear negative bias in anxiety. It is this circuitry that is the focus of the and anxiety. In the psychological literature, anxiety is defined as present review. a prolonged state of heightened anticipatory arousal, often prompted by distal or unpredictable threats (Davis et al., 2010). Fear, on the contrary, is conceptualised as a ‘fight or flight’ reac- Negative bias in anxiety tion and typically involves active defence against immediate Anxiety disorders are characterised by a general ‘negative bias’ threat, usually dissipating upon removal of the threat. In ‘real- in both attention and memory towards affectively negative world’ terms, a person’s reaction to a spider on the table in front (/threatening/aversive) information that promotes and upholds of them might elicit a fear response (which may be exacerbated the anxious state while having knock-on effects in a wide range in cases of phobia), whereas the knowledge that a spider might be of other cognitive functions (e.g. learning, inhibitory control; in the room but uncertainty of its location might elicit anxiety. Craske et al., 2017; Robinson et al., 2013a). It has been widely Elevations in both responses could of course play a key role in shown, for instance, that people with anxiety tend to interpret driving negative affective bias (although see Fox and Shackman, neutral information in a more negative light, have maladaptive 2017; Shackman and Fox, 2016, for suggestions of why this attention biases towards threat even when threats are not imme- explicit distinction between fear and anxiety might be problem- diately present or relevant (this may be particularly prominent in atic with regard to the underlying neurobiology). some subtypes of anxiety such as social anxiety disorder; Abend Animal models are particularly useful in the investigation of et al., 2017) and have a bias towards learning about negative the neural basis of anxiety and fear because of their cross-spe- information (Abend et al., 2017; Hakamata et al., 2010; Monk cies overlap in the neural circuitry underlying these processes et al., 2006; Okon-Singer and Aue, 2017; Roy et al., 2008). It is (Davis, 2000; LeDoux, 2000); cross-species functional homo- of course possible to break negative biases into specific subtypes logues of brain circuitry can inform translational research across of bias (for a comprehensive attempt to do this, see Grupe and humans and rodents and can provide testable models, even if the Nitschke, 2013), but here we aim to build a broad preliminary circuits themselves are not directly conserved across species. model across disparate animal and human experimental data, For instance, startle response, in which whole-body jump is typi- along with clinical data, with the goal of drawing holistic con- cally used as an index in rodents, is paralleled by an eye-blink clusions. Similarly, the term negative bias of course encom- response in humans (Davis, 2001) and is a reliable experimental passes myriad constructs, including a distinction made by many measure of aversive responding on some cognitive tasks researchers between fear and anxiety (Davis et al., 2010), as well (Aylward and Robinson, 2017) but not necessarily on others as subcategories of learned versus prepotent fears (LeDoux, (Bradford et al., 2015). This startle reflex is heightened by both 2000; Phelps et al., 2004). However, in this review, we broadly fear and anxiety states across rats and humans (Grillon and focus across the Negative Valence Systems domain of the RDoC Davis, 2007; Grillon et al., 1991), but the subcortical circuitry (Insel et al., 2010) in an attempt to identify common patterns and responsible may differ (Robinson et al., 2012a). Specifically, build a simple model of how negative bias is generated at the fear responses are associated with the central amygdala, anxiety neurobiological level. If we were to be too fine-grained in our responses are associated with the bed nucleus of the stria termi- definitions of negative bias, we will find ourselves with multiple nalis, and both fear and anxiety responses are associated with the non-overlapping studies, which would limit interpretation. basolateral amygdala (BLA; Davis et al., 2010; Sengupta et al., To demonstrate our aim of bridging disparate experimental 2017; Tovote et al., 2015), although this distinction has been literature and clinical utility, consider Beck’s (1967) early challenged and warrants further investigation (Gungor and Paré, theory of elevated negative bias (or ‘negative schemata’). In 2016; Shackman and Fox, 2016). Nevertheless, taken together, this framework, Beck proposed a cognitive triad illustrating a rodent and human work points to the clear role of subcortical cycle among a negative view of the world, the self and the regions, and the extended amygdala in particular (along with its future. This model encompasses wide-ranging cognitive func- inter-connections and external projections), in driving aversive/ tions from memory to attention but nonetheless forms the fear responding and hence negative bias (Boeke et al., 2017; basis of successful psychological treatments such as cogni- Campese et al., 2015, 2017; Sengupta et al., 2017; Terburg et al., tive-behavioural therapy (CBT), highlighting potential advan- 2012; Tovote et al., 2015). tages in taking a broad approach to linking cognitive research However, regions of the brain rarely, if ever, work in isolation. with clinical practice. Within the hierarchy of neural processing, these subcortical We therefore review converging evidence across humans and regions also interact with ‘higher’ cortical areas. This is perhaps animal models suggesting that negative bias may arise, at least in best illustrated by work exploring the weakening of learned aver- part, from activity within prefrontal regions and their interactions sive responses during extinction. During fear extinction, a cue with subcortical regions (Etkin and Wager, 2007; Shin and which previously indicated the onset of an aversive event no Liberzon, 2009). These circuits may underpin the ability to longer predicts a negative outcome, so the individual must ‘extin- engage or disengage attention from threats and may be critical to guish’ their original aversive response. Animal models of condi- understanding the mechanistic basis of negative bias as well as its tioned fear extinction indeed implicate subcortical regions such Carlisi and Robinson 3 as the BLA, but they extend the circuitry to encompass medial information flow between these regions in a circuit that drives the prefrontal cortical regions as well. In particular, within the rodent overall output. PFC, subdivisions of the infralimbic (IL) and the prelimbic (PL) Collectively, animal work therefore suggests a putative neural have been posited to play distinct roles in the expression and mechanism of negative bias; one (bi-directional PL–amygdala) extinction of conditioned fear, with the IL supporting fear extinc- circuit may serve to facilitate negative bias, while another (bi- tion as expressed by the amygdala and the PL conversely promot- directional IL–amygdala) circuit may serve to suppress negative ing fear expression as expressed by the amygdala (Klavir et al., bias (Calhoon and Tye, 2015). This simplified heuristic provides 2017; Morgan and LeDoux, 1995; Sierra-Mercado et al., 2011; a framework with which to consider neurobiological research in Vidal-Gonzalez et al., 2006). In monkeys, activity in the dorsal anxious humans. anterior cingulate (dACC) is correlated with the BLA during fear learning and memory acquisition (Klavir et al., 2013; Livneh and Paz, 2012), and in rodents, PL response to cues predicting an Neurobiological basis of negative bias aversive event increases post-fear learning (Burgos-Robles et al., in humans 2009). Using pharmacological or electrical stimulation and optogenetic approaches, it has been shown that, on the other Perhaps unusually for a symptom related to psychiatric disorder, hand, increased activity in the rodent IL predicts fear extinction negative bias in anxiety can be adaptive. For example, when one in the amygdala (Do-Monte et al., 2015; Klavir et al., 2017; is walking home late at night and hears an unexpected noise Milad and Quirk, 2002, 2012). Having said that, recent rodent down a dark alley, an appraisal of this situation as potentially work has suggested that this dissociation may not be as clear-cut threatening raises awareness, preparing the body’s fight or flight as previously thought. For example, the role of the IL in fear response in the event of immediate danger. In other words, nega- extinction has been challenged in optogenetics work showing tive bias in anxiety can promote harm avoidance. However, if this that extinction recall was intact after the silencing of IL neurons heightened anxiety and negative bias does not subside when one and that stimulation of ventromedial prefrontal cortex (vmPFC) is subsequently safe at home, this response becomes maladaptive inputs to the amygdala facilitated extinction memory formation and can impair daily functioning (i.e. it transitions into a patho- but not retrieval (Bukalo et al., 2015; Do-Monte et al., 2015). logical state). Thus, it was suggested by the pioneers of CBT that Moreover, the distinction between PL/dorsal and IL/ventral pre- a biased appraisal of threat (i.e. negative bias) leading to catastro- frontal regions being responsible for fear expression and suppres- phising or excessive worry is a central characteristic of anxiety sion, respectively, has been challenged (Giustino and Maren, disorders (Beck and Clark, 1997). Dorsal regions of the PFC 2015). For instance, rodent work has shown that these regions (dorsomedial PFC (dmPFC) and dACC) seem to be associated have structurally similar projections to the amygdala (Cho et al., with this behavioural response at the neural level; these regions 2013; Gutman et al., 2012; Hübner et al., 2014; Pinard et al., are activated during conscious threat appraisal in healthy indi- 2012), and functionally dichotomous distinctions between these viduals and have been shown to be overactive during threat regions have been shown in the opposite direction than was ini- appraisal in pathological anxiety (for review, see Kalisch and tially postulated (Chang et al., 2010). Gerlicher, 2014). This rodent research nevertheless highlights a key potential Consistent with the animal work highlighted above, patho- mechanism of negative bias; namely that the overall expression logical negative bias in humans may in fact result in part from an of aversive responding may be held in the balance of opposing inability to extinguish conditioned fear responses driven in turn circuitry. Thus, whether negative bias is expressed or dampened by this altered, worry-related PFC responding (Milad et al., 2007; may depend on whether one of these circuits is able to override Rothbaum and Davis, 2003). Healthy subjects show increased the other, with sub-regions of the medial prefrontal cortex activation in the vmPFC during acquisition and retrieval of (mPFC) playing a key role in arbitrating this response. Indeed, extinction (Kalisch et al., 2006; Milad and Rauch, 2007; Phelps across a range of paradigms in rodents, the PFC has been shown et al., 2004), which has led to the suggestion that this region in to play a regulatory role over BLA activation during fear expres- the human brain might be functionally (albeit perhaps not struc- sion, social interaction and anxiety-related behaviours (Bickart turally) homologous to the rodent IL. On the other hand, patients et al., 2014; Bremner, 2004; Davis, 1998; Felix-Ortiz et al., 2013; with anxiety disorders have shown reduced activation in the Janak and Tye, 2015; Kling and Steklis, 1976). vmPFC along with increased activation in the dACC, leading to That said, the implicit implication that the PFC is a ‘top- the suggestion that the dACC might be functionally homologous down’ regulator of the amygdala during fear extinction should be to the rodent PL (Milad et al., 2009). challenged. Optogenetics research in rodents has shown bi-direc- Consistent with this proposition, at the neural level, the dACC tional effects of modulating BLA projections to the IL and PL and adjacent dmPFC have been implicated in the appraisal and during a number of behavioural assays assessing anxiety-like expression of fear (Etkin et al., 2011; Vogt, 2005), as well as the behaviour (Courtin et al., 2014; Felix-Ortiz et al., 2016; Herry anticipation of emotional stimuli (Erk et al., 2006). Moreover, et al., 2008; Laviolette et al., 2005). Specifically, BLA activity Kalisch and Gherlicher (2014) argue that the dACC/dmPFC can projecting up to PL regions is increased during fear conditioning be further subdivided into an anterior part, the rostral dACC/ (Senn et al., 2014), while fear extinction also enhances activity in dmPFC and a posterior part, with the rostral but not posterior part BLA projections up to the IL (Milad and Quirk, 2002; Senn et al., implicated in conscious threat appraisal and worry (Kalisch and 2014). Moreover, recordings from non-human primates further Gerlicher, 2014; Mechias et al., 2010). Thus, activity in dorsal support this bi-directional effect during fear learning (Klavir PFC regions is broadly associated with increased negative bias. et al., 2013). In other words, it is not so much that the PFC ‘regu- The dACC has also been linked to the adaptive control of behav- lates’ the amygdala but rather the reciprocal relationship of iour as well as the risk of development of anxiety disorders 4 Brain and Neuroscience Advances (Cavanagh et al., 2017; Goodkind et al., 2015; McTeague et al., responding (as indexed by a threat-by-valence interaction in reac- 2017; Meyer, 2017). tion times, driven by a valence-specific reduced reaction time to Regarding ventral regions, studies in healthy adults (Bishop, fearful faces under threat vs safe conditions), suggesting a key 2007) as well as adults (Etkin and Wager, 2007; Milad et al., mediating role for dACC/dmPFC–amygdala circuitry in driving 2007; Price and Drevets, 2012) and children and adolescents negative bias. Critically, coupling within this same circuitry was (Guyer et al., 2008; Monk et al., 2006; Strawn et al., 2012) with shown to be elevated at baseline in individuals with an anxiety anxiety disorders have shown abnormal function in orbitofrontal disorder (Robinson et al., 2014; in the absence of induced anxi- cortex (OFC) and ventrolateral PFC (VLPFC). The cause and ety), suggesting that the same circuitry which can be selectively effect of such abnormalities have been studied in non-human pri- engaged and disengaged in healthy controls is more persistently mates through lesions to the anterior OFC and VLPFC (Agustín- engaged in patients with clinical anxiety, thereby providing a Pavón et al., 2012; Izquierdo and Murray, 2005; Kalin et al., route by which adaptive anxiety can transition into a maladaptive 2007; Machado and Bachevalier, 2008), with findings largely state. Across both studies, however, the correlation between the showing increased anxiety during fear conditioning paradigms amygdala and dorsal cortical regions was positive. In other words, when these regions are lesioned. Thus, broadly speaking, activity activity in the dorsal cortical regions increases as activity increases in ventral cortical regions is associated with reduced negative in the amygdala and vice versa. The role that this circuit seems to bias (although it should be noted that this association may not be play in threat responding therefore appears somewhat analogous as consistent as previously thought (cf. Shackman et al., 2011). to the role of the PL in rodents. Thus, a human functional homo- The hippocampus is another structure that has been hypothe- logue of the rodent PL–amygdala circuit may drive increased sised to play a critical role in the pathophysiology of anxiety. threat responding and negative affective biases in anxiety Specifically, this region is a key mediator of the acquisition and disorders. expression of learned fear, as demonstrated by a number of early However, rodent work has also highlighted the contrasting lesion studies showing that hippocampal lesions dampened fear role of the inhibitory IL circuit (Kim et al., 2011a). To this end, response to previous learned associations (Kim and Fanselow, another study in humans (Vytal et al., 2014) expanded this puta- 1992; Phillips and LeDoux, 1992; Selden et al., 1991). Studies in tive circuitry to encompass a reciprocal inhibitory circuit. both human and rodents suggest that this region integrates con- Specifically, inducing anxiety during an adapted resting-state textual information during fear conditioning and may regulate scan replicated positive dmPFC–amygdala coupling, but at the context-dependent recall after extinction (Giustino and Maren, same time enhanced negative coupling between a ventral medial 2015). The rodent PL and IL receive excitatory inputs from both prefrontal region and the amygdala. In other words, while the dorsal and ventral hippocampus (Little and Carter, 2013), and increased dorsal activation was associated with increased amyg- it has been suggested that, similar to the amygdala, these projec- dala activation, increased ventral activation was associated with tions may inhibit downstream mPFC outputs (Sotres-Bayon decreased amygdala activity (Vytal et al., 2014). Earlier positron et al., 2012). In humans, Linnman et al. (2011) demonstrated that emission tomography (PET) studies in humans have shown simi- fear (elicited by electric shock expectation) was associated with larly contrasting relationships between prefrontal and subcortical increased connectivity between the hippocampus and the vmPFC, regions (Linnman et al., 2012a, 2012b). For example, during and decreased connectivity between the hippocampus and the red extinction training, resting amygdala metabolism positively pre- nucleus midbrain region, suggesting that the hippocampus may dicted vmPFC activation and negatively predicted dACC activa- facilitate a switch between what they term a ‘fear’ network and a tion, but during extinction recall, these relationships were in the ‘resting’ network. opposite direction (Linnman et al., 2012a). In a rodent study However, as highlighted by the rodent literature above, nega- investigating the impact of early-life environmental stress, tive bias is not so much driven by regions acting in isolation. Johnson et al. (2018) showed that stress was related to increased Rather, it is the cortical–subcortical circuitry that is important for amygdala–PFC and amygdala–hippocampus coupling and that anxiety response. For example, Kalin et al. (2016) used a viral this connectivity was related to anxiety-like behaviours in a vector approach in primates to demonstrate a relationship between translational model of early-life stress. Thus, these studies dem- overexpression of corticotropin-releasing hormone (CRH) in the onstrate that the relationship between distinct mPFC regions may dorsal amygdala and increased defensive behaviour during expo- have opposing effects on aversive responding. Another study of sure to threat. Moreover, this link between metabolism and behav- resting-state functional connectivity in healthy humans showed iour has also been observed in rodents and was associated with that those individuals who reported high levels of anxiety were functional connectivity between the dorsal amygdala and OFC characterised by negatively correlated amygdala–vmPFC con- (Regev et al., 2012). To this end, connectivity between the dACC/ nectivity, while this connectivity was positively correlated in dmPFC and the amygdala has been implicated in the pathophysi- those reporting low levels of anxiety (Kim et al., 2011b). ology of anxiety in humans. Structurally, the integrity of white Moreover, amygdala–dmPFC connectivity was negatively cor- matter tracts between the amygdala and the PFC has been shown related only in those reporting low anxiety. More dorsal regions to predict individual differences in trait anxiety (Kim and Whalen, of the PFC (like the PL in rodents) may increase aversive 2009). Functionally, Robinson et al. (2012a) studied the role of responding, while more ventral regions (like the IL in rodents) these regions in negative bias during induced anxiety in healthy may reduce aversive responding. Moreover, the nature of these individuals. They found that connectivity increased during the functional imaging connectivity analyses means that they are processing of threatening stimuli (fearful faces) selectively in the non-directional. In other words, it is not possible to say whether context of induced anxiety. Moreover, the strength of this connec- one region is driving the other – it is simply a correlation. Given tivity was positively correlated with participants’ subjective rat- the bi-directional nature of the rodent work highlighted above, ings of anxiety, as well as the extent of negative bias in behavioural these circuits therefore should not be considered ‘top-down’ or Carlisi and Robinson 5 ‘bottom-up’; rather, the overall reciprocal cortical–subcortical One influential idea common to the human and rodent devel- interaction likely drives the ultimate behavioural expression. opmental literature is that learned fear associations (i.e. memo- Together, these findings highlight the value of translational ries) from early life are important contributing factors to the research. A model of cortical–subcortical interactions during subsequent development of anxiety disorders (Britton et al., negative bias inspired by rodent work provides a framework 2011; Glenn et al., 2012; Jacobs and Nadel, 1985). Cross-species within which to consider the role of neural circuitry in negative animal work (Harlow and Harlow, 1965; Hess et al., 1962) has bias in humans. shown that fear learning is characterised by approach behaviour (such as maternal attachment or odour approach) in infants, but is characterised by almost diametrically opposed avoidance behav- Development iour (such as maternal or odour avoidance or avoidance of ele- vated/open areas in typical rodent paradigms) in adults (for an The work reviewed above therefore suggests that medial prefron- extensive review, see Ganella and Kim, 2014). This also suggests tal–amygdala interactions may drive the negative bias symptoms that at some point during development, there is a change in the that are a core feature of anxiety disorders. However, how these underlying neurobiology promoting this behaviour. mechanisms develop and persist across the lifespan remains Integrating this within the circuitry framework of the pre- unclear. If we want to target these symptoms and intervene early, sent review, Chan et al. (2011) inactivated PL in juvenile, pre- it is important to determine when and how alterations to these adolescent and adult rats and found that PL inactivation circuits emerge. significantly reduced freezing behaviour, as would be pre- Within the general population, pathological anxiety com- dicted by the above reviewed evidence, but that it only did so monly emerges during childhood or adolescence and reflects a in adolescent and adult rats, suggesting that the role of differ- combination of genetic factors and early-life experiences (Pine, ent medial prefrontal regions in negative bias changes across 2007). ‘Anxious temperament’ is considered to be a stable trait development. In other words, differences in fear responding, across time, and those with extreme levels of such traits are at a mediated by amygdala–medial prefrontal pathways, may par- higher risk for developing clinical or pathological anxiety tially be a result of a more protracted course of development (Arnaudova et al., 2013; Jones, 2013; Nugent et al., 2011). and reorganisation in these cortical–subcortical pathways Similarly, the stable traits of ‘behavioural inhibition’, a tempera- (Arruda-Carvalho et al., 2017; Ganella and Kim, 2014; ment characterised by a tendency to withdraw from new situa- LeDoux, 2000; Pattwell et al., 2016). Similar developmental tions (Kagan et al., 1987; Svihra and Katzman, 2004) and, more changes in prefrontal–subcortical negative bias circuitry are broadly, ‘dispositional negativity’ (Shackman et al., 2016) are also seen in humans; in typically developing humans, mPFC– thought to be early phenotypes of anxiety disorders. There is evi- amygdala connections are immature during childhood and dence that anxiety-related amygdala abnormalities and affected strengthen to adult levels during adolescence (Gee et al., top-down prefrontal regulation originate early in development 2013a, 2013b), and structural changes in white matter have (Clauss and Blackford, 2012; Kalin, 2017). Moreover, it has been been shown to mediate amygdala function in adolescents estimated that 50% of children showing increased behavioural (Swartz et al., 2014). Moreover, early perturbations in medial inhibition in childhood will later develop stress-related psycho- prefrontal circuitry have been implicated in the development of pathology (Clauss and Blackford, 2012). This is paralleled by anxiety and depression. For example, a preliminary study in findings of reduced amygdala–dorsolateral prefrontal cortex adolescents with depression found that patients had decreased (dlPFC) coupling in preadolescent children diagnosed with an functional connectivity in a subgenual (ventral) anterior cingu- anxiety disorder as well as in young non-human primates with late (ACC)-based network compared to healthy adolescents elevated levels of traits related to anxious temperament (includ- (Cullen et al., 2009). Moreover, negative coupling within ing heightened behavioural inhibition; Birn et al., 2014). vmPFC–amygdala circuitry during fear extinction was recently Although more longitudinal studies are needed to confirm this, shown only in adults and not adolescents (Ganella et al., 2017). this evidence suggests that rapid changes in the mPFC and the These ventral regions may reflect overlapping human homo- later maturation of amygdala–cortical connections during adoles- logues of the rodent IL. Thus, a developmental delay in the cence, a period recently suggested to encompass 10–24 years of ability to engage the circuitry that can dampen negative bias age (Sawyer et al., 2018), may contribute to the emergence of might explain the emergence of anxiety disorders during ado- anxiety during a specific developmental window (Andersen, lescence. This work is in its infancy, but the concept of recipro- 2003; Casey et al., 2008). Indeed, prospective studies in humans cal cortical–subcortical circuits again provides a framework (Giedd et al., 1999; Jones et al., 2017; Kalin, 2017; Swartz and with which to consider the emergence of anxiety and negative Monk, 2013) as well as rodent studies (Arruda-Carvalho et al., bias during development. 2017; Cohen et al., 2013; Gee et al., 2013a; Pattwell et al., 2016) have shown that this period constitutes a window of heightened risk for the development of anxiety. However, vast structural Treatment brain changes have also been observed during childhood, sug- gesting that children are subject to a heightened vulnerability to If prefrontal–subcortical circuitry is critical in the development environmental impacts which may influence the development of and manifestation of negative bias, then modulation of this cir- anxiety even before the onset of adolescence. Indeed, behaviour- cuitry should serve to modify negative affective biases and hence ally inhibited temperament has been observed in young children treat symptoms. The first-line treatments for clinical anxiety are who later develop anxiety, with similar neural circuitry altera- serotonergic medication and psychological therapy. Emerging tions linking these phenotypes (Buzzell et al., 2017; Gold et al., evidence suggests that successful response to both types of treat- 2016; Sylvester et al., 2016). ment may also depend on this prefrontal–subcortical circuitry. 6 Brain and Neuroscience Advances assessed whether CBT combined with attention bias modifica- The role of serotonin in pharmacological tion therapy (ABMT) was more clinically effective than CBT treatment alone and whether this treatment response could be predicted Serotonin (5-hydroxytryptamine (5-HT)) has long been impli- through pre-treatment amygdala-based functional connectivity cated in the neuropsychopharmacology of anxiety (Dayan and (White et al., 2017). This study found that patients differed from Huys, 2009; Harmer et al., 2009, 2011), largely because selective controls in amygdala–insula connectivity on a threat attention serotonin reuptake inhibitors (SSRIs) are the most common and task. Moreover, while both CBT groups showed clinical improve- effective pharmacological treatment for anxiety disorders ment, the combined CBT + ABMT group showed the greatest (Harmer et al., 2009, 2011). It is thought that serotonin plays a reduction in symptoms and that baseline amygdala functional particular role in maintaining the balance between the processing connectivity differentially predicted the level of treatment of appetitive and aversive information (Cools et al., 2008; response in patients. Crockett et al., 2009; Robinson et al., 2012b) and more precisely However, whether these changes in cortical–subcortical cir- in the inhibition of PFC-linked neural circuitry important for cuits are driven by CBT, or whether they simply reflect reduced driving negative bias (Crockett et al., 2009; Robinson et al., overall anxiety and negative bias per se, is unclear. To this end, 2013b). basic research has attempted to determine causality. Specifically, The impact of serotonin in healthy humans can be studied by it has been shown that in healthy individuals, simple attentional acute tryptophan depletion – a dietary manipulation that tempo- instruction can alter the engagement of affective-bias-related rarily reduces serotonin levels (Crockett et al., 2012). Reduced dmPFC–amygdala circuitry (Robinson et al., 2016). When sub- serotonin has been shown to increase positive coupling within jects are instructed to pay attention to neutral aspects of com- the same circuit shown to be elevated by induced anxiety (Vytal pound cues (rather than the affectively salient components of the et al., 2014) and at baseline in individuals with an anxiety disor- same cues), anxiety-induced amygdala–dmPFC coupling (as der (Robinson et al., 2014), suggesting that serotonergic drugs seen above; Robinson et al., 2012a) is down-regulated. This sug- (which putatively elevate serotonin availability) may work by gests that psychological treatments such as CBT may reduce reducing activity within this dorsal prefrontal circuit (Robinson negative bias by down-regulating the dorsal PFC–amygdala cir- et al., 2013b), thus reducing negative bias. By contrast, a study cuitry that promotes negative bias. In the context of threat pro- using a different paradigm showed that tryptophan depletion can cessing, there has been limited work showing whole-brain also decrease coupling between the amygdala and a more ventral increased ventrolateral prefrontal activation in anxious youth prefrontal region (Passamonti et al., 2012). Moreover, direct who underwent CBT relative to controls (Maslowsky et al., reductions in ventrally located orbitofrontal serotonin in the mar- 2010), as well as reduced dorsomedial prefrontal activation post- moset can increase negative bias (Rygula et al., 2015). Within the CBT relative to pre-CBT in individuals with social anxiety framework described above, this suggests that serotonin can also (Klumpp et al., 2013). Nevertheless, the role that CBT plays in serve to promote ventral PFC circuits that inhibit aversive pro- dorsal versus ventral prefrontal–amygdala circuitry in humans cessing while inhibiting dorsal PFC circuits that promote aver- has not been systematically studied. Moreover, recovery rates of sive responding. As such, pharmacological treatments may work patients with anxiety undergoing psychological treatment are less by restoring the balance between the circuits that, respectively, than 50% (Community and Mental Health team, 2014), so it is promote and inhibit negative bias. Recent work also suggests that plausible that these mechanisms are again only relevant in a sub- the influence of serotonergic drugs on this circuitry might be set of patients. mediated by genetic factors (Perna et al., 2005; Santangelo et al., 2016), which may in turn explain why such medications only work for a subset of anxious patients. Conclusions and future directions In this review, we have outlined evidence across animals and humans suggesting that bi-directional prefrontal–subcortical cir- Psychological treatment cuits and their interactions may drive elevated aversive process- CBT is the most common psychological intervention used to treat ing, or negative bias, in anxiety. Specifically, we point to anxiety and is based on the premise that negative biases in converging evidence within the Negative Valance Systems thoughts and actions can be shifted through cognitive reappraisal domain of the RDoC which suggests that ventral PFC–subcortical and emotion regulation strategies (Beck and Clark, 1997). There circuitry in humans may be associated with reduced negative have been numerous studies (see review by Brooks and Stein, bias, while more dorsal PFC–subcortical circuitry may be associ- 2015) which suggest that CBT modulates prefrontal–subcortical ated with increased negative bias. Moreover, we provide evi- interactions. Indeed, baseline medial prefrontal and amygdala dence suggesting that the emergence of anxiety in adolescence activity might even predict treatment response to CBT in anxiety may be a result of differential developmental trajectories of these (Klumpp et al., 2017). For instance, Shou et al. (2017) showed circuits and that both pharmacological and psychological inter- that functional connectivity between the amygdala and the ventions might be effective by modulating the overall balance of fronto-parietal network increased in patients with major depres- these circuits in driving negative affective bias. These findings sive disorder (MDD) or PTSD who underwent a course of CBT are summarised in Figure 1. compared to controls, supporting a mechanism by which this cir- Nevertheless, it is still unclear exactly how we bridge the gap cuitry may interact with psychological intervention (although it between brain and behaviour. Although we can associate these should be noted that this study did not include a patient group that circuits with negative bias, we cannot yet say how exactly the did not undergo CBT, so the specificity of these results is difficult underlying neuronal activity is translated into observable behav- to quantify). Similarly, a study of adolescents with anxiety iour. One particularly promising avenue in this regard is the Carlisi and Robinson 7 data (Thompson et al., 2014). Moreover, the UK Biobank (Sudlow et al., 2015) is a consortium across 22 research centres in the United Kingdom with genetic and longitudinal physical health and behavioural data on over 500,000 participants, all of which has been made open access. These are early efforts, par- ticularly in the field of anxiety disorders, but promising mega- and meta-analyses have already come out of such efforts in other fields of psychiatry such as obsessive–compulsive disorder and schizophrenia (Boedhoe et al., 2016; De Wit et al., 2014; Van Erp et al., 2016). Finally, it is important to investigate how these effects change over time. That is, are these mechanisms stable, or do they change across development to influence symptom onset and persistence? Longitudinal studies are critical for understanding these ques- tions. There have been longitudinal studies investigating brain changes over time in adolescents (e.g. the IMAGEN study; Schumann et al., 2010), but this investigation needs to be scaled up to larger populations and multiple time points and age ranges if we are to truly understand the developmental changes that occur across the life course of anxiety disorders. One promising example of this work currently underway is the Adolescent Brain Cognitive Development study (ABCD; https://abcdstudy.org/ index.html), which is the largest long-term longitudinal study of brain development in the United States, currently in the process of collecting biological and behavioural data on over 10,000 chil- dren aged 9–10. Similarly, to gain an understanding of the under- lying genetic contributions of anxiety, it is important to investigate Figure 1. Schematic summarising findings and proposed simplified the extent to which certain features and symptoms are heritable. model of negative affective bias in anxiety. Bi-directional excitatory This can be achieved through longitudinal twin studies (e.g. the connections between dorsal regions of the mPFC/ACC and the amygdala Twins Early Development Study (TEDS; Oliver and Plomin, promote negative bias, while inhibitory connections between ventral 2007) and the Tennessee Twin Study (Lahey et al., 2008)), but regions coupled with the amygdala inhibit negative bias. The ventral many of the existing studies do not focus on brain imaging due to inhibitory circuit may only fully develop in adulthood, meaning limited time and resources and the high cost involved in neuroim- that adolescence is a period of high vulnerability to negative bias. aging research. Regardless, observational population-based stud- Successful treatments (SSRIs and CBT) may be effective via promotion ies are an important complimentary approach to the small-scale of the ventral circuit and inhibition of the dorsal circuit. case–control designs more frequently implemented in neuroim- ACC: anterior cingulate cortex; mPFC: medial prefrontal cortex; CBT: cognitive- aging research on anxiety. behavioural therapy; SSRI: selective serotonin reuptake inhibitor. In conclusion, work has begun to delineate overlapping neu- ral networks involving the PFC and subcortical regions includ- ing the amygdala that may drive aversive responding and nascent field of computational psychiatry, which attempts to negative bias in both animals and humans. There is also promis- bridge the gap between brain activation and observable symp- ing evidence that pharmacological and psychological interven- toms (Huys et al., 2016). Specifically, using mathematical theo- tions can shape this circuitry and hence ameliorate negative ries of cognition and human behaviour, computational psychiatry affective bias. Future research should expand these findings to aims to objectively quantify the calculations generated by neu- larger populations and investigate how these neural underpin- rons which shape behaviour (Huys et al., 2016). This work is in nings arise in childhood/adolescence and change over time to its infancy but has begun to delineate the computational basis of shape behaviour. common symptoms in anxiety linked to negative bias, such as avoidance (Mkrtchian et al., 2017), risk aversion (Charpentier et al., 2017) and goal-directed behaviour (Carlisi et al., 2017; Declaration of conflicting interests Gillan et al., 2014). The author(s) declared no potential conflicts of interest with respect to Furthermore, if we are to understand current findings in a the research, authorship and/or publication of this article. truly generalisable context, it is critical to investigate these mech- anisms in large-scale populations. Cohort studies are an ideal Funding way to examine these questions at the population level, often C.O.C. is supported by a Wellcome Trust Sir Henry Wellcome Postdoctoral sampling from a diverse community of individuals. Data sharing Fellowship (206459/Z/17/Z) and O.J.R. is supported by a Medical efforts have attempted to address this. For example, the ENIGMA Research Council Career Development Award (MR/K024280/1). consortium is an international collaboration of research centres which aims to combine neuroimaging and genetic datasets from ORCID iD sites around the world in an attempt to amass sample sizes large enough to detect very small effects in brain imaging and genetic Christina O. Carlisi https://orcid.org/0000-0002-0942-8586 8 Brain and Neuroscience Advances Buzzell GA, Troller-Renfree SV, Barker TV, et al. (2017) A neurobe- References havioral mechanism linking behaviorally inhibited temperament and Abend R, de Voogd L, Salemink E, et al. (2017) Association between later adolescent social anxiety. Journal of the American Academy of attention bias to threat and anxiety symptoms in children and adoles- Child & Adolescent Psychiatry 56(12): 1097–1105. cents. Depression and Anxiety 35(3): 229–238. Calhoon GG and Tye KM (2015) Resolving the neural circuits of anxiety. Agustín-Pavón C, Braesicke K, Shiba Y, et al. (2012) Lesions of ventro- Nature Neuroscience 18: 1394–1404. lateral prefrontal or anterior orbitofrontal cortex in primates heighten Campese VD, Gonzaga R, Moscarello JM, et al. (2015) Modulation of negative emotion. Biological Psychiatry 72(4): 266–272. instrumental responding by a conditioned threat stimulus requires Andersen SL (2003) Trajectories of brain development: Point of vulner- lateral and central amygdala. Frontiers in Behavioral Neuroscience ability or window of opportunity? Neuroscience & Biobehavioral 9: 293. Reviews 27(1–2): 3–18. Campese VD, Soroeta JM, Vazey EM, et al. (2017) Noradrenergic regu- Arnaudova I, Krypotos A-M, Effting M, et al. (2013) Individual differ- lation of central amygdala in aversive pavlovian-to-instrumental ences in discriminatory fear learning under conditions of ambiguity: transfer. eNeuro 4(5): pii. A vulnerability factor for anxiety disorders? Frontiers in Psychology Carlisi CO, Norman L, Murphy CM, et al. (2017) Shared and disorder- 4: 298. specific neurocomputational mechanisms of decision-making in Arruda-Carvalho M, Wu W-C, Cummings KA, et al. (2017) Optogenetic autism spectrum disorder and obsessive-compulsive disorder. Cere- examination of prefrontal-amygdala synaptic development. Journal bral Cortex 27(12): 5804–5816. of Neuroscience 37(11): 2976–2985. Casey B, Jones RM and Hare TA (2008) The adolescent brain. Annals of Aue T and Okon-Singer H (2015) Expectancy biases in fear and anxiety the New York Academy of Sciences 1124(1): 111–126. and their link to biases in attention. Clinical Psychology Review 42: Caspi A, Houts RM, Belsky DW, et al. (2014) The p factor: One gen- 83–95. eral psychopathology factor in the structure of psychiatric disorders? Aylward J and Robinson OJ (2017) Towards an emotional ‘stress test’: Clinical Psychological Science 2(2): 119–137. A reliable, non-subjective cognitive measure of anxious responding. Cavanagh JF, Meyer A and Hajcak G (2017) Error-specific cognitive Scientific Reports 7: 40094. control alterations in generalized anxiety disorder. Biological Psy- Beck AT (1967) Depression: Clinical, experimental, and theoretical chiatry 2(5): 413–420. aspects. Philadelphia: University of Pennsylvania Press. Chan T, Kyere K, Davis BR, et al. (2011) The role of the medial prefron- Beck AT and Clark DA (1997) An information processing model of tal cortex in innate fear regulation in infants, juveniles, and adoles- anxiety: Automatic and strategic processes. Behaviour Research and cents. Journal of Neuroscience 31(13): 4991–4999. Therapy 35(1): 49–58. Chang C-H, Berke JD and Maren S (2010) Single-unit activity in the Beesdo K, Knappe S and Pine DS (2009) Anxiety and anxiety disorders medial prefrontal cortex during immediate and delayed extinction of in children and adolescents: Developmental issues and implications fear in rats. PLoS ONE 5(8): e11971. for DSM-V. Psychiatric Clinics of North America 32(3): 483–524. Charpentier CJ, Aylward J, Roiser JP, et al. (2017) Enhanced risk aver- Bickart KC, Dickerson BC and Barrett LF (2014) The amygdala as a sion, but not loss aversion, in unmedicated pathological anxiety. Bio- hub in brain networks that support social life. Neuropsychologia 63: logical Psychiatry 81(12): 1014–1022. 235–248. Cho J-H, Deisseroth K and Bolshakov VY (2013) Synaptic encod- Birn RM, Shackman AJ, Oler JA, et al. (2014) Evolutionarily conserved ing of fear extinction in mPFC-amygdala circuits. Neuron 80(6): prefrontal-amygdalar dysfunction in early-life anxiety. Molecular 1491–1507. Psychiatry 19(8): 915–922. Cisler JM and Koster EH (2010) Mechanisms of attentional biases Bishop SJ (2007) Neurocognitive mechanisms of anxiety: An integrative towards threat in anxiety disorders: An integrative review. Clinical account. Trends in Cognitive Sciences 11(7): 307–316. Psychology Review 30(2): 203–216. Boedhoe PS, Schmaal L, Abe Y, et al. (2016) Distinct subcortical volume Clark LA, Cuthbert B, Lewis-Fernández R, et al. (2017) Three approaches alterations in pediatric and adult OCD: A worldwide meta-and mega- to understanding and classifying mental disorder: ICD-11, DSM-5, analysis. American Journal of Psychiatry 174(1): 60–69. and the National Institute of Mental Health’s Research Domain Cri- Boeke EA, Moscarello JM, LeDoux JE, et al. (2017) Active avoid- teria (RDoC). Psychological Science in the Public Interest 18(2): ance: Neural mechanisms and attenuation of pavlovian conditioned 72–145. responding. Journal of Neuroscience 37(18): 4808–4818. Clauss JA and Blackford JU (2012) Behavioral inhibition and risk for Bradford DE, Starr MJ, Shackman AJ, et al. (2015) Empirically based developing social anxiety disorder: A meta-analytic study. Journal comparisons of the reliability and validity of common quantification of the American Academy of Child & Adolescent Psychiatry 51(10): approaches for eyeblink startle potentiation in humans. Psychophysi- 1066–1075. ology 52(12): 1669–1681. Cohen MM, Tottenham N and Casey BJ (2013) Translational develop- Bremner JD (2004) Brain imaging in anxiety disorders. Expert Review of mental studies of stress on brain and behavior: Implications for ado- Neurotherapeutics 4(2): 275–284. lescent mental health and illness? Neuroscience 249: 53–62. Britton JC, Lissek S, Grillon C, et al. (2011) Development of anxiety: Community and Mental Health team (2014) Psychological therapies, The role of threat appraisal and fear learning. Depression and Anxi- annual report on the use of IAPT services: England– 2013/14. Avail- ety 28(1): 5–17. able at: https://digital.nhs.uk/catalogue/PUB14899 Brooks SJ and Stein DJ (2015) A systematic review of the neural bases of Cools R, Roberts AC and Robbins TW (2008) Serotoninergic regulation psychotherapy for anxiety and related disorders. Dialogues in Clini- of emotional and behavioural control processes. Trends in Cognitive cal Neuroscience 17(3): 261–279. Sciences 12(1): 31–40. Brown R, Galanter E, Hess EH, et al. (1962). New directions in psychology. Copeland WE, Angold A, Shanahan L, et al. (2014) Longitudinal patterns Oxford: Holt, Rinehart, & Winston. of anxiety from childhood to adulthood: The great smoky mountains Bukalo O, Pinard CR, Silverstein S, et al. (2015) Prefrontal inputs to study. Journal of the American Academy of Child & Adolescent Psy- the amygdala instruct fear extinction memory formation. Science chiatry 53(1): 21–33. Advances 1(6): e1500251. Courtin J, Chaudun F, Rozeske RR, et al. (2014) Prefrontal parvalbumin Burgos-Robles A, Vidal-Gonzalez I and Quirk GJ (2009) Sustained con- interneurons shape neuronal activity to drive fear expression. Nature ditioned responses in prelimbic prefrontal neurons are correlated 505(7481): 92–96. with fear expression and extinction failure. Journal of Neuroscience Craske MG, Stein MB, Eley TC, et al. (2017) Anxiety disorders. Nature 29(26): 8474–8482. Reviews Disease Primers 3: 17024. Carlisi and Robinson 9 Crockett M, Clark L, Roiser J, et al. (2012) Converging evidence for Gee DG, Humphreys KL, Flannery J, et al. (2013b) A developmental central 5-HT effects in acute tryptophan depletion. Molecular Psy- shift from positive to negative connectivity in human amygdala–pre- chiatry 17(2): 121–123. frontal circuitry. Journal of Neuroscience 33(10): 4584–4593. Crockett MJ, Clark L and Robbins TW (2009) Reconciling the role of Giedd JN, Blumenthal J, Jeffries NO, et al. (1999) Brain development serotonin in behavioral inhibition and aversion: Acute tryptophan during childhood and adolescence: A longitudinal MRI study. depletion abolishes punishment-induced inhibition in humans. Jour- Nature Neuroscience 2(10): 861–863. nal of Neuroscience 29(38): 11993–11999. Gillan CM, Morein-Zamir S, Kaser M, et al. (2014) Counterfactual Cullen KR, Gee DG, Klimes-Dougan B, et al. (2009) A preliminary study processing of economic action-outcome alternatives in obsessive- of functional connectivity in comorbid adolescent depression. Neu- compulsive disorder: Further evidence of impaired goal-directed roscience Letters 460(3): 227–231. behavior. Biological Psychiatry 75(8): 639–646. Cuthbert BN and Insel TR (2013) Toward the future of psychiatric diag- Giustino TF and Maren S (2015) The role of the medial prefrontal cortex nosis: The seven pillars of RDoC. BMC Medicine 11: 126–126. in the conditioning and extinction of fear. Frontiers in Behavioral Davis M (1998) Are different parts of the extended amygdala involved in Neuroscience 9: 298. fear versus anxiety? Biological Psychiatry 44(12): 1239–1247. Glenn CR, Klein DN, Lissek S, et al. (2012) The development of fear Davis M (2000) The role of the amygdala in conditioned and uncon- learning and generalization in 8–13 year-olds. Developmental Psy- ditioned fear and anxiety. In: Aggleton JP (ed.) The Amygdala: chobiology 54(7): 675–684. A Functional Analysis. New York: Oxford University Press, Gold AL, Brotman MA, Adleman NE, et al. (2016) Comparing brain pp. 213–288. morphometry across multiple childhood psychiatric disorders. Jour- Davis M (2001) Fear-Potentiated Startle in Rats: Current Protocols in nal of the American Academy of Child & Adolescent Psychiatry Neuroscience. Hoboken, NJ: John Wiley & Sons. 55(12): 1027–1037. Davis M, Walker DL, Miles L, et al. (2010) Phasic vs sustained fear in Goodkind M, Eickhoff SB, Oathes DJ, et al. (2015) Identification of a rats and humans: Role of the extended amygdala in fear vs anxiety. common neurobiological substrate for mental illness. JAMA Psy- Neuropsychopharmacology 35(1): 105–135. chiatry 72(4): 305–315. Dayan P and Huys QJM (2009) Serotonin in affective control. Annual Griebel G and Holmes A (2013) 50 years of hurdles and hope in anxio- Review of Neuroscience 32(2): 95–126. lytic drug discovery. Nature Reviews Drug Discovery 12(9): 667– De Wit SJ, Alonso P, Schweren L, et al. (2014) Multicenter voxel-based 687. morphometry mega-analysis of structural brain scans in obsessive- Grillon C, Ameli R, Woods SW, et al. (1991) Fear-potentiated startle in compulsive disorder. American Journal of Psychiatry 171(3): humans: Effects of anticipatory anxiety on the acoustic blink reflex. 340–349. Psychophysiology 28(5): 588–595. DiLuca M and Olesen J (2014) The cost of brain diseases: A burden or a Grillon C and Davis M (2007) Effects of stress and shock anticipation challenge? Neuron 82(6): 1205–1208. on prepulse inhibition of the startle reflex. Psychophysiology 34(5): Dodd HF, Vogt J, Turkileri N, et al. (2017) Task relevance of emotional 511–517. information affects anxiety-linked attention bias in visual search. Grupe DW and Nitschke JB (2011) Uncertainty is associated with biased Biological Psychology 122: 13–20. expectancies and heightened responses to aversion. Emotion 11(2): Do-Monte FH, Manzano-Nieves G, Quiñones-Laracuente K, et al. (2015) 413–424. Revisiting the role of infralimbic cortex in fear extinction with opto- Grupe DW and Nitschke JB (2013) Uncertainty and anticipation in anxi- genetics. Journal of Neuroscience 35(8): 3607–3615. ety: An integrated neurobiological and psychological perspective. Erk S, Abler B and Walter H (2006) Cognitive modulation of emotion Nature Reviews Neuroscience 14(7): 488–501. anticipation. European Journal of Neuroscience 24(4): 1227–1236. Gungor NZ and Paré D (2016) Functional heterogeneity in the bed Etkin A and Wager TD (2007) Functional neuroimaging of anxiety: A nucleus of the Stria Terminalis. Journal of Neuroscience 36(31): meta-analysis of emotional processing in PTSD, social anxiety dis- 8038–8049. order, and specific phobia. American Journal of Psychiatry 164(10): Gutman DA, Keifer OP Jr, Magnuson ME, et al. (2012) A DTI tractogra- 1476–1488. phy analysis of infralimbic and prelimbic connectivity in the mouse Etkin A, Egner T and Kalisch R (2011) Emotional processing in anterior using high-throughput MRI. Neuroimage 63(2): 800–811. cingulate and medial prefrontal cortex. Trends in Cognitive Sciences Guyer AE, Monk CS, McClure-Tone EB, et al. (2008) A developmental 15(2): 85–93. examination of amygdala response to facial expressions. Journal of Felix-Ortiz A, Burgos-Robles A, Bhagat N, et al. (2016) Bidirectional Cognitive Neuroscience 20(9): 1565–1582. modulation of anxiety-related and social behaviors by amygdala Hakamata Y, Lissek S, Bar-Haim Y, et al. (2010) Attention bias modifi- projections to the medial prefrontal cortex. Neuroscience 321: cation treatment: A meta-analysis toward the establishment of novel 197–209. treatment for anxiety. Biological Psychiatry 68(11): 982–990. Felix-Ortiz AC, Beyeler A, Seo C, et al. (2013) BLA to vHPC inputs Harlow HF and Harlow MK (1965) Chapter 8: The affectional sys- modulate anxiety-related behaviors. Neuron 79(4): 658–664. tems. In: Schrier AM, Stollnitz F and Harlow HF (eds) Behavior Fox AS and Shackman AJ (2017) The central extended amygdala in fear of Nonhuman Primates. Cambridge, MA: Academic Press, pp. and anxiety: Closing the gap between mechanistic and neuroimaging 287–334. research. Neuroscience Letters. Epub ahead of print 30 November. Harmer CJ, Cowen PJ and Goodwin GM (2011) Efficacy markers in DOI: 10.1016/j.neulet.2017.11.056. depression. Journal of Psychopharmacology 25(9): 1148–1158. Ganella DE and Kim JH (2014) Developmental rodent models of fear Harmer CJ, Goodwin GM and Cowen PJ (2009) Why do antidepres- and anxiety: From neurobiology to pharmacology. British Journal of sants take so long to work? A cognitive neuropsychological model Pharmacology 171(20): 4556–4574. of antidepressant drug action. British Journal of Psychiatry 195(2): Ganella DE, Barendse MEA, Kim JH, et al. (2017) Prefrontal-amygdala 102–108. connectivity and state anxiety during fear extinction recall in adoles- Herry C, Ciocchi S, Senn V, et al. (2008) Switching on and off fear by cents. Frontiers in Human Neuroscience 11: 584. distinct neuronal circuits. Nature 454(1): 600–606. Gee DG, Gabard-Durnam LJ, Flannery J, et al. (2013a) Early develop- Hess EH, Brown R, Galanter E, et al. (1962) New directions in psychology. mental emergence of human amygdala–prefrontal connectivity after Hettema JM, Prescott CA, Myers JM, et al. (2005) The structure of maternal deprivation. Proceedings of the National Academy of Sci- genetic and environmental risk factors for anxiety disorders in men ences 110(39): 15638–15643. and women. Archives of General Psychiatry 62(2): 182–189. 10 Brain and Neuroscience Advances Hübner C, Bosch D, Gall A, et al. (2014) Ex vivo dissection of optoge- Kim MJ, Loucks RA, Palmer AL, et al. (2011b) The structural and func- netically activated mPFC and hippocampal inputs to neurons in the tional connectivity of the amygdala: From normal emotion to patho- basolateral amygdala: Implications for fear and emotional memory. logical anxiety. Behavioural Brain Research 223(2): 403–410. Frontiers in Behavioral Neuroscience 8: 64. Klavir O, Genud-Gabai R and Paz R (2013) Functional connectivity Huys QJM, Maia TV and Frank MJ (2016) Computational psychiatry as between amygdala and cingulate cortex for adaptive aversive learn- a bridge from neuroscience to clinical applications. Nature Neurosci- ing. Neuron 80(5): 1290–1300. ence 19(3): 404–413. Klavir O, Prigge M, Sarel A, et al. (2017) Manipulating fear associations Insel T, Cuthbert B, Garvey M, et al. (2010) Research domain crite- via optogenetic modulation of amygdala inputs to prefrontal cortex. ria (RDoC): Toward a new classification framework for research Nature Neuroscience 20(6): 836–844. on mental disorders. American Journal of Psychiatry 167(7): Kling A and Steklis H (1976) A neural substrate for affiliative behav- 748–751. ior in nonhuman primates. Brain, Behavior and Evolution 13(2–3): Izquierdo A and Murray EA (2005) Opposing effects of amygdala and 216–238. orbital prefrontal cortex lesions on the extinction of instrumental Klumpp H, Fitzgerald DA and Phan KL (2013) Neural predictors and responding in macaque monkeys. European Journal of Neuroscience mechanisms of cognitive behavioral therapy on threat processing in 22(9): 2341–2346. social anxiety disorder. Progress in Neuro-psychopharmacology & Jacobs W and Nadel L (1985) Stress-induced recovery of fears and pho- Biological Psychiatry 45: 83–91. bias. Psychological Review 92(4): 512–531. Klumpp H, Fitzgerald JM, Kinney KL, et al. (2017) Predicting cognitive Janak PH and Tye KM (2015) From circuits to behaviour in the amyg- behavioral therapy response in social anxiety disorder with anterior dala. Nature 517(7534): 284–292. cingulate cortex and amygdala during emotion regulation. Neuroim- Johnson FK, Delpech J-C, Thompson GJ, et al. (2018) Amygdala hyper- age Clinical 15: 25–34. connectivity in a mouse model of unpredictable early life stress. Kotov R, Krueger RF, Watson D, et al. (2017) The hierarchical tax- Translational Psychiatry 8(1): 49. onomy of psychopathology (HiTOP): A dimensional alternative Jones PB (2013) Adult mental health disorders and their age at onset. to traditional nosologies. Journal of Abnormal Psychology 126(4): British Journal of Psychiatry 202(1): s5–s10. 454–477. Jones SA, Morales AM, Lavine JB, et al. (2017) Convergent neurobio- Krueger RF and Eaton NR (2015) Transdiagnostic factors of mental dis- logical predictors of emergent psychopathology during adolescence. orders. World Psychiatry 14(1): 27–29. Birth Defects Research 109(20): 1613–1622. Lahey BB, Applegate B, Hakes JK, et al. (2012) Is there a general factor Kaczkurkin A, Moore T, Calkins M, et al. (2017) Common and dissociable of prevalent psychopathology during adulthood? Journal of Abnor- regional cerebral blood flow differences associate with dimensions of mal Psychology 121(4): 971–977. psychopathology across categorical diagnoses. Molecular Psychiatry. Lahey BB, Rathouz PJ, Van Hulle C, et al. (2008) Testing structural mod- Epub ahead of print 19 September. DOI: 10.1038/mp.2017.174. els of DSM-IV symptoms of common forms of child and adolescent Kagan J, Reznick JS and Snidman N (1987) The physiology and psychol- psychopathology. Journal of Abnormal Child Psychology 36(2): ogy of behavioral inhibition in children. Child Development 58(6): 187–206. 1459–1473. Lahey BB, Zald DH, Perkins SF, et al. (2017) Measuring the hierarchical Kalin NH (2017) Mechanisms underlying the early risk to develop anxi- general factor model of psychopathology in young adults. Interna- ety and depression: A translational approach. European Neuropsy- tional Journal of Methods in Psychiatric Research 27(1). chopharmacology 27(6): 543–553. Laviolette SR, Lipski WJ and Grace AA (2005) A subpopulation of neu- Kalin NH, Fox AS, Kovner R, et al. (2016) Overexpressing corticotro- rons in the medial prefrontal cortex encodes emotional learning with pin-releasing factor in the primate amygdala increases anxious tem- burst and frequency codes through a dopamine D4 receptor-depen- perament and alters its neural circuit. Biological Psychiatry 80(5): dent basolateral amygdala input. Journal of Neuroscience 25(26): 345–355. 6066–6075. Kalin NH, Shelton SE and Davidson RJ (2007) Role of the primate orbi- LeDoux JE (2000) Emotion circuits in the brain. Annual Review of Neu- tofrontal cortex in mediating anxious temperament. Biological Psy- roscience 23(1): 155–184. chiatry 62(10): 1134–1139. Linnman C, Rougemont-Bücking A, Beucke JC, et al. (2011) Uncon- Kalisch R and Gerlicher AMV (2014) Making a mountain out of a mole- ditioned responses and functional fear networks in human classical hill: On the role of the rostral dorsal anterior cingulate and dorsome- conditioning. Behavioural Brain Research 221(1): 237–245. dial prefrontal cortex in conscious threat appraisal, catastrophizing, Linnman C, Zeidan MA, Furtak SC, et al. (2012a) Resting amygdala and worrying. Neuroscience & Biobehavioral Reviews 42: 1–8. and medial prefrontal metabolism predicts functional activation of Kalisch R, Wiech K, Critchley HD, et al. (2006) Levels of appraisal: A the fear extinction circuit. American Journal of Psychiatry 169(4): medial prefrontal role in high-level appraisal of emotional material. 415–423. Neuroimage 30(4): 1458–1466. Linnman C, Zeidan MA, Pitman RK, et al. (2012b) Resting cerebral Kessler R, Avenevoli S, Costello E, et al. (2012) Prevalence, persis- metabolism correlates with skin conductance and functional brain tence, and sociodemographic correlates of DSM-IV disorders in activation during fear conditioning. Biological Psychology 89(2): the national comorbidity survey replication adolescent supplement. 450–459. Archives of General Psychiatry 69(4): 372–380. Little JP and Carter AG (2013) Synaptic mechanisms underlying strong Kessler R, Demler R, Olfson F, et al. (2005) Prevalence and treatment of reciprocal connectivity between the medial prefrontal cortex and baso- mental disorders, 1990 to 2003. New England Journal of Medicine lateral amygdala. Journal of Neuroscience 33(39): 15333–15342. 352(24): 2515–2523. Livneh U and Paz R (2012) Amygdala-prefrontal synchronization under- Kim JJ and Fanselow MS (1992) Modality-specific retrograde amnesia of lies resistance to extinction of aversive memories. Neuron 75(1): fear. Science 256(5057): 675–677. 133–142. Kim MJ and Whalen PJ (2009) The structural integrity of an amygdala– Machado CJ and Bachevalier J (2008) Behavioral and hormonal prefrontal pathway predicts trait anxiety. Journal of Neuroscience reactivity to threat: Effects of selective amygdala, hippocampal 29(37): 11614–11618. or orbital frontal lesions in monkeys. Psychoneuroendocrinology Kim MJ, Gee DG, Loucks RA, et al. (2011a) Anxiety dissociates dorsal 33(7): 926–941. and ventral medial prefrontal cortex functional connectivity with the McTeague LM, Huemer J, Carreon DM, et al. (2017) Identification amygdala at rest. Cerebral Cortex 21(7): 1667–1673. of common neural circuit disruptions in cognitive control across Carlisi and Robinson 11 psychiatric disorders. American Journal of Psychiatry 174(7): Pine DS (2007) Research review: A neuroscience framework for pedi- 676–685. atric anxiety disorders. Journal of Child Psychology and Psychiatry Maslowsky J, Mogg K, Bradley BP, et al. (2010) A preliminary investiga- 48(7): 631–648. tion of neural correlates of treatment in adolescents with generalized Price JL and Drevets WC (2012) Neural circuits underlying the patho- anxiety disorder. Journal of Child and Adolescent Psychopharma- physiology of mood disorders. Trends in Cognitive Sciences 16(1): cology 20(2): 105–111. 61–71. Mechias M-L, Etkin A and Kalisch R (2010) A meta-analysis of Regev L, Tsoory M, Gil S, et al. (2012) Site-specific genetic manipula- instructed fear studies: Implications for conscious appraisal of threat. tion of amygdala corticotropin-releasing factor reveals its imperative Neuroimage 49(2): 1760–1768. role in mediating behavioral response to challenge. Biological Psy- Meyer A (2017) A biomarker of anxiety in children and adolescents: A chiatry 71(4): 317–326. review focusing on the error-related negativity (ERN) and anxiety Robinson OJ, Charney DR, Overstreet C, et al. (2012a) The adaptive across development. Developmental Cognitive Neuroscience 27(1): threat bias in anxiety: Amygdala–dorsomedial prefrontal cortex cou- 58–68. pling and aversive amplification. Neuroimage 60(1): 523–529. Milad MR and Quirk GJ (2002) Neurons in medial prefrontal cortex sig- Robinson OJ, Cools R and Sahakian BJ (2012b) Tryptophan depletion nal memory for fear extinction. Nature 420(6911): 70–74. disinhibits punishment but not reward prediction: Implications for Milad MR and Quirk GJ (2012) Fear extinction as a model for transla- resilience. Psychopharmacology 219(2): 599–605. tional neuroscience: Ten years of progress. Annual Review of Psy- Robinson OJ, Krimsky M, Lieberman L, et al. (2014) The dorsal medial chology 63: 129–151. prefrontal (anterior cingulate) cortex–amygdala aversive amplifica- Milad MR and Rauch SL (2007) The role of the orbitofrontal cortex tion circuit in unmedicated generalised and social anxiety disorders: in anxiety disorders. Annals of the New York Academy of Sciences An observational study. Lancet Psychiatry 1(4): 294–302. 1121: 546–561. Robinson OJ, Krimsky M, Lieberman L, et al. (2016) Anxiety-poten- Milad MR, Pitman RK, Ellis CB, et al. (2009) Neurobiological basis of tiated amygdala-medial frontal coupling and attentional control. failure to recall extinction memory in posttraumatic stress disorder. Translational Psychiatry 6(6): e833. Biological Psychiatry 66(12): 1075–1082. Robinson OJ, Overstreet C, Allen PS, et al. (2013b) The role of serotonin Milad MR, Quirk GJ, Pitman RK, et al. (2007) A role for the human dor- in the neurocircuitry of negative affective bias: Serotonergic modu- sal anterior cingulate cortex in fear expression. Biological Psychiatry lation of the dorsal medial prefrontal-amygdala ‘aversive amplifica- 62(10): 1191–1194. tion’ circuit. Neuroimage 78(1): 217–223. Mkrtchian A, Aylward J, Dayan P, et al. (2017) Modeling avoidance in Robinson OJ, Vytal K, Cornwell B, et al. (2013a) The impact of anxiety mood and anxiety disorders using reinforcement learning. Biological upon cognition: Perspectives from human threat of shock studies. Psychiatry 82(7): 532–539. Frontiers in Human Neuroscience 7: 203. Monk C, Nelson E, McClure E, et al. (2006) Ventrolateral prefrontal Rothbaum BO and Davis M (2003) Applying learning principles to the cortex activation and attentional bias in response to angry faces in treatment of post-trauma reactions. Annals of the New York Academy adolescents with generalized anxiety disorder. American Journal of of Sciences 1008: 112–121. Psychiatry 163(6): 1091–1097. Roy AK, Vasa RA, Bruck M, et al. (2008) Attention bias toward threat Morgan MA and LeDoux JE (1995) Differential contribution of dorsal in pediatric anxiety disorders. Journal of the American Academy of and ventral medial prefrontal cortex to the acquisition and extinc- Child & Adolescent Psychiatry 47(10): 1189–1196. tion of conditioned fear in rats. Behavioral Neuroscience 109(4): Roy-Byrne P (2015) Treatment-refractory anxiety; definition, risk fac- 681–688. tors, and treatment challenges. Dialogues in Clinical Neuroscience Nugent NR, Tyrka AR, Carpenter LL, et al. (2011) Gene–environment 17(2): 191–206. interactions: Early life stress and risk for depressive and anxiety dis- Rygula R, Clarke HF, Cardinal RN, et al. (2015) Role of central sero- orders. Psychopharmacology 214(1): 175–196. tonin in anticipation of rewarding and punishing outcomes: Effects Okon-Singer H and Aue T (2017) Neurocognitive mechanisms modulat- of selective amygdala or orbitofrontal 5-HT depletion. Cerebral ing attention bias in anxiety: Current perspectives. Biological Psy- Cortex 25(9): 3064–3076. chology: 122: 1–3. Santangelo AM, Roberts AC, Ferguson-Smith AC, et al. (2016) Novel Oliver BR and Plomin R (2007) Twins’ early development study (TEDS): primate model of serotonin transporter genetic polymorphisms asso- A multivariate, longitudinal genetic investigation of language, cog- ciated with gene expression, anxiety and sensitivity to antidepres- nition and behavior problems from childhood through adolescence. sants. Neuropsychopharmacology 41(9): 2366–2376. Twin Research and Human Genetics 10(1): 96–105. Sawyer SM, Azzopardi PS, Wickremarathne D, et al. (2018) The age of Passamonti L, Crockett MJ, Apergis-Schoute AM, et al. (2012) Effects adolescence. Lancet Child & Adolescent Health 2(9): 223–228. of acute tryptophan depletion on prefrontal-amygdala connectivity Schumann G, Loth E, Banaschewski T, et al. (2010) The IMAGEN while viewing facial signals of aggression. Biological Psychiatry study: Reinforcement-related behaviour in normal brain function and 71(1): 36–43. psychopathology. Molecular Psychiatry 15(12): 1128–1139. Pattwell SS, Liston C, Jing D, et al. (2016) Dynamic changes in neural Selden N, Everitt B, Jarrard L, et al. (1991) Complementary roles for the circuitry during adolescence are associated with persistent attenua- amygdala and hippocampus in aversive conditioning to explicit and tion of fear memories. Nature Communications 7: 11475. contextual cues. Neuroscience 42(2): 335–350. Perna G, Favaron E, Di Bella D, et al. (2005) Antipanic efficacy of parox- Sengupta A, Yau JO, Bressel PJ-RD, et al. (2017) Basolateral amygdala etine and polymorphism within the promoter of the serotonin trans- neurons maintain aversive emotional salience. Journal of Neurosci- porter gene. Neuropsychopharmacology 30(12): 2230–2235. ence 38(12): 2460–2417. Phelps EA, Delgado MR, Nearing KI, et al. (2004) Extinction learning in Senn V, Wolff SB, Herry C, et al. (2014) Long-range connectivity defines humans: Role of the amygdala and vmPFC. Neuron 43(6): 897–905. behavioral specificity of amygdala neurons. Neuron 81(2): 428–437. Phillips R and LeDoux J (1992) Differential contribution of amygdala Shackman AJ and Fox AS (2016) Contributions of the central extended and hippocampus to cued and contextual fear conditioning. Behav- amygdala to fear and anxiety. Journal of Neuroscience 36(31): ioral Neuroscience 106(2): 274–285. 8050–8063. Pinard CR, Mascagni F and McDonald AJ (2012) Medial prefrontal cor- Shackman AJ, Salomons TV, Slagter HA, et al. (2011) The integration tical innervation of the intercalated nuclear region of the amygdala. of negative affect, pain and cognitive control in the cingulate cortex. Neuroscience 205: 112–124. Nature Reviews Neuroscience 12(3): 154–167. 12 Brain and Neuroscience Advances Shackman AJ, Tromp DP, Stockbridge MD, et al. (2016) Dispositional Sylvester CM, Barch DM, Harms MP, et al. (2016) Early childhood negativity: An integrative psychological and neurobiological per- behavioral inhibition predicts cortical thickness in adulthood. spective. Psychol Bull 142(12): 1275–1314. Journal of the American Academy of Child & Adolescent Psychiatry Shin LM and Liberzon I (2009) The neurocircuitry of fear, stress, and 55(2): 122–129. anxiety disorders. Neuropsychopharmacology 35(1): 169–191. Terburg D, Morgan BE, Montoya ER, et al. (2012) Hypervigilance for Shou H, Yang Z, Satterthwaite TD, et al. (2017) Cognitive behavioral fear after basolateral amygdala damage in humans. Translational therapy increases amygdala connectivity with the cognitive con- Psychiatry 2(5): e115. trol network in both MDD and PTSD. Neuroimage Clinical 14(1): Thompson PM, Stein JL, Medland SE, et al. (2014) The ENIGMA con- 464–470. sortium: Large-scale collaborative analyses of neuroimaging and Sierra-Mercado D, Padilla-Coreano N and Quirk GJ (2011) Dissociable genetic data. Brain Imaging and Behavior 8(2): 153–182. roles of prelimbic and infralimbic cortices, ventral hippocampus, and Tovote P, Fadok JP and Lüthi A (2015) Neuronal circuits for fear and basolateral amygdala in the expression and extinction of conditioned anxiety. Nature Reviews Neuroscience 16(10): 317. fear. Neuropsychopharmacology 36(2): 529–538. Van Erp TG, Hibar DP, Rasmussen JM, et al. (2016) Subcortical brain Sotres-Bayon F, Sierra-Mercado D, Pardilla-Delgado E, et al. (2012) volume abnormalities in 2028 individuals with schizophrenia and Gating of fear in prelimbic cortex by hippocampal and amygdala 2540 healthy controls via the ENIGMA consortium. Molecular Psy- inputs. Neuron 76(4): 804–812. chiatry 21(4): 547–553. Strawn JR, Bitter SM, Weber WA, et al. (2012) Neurocircuitry of gen- Vidal-Gonzalez I, Vidal-Gonzalez B, Rauch SL, et al. (2006) Micro- eralized anxiety disorder in adolescents: A pilot functional neuro- stimulation reveals opposing influences of prelimbic and infralimbic imaging and functional connectivity study. Depression and Anxiety cortex on the expression of conditioned fear. Learning & Memory 29(11): 939–947. 13(6): 728–733. Sudlow C, Gallacher J, Allen N, et al. (2015) UK biobank: An open Vogt BA (2005) Pain and emotion interactions in subregions of the cin- access resource for identifying the causes of a wide range of gulate gyrus. Nature Reviews Neuroscience 6(7): 533–544. complex diseases of middle and old age. PLoS Medicine 12(3): Vos T, Allen C, Arora M, et al. (2016) Global, regional, and national e1001779. incidence, prevalence, and years lived with disability for 310 dis- Svihra M and Katzman MA (2004) Behavioural inhibition: A predictor of eases and injuries, 1990–2015: A systematic analysis for the Global anxiety. Paediatrics & Child Health 9(8): 547–550. Burden of Disease Study 2015. Lancet 388(10053): 1545–1602. Swartz JR and Monk CS (2013) The role of corticolimbic circuitry in Vytal KE, Overstreet C, Charney DR, et al. (2014) Sustained anxiety the development of anxiety disorders in children and adolescents. increases amygdala–dorsomedial prefrontal coupling: A mechanism In: Andersen SL and Pine DS (eds) The Neurobiology of Childhood. for maintaining an anxious state in healthy adults. Journal of Psy- Berlin: Springer, pp. 133–148. chiatry & Neuroscience 39(5): 321–329. Swartz JR, Carrasco M, Wiggins JL, et al. (2014) Age-related changes in White LK, Sequeira S, Britton JC, et al. (2017) Complementary features the structure and function of prefrontal cortex–amygdala circuitry in of attention bias modification therapy and cognitive-behavioral ther- children and adolescents: A multi-modal imaging approach. Neuro- apy in pediatric anxiety disorders. American Journal of Psychiatry image 86(2): 212–220. 174(8): 775–784. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Brain and Neuroscience Advances SAGE

The role of prefrontal–subcortical circuitry in negative bias in anxiety: Translational, developmental and treatment perspectives:

Loading next page...
 
/lp/sage/the-role-of-prefrontal-subcortical-circuitry-in-negative-bias-in-GKJiLjoufg

References (388)

Publisher
SAGE
Copyright
Copyright © 2022 by SAGE Publications Ltd and British Neuroscience Association, unless otherwise noted. Manuscript content on this site is licensed under Creative Commons Licenses
ISSN
2398-2128
eISSN
2398-2128
DOI
10.1177/2398212818774223
Publisher site
See Article on Publisher Site

Abstract

Anxiety disorders are the most common cause of mental ill health in the developed world, but our understanding of symptoms and treatments is not presently grounded in knowledge of the underlying neurobiological mechanisms. In this review, we discuss accumulating work that points to a role for prefrontal–subcortical brain circuitry in driving a core psychological symptom of anxiety disorders – negative affective bias. Specifically, we point to converging work across humans and animal models, suggesting a reciprocal relationship between dorsal and ventral prefrontal–amygdala circuits in promoting and inhibiting negative bias, respectively. We discuss how the developmental trajectory of these circuits may lead to the onset of anxiety during adolescence and, moreover, how effective pharmacological and psychological treatments may serve to shift the balance of activity within this circuitry to ameliorate negative bias symptoms. Together, these findings may bring us closer to a mechanistic, neurobiological understanding of anxiety disorders and their treatment. Keywords Anxiety, circuit, negative bias, prefrontal cortex Received: 19 December 2017; accepted: 9 April 2018 Introduction Anxiety disorders are the most common cause of mental illness in consist of shared symptoms which are similar manifestations of the developed world, with large social, economic and psychologi- relatively few underlying dimensions (Caspi et al., 2014; Clark cal impacts (DiLuca and Olesen, 2014; Shin and Liberzon, 2009; et al., 2017; Kaczkurkin et al., 2017; Kotov et al., 2017; Krueger Vos et al., 2016). A propensity towards the development of anxi- and Eaton, 2015; Lahey et al., 2012, 2017). ety disorders is heritable (Hettema et al., 2005), often begins in Recent efforts such as the Research Domain Criteria (RDoC; childhood or adolescence (Beesdo et al., 2009; Pine, 2007) and Insel et al., 2010), therefore, attempt to re-frame the investigation persists into adulthood (Copeland et al., 2014; Craske et al., of psychiatric disorders by advocating a trans-diagnostic approach 2017). It is estimated that close to one in four people will suffer focusing on the neurobiological mechanisms underpinning symp- from an anxiety disorder – including generalised anxiety (GAD), toms that cut across traditional categorical diagnoses. In particu- post-traumatic stress disorder (PTSD), social anxiety or phobias lar, one domain within the RDoC, Negative Valence Systems, – in their lifetime (Kessler et al., 2005, 2012), but currently avail- includes responses to aversive situations such as fear, anxiety, able psychological and pharmacological treatments are effective sustained threat and loss (/reward omission) that overlap with a for less than half of these individuals (Roy-Byrne, 2015; key concept from the clinical psychology literature – negative Community and Mental Health team, 2014) and progress in the affective bias. Negative biases in cognition are thought to promote discovery of anxiolytic drugs has been slow (Griebel and Holmes, and uphold key symptoms of many psychiatric conditions but are 2013). One reason for this treatment gap is that we have a limited understanding of the biological mechanisms by which anxiety Division of Psychology and Language Sciences, University College symptoms emerge or how these mechanisms are modulated by London, London, UK our current interventions. As such, we struggle to develop new Institute of Cognitive Neuroscience, University College London, treatments that can modulate known biological targets. Moreover, London, UK it is increasingly clear that our current diagnoses, based largely on self-reported symptoms, do not map clearly onto underlying biol- Corresponding author: ogy or indeed onto the latent structure of the self-reported symp- Oliver J. Robinson, Institute of Cognitive Neuroscience, University College toms themselves (Cuthbert and Insel, 2013; Kotov et al., 2017). London, Alexandra House, 17-19 Queen Square, London WC1N 3AZ, UK. Indeed, factor analyses suggest that many categorical disorders Email: o.robinson@ucl.ac.uk Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). 2 Brain and Neuroscience Advances especially prominent in anxiety disorders, perpetuated by antici- development and treatment (Aue and Okon-Singer, 2015; Cisler pation of – and uncertainty about – future events (Grupe and and Koster, 2010; Dodd et al., 2017). Nitschke, 2011, 2013). Although none of our current treatments or diagnoses are based on a mechanistic neurobiological understand- Animal models of negative bias ing of negative bias, recent work has begun to delineate the role that interactions between the prefrontal cortex (PFC) and subcor- Research falling within the Negative Valence Systems domain of tical regions such as the amygdala play in the manifestation of RDoC in animal models often makes a distinction between fear negative bias in anxiety. It is this circuitry that is the focus of the and anxiety. In the psychological literature, anxiety is defined as present review. a prolonged state of heightened anticipatory arousal, often prompted by distal or unpredictable threats (Davis et al., 2010). Fear, on the contrary, is conceptualised as a ‘fight or flight’ reac- Negative bias in anxiety tion and typically involves active defence against immediate Anxiety disorders are characterised by a general ‘negative bias’ threat, usually dissipating upon removal of the threat. In ‘real- in both attention and memory towards affectively negative world’ terms, a person’s reaction to a spider on the table in front (/threatening/aversive) information that promotes and upholds of them might elicit a fear response (which may be exacerbated the anxious state while having knock-on effects in a wide range in cases of phobia), whereas the knowledge that a spider might be of other cognitive functions (e.g. learning, inhibitory control; in the room but uncertainty of its location might elicit anxiety. Craske et al., 2017; Robinson et al., 2013a). It has been widely Elevations in both responses could of course play a key role in shown, for instance, that people with anxiety tend to interpret driving negative affective bias (although see Fox and Shackman, neutral information in a more negative light, have maladaptive 2017; Shackman and Fox, 2016, for suggestions of why this attention biases towards threat even when threats are not imme- explicit distinction between fear and anxiety might be problem- diately present or relevant (this may be particularly prominent in atic with regard to the underlying neurobiology). some subtypes of anxiety such as social anxiety disorder; Abend Animal models are particularly useful in the investigation of et al., 2017) and have a bias towards learning about negative the neural basis of anxiety and fear because of their cross-spe- information (Abend et al., 2017; Hakamata et al., 2010; Monk cies overlap in the neural circuitry underlying these processes et al., 2006; Okon-Singer and Aue, 2017; Roy et al., 2008). It is (Davis, 2000; LeDoux, 2000); cross-species functional homo- of course possible to break negative biases into specific subtypes logues of brain circuitry can inform translational research across of bias (for a comprehensive attempt to do this, see Grupe and humans and rodents and can provide testable models, even if the Nitschke, 2013), but here we aim to build a broad preliminary circuits themselves are not directly conserved across species. model across disparate animal and human experimental data, For instance, startle response, in which whole-body jump is typi- along with clinical data, with the goal of drawing holistic con- cally used as an index in rodents, is paralleled by an eye-blink clusions. Similarly, the term negative bias of course encom- response in humans (Davis, 2001) and is a reliable experimental passes myriad constructs, including a distinction made by many measure of aversive responding on some cognitive tasks researchers between fear and anxiety (Davis et al., 2010), as well (Aylward and Robinson, 2017) but not necessarily on others as subcategories of learned versus prepotent fears (LeDoux, (Bradford et al., 2015). This startle reflex is heightened by both 2000; Phelps et al., 2004). However, in this review, we broadly fear and anxiety states across rats and humans (Grillon and focus across the Negative Valence Systems domain of the RDoC Davis, 2007; Grillon et al., 1991), but the subcortical circuitry (Insel et al., 2010) in an attempt to identify common patterns and responsible may differ (Robinson et al., 2012a). Specifically, build a simple model of how negative bias is generated at the fear responses are associated with the central amygdala, anxiety neurobiological level. If we were to be too fine-grained in our responses are associated with the bed nucleus of the stria termi- definitions of negative bias, we will find ourselves with multiple nalis, and both fear and anxiety responses are associated with the non-overlapping studies, which would limit interpretation. basolateral amygdala (BLA; Davis et al., 2010; Sengupta et al., To demonstrate our aim of bridging disparate experimental 2017; Tovote et al., 2015), although this distinction has been literature and clinical utility, consider Beck’s (1967) early challenged and warrants further investigation (Gungor and Paré, theory of elevated negative bias (or ‘negative schemata’). In 2016; Shackman and Fox, 2016). Nevertheless, taken together, this framework, Beck proposed a cognitive triad illustrating a rodent and human work points to the clear role of subcortical cycle among a negative view of the world, the self and the regions, and the extended amygdala in particular (along with its future. This model encompasses wide-ranging cognitive func- inter-connections and external projections), in driving aversive/ tions from memory to attention but nonetheless forms the fear responding and hence negative bias (Boeke et al., 2017; basis of successful psychological treatments such as cogni- Campese et al., 2015, 2017; Sengupta et al., 2017; Terburg et al., tive-behavioural therapy (CBT), highlighting potential advan- 2012; Tovote et al., 2015). tages in taking a broad approach to linking cognitive research However, regions of the brain rarely, if ever, work in isolation. with clinical practice. Within the hierarchy of neural processing, these subcortical We therefore review converging evidence across humans and regions also interact with ‘higher’ cortical areas. This is perhaps animal models suggesting that negative bias may arise, at least in best illustrated by work exploring the weakening of learned aver- part, from activity within prefrontal regions and their interactions sive responses during extinction. During fear extinction, a cue with subcortical regions (Etkin and Wager, 2007; Shin and which previously indicated the onset of an aversive event no Liberzon, 2009). These circuits may underpin the ability to longer predicts a negative outcome, so the individual must ‘extin- engage or disengage attention from threats and may be critical to guish’ their original aversive response. Animal models of condi- understanding the mechanistic basis of negative bias as well as its tioned fear extinction indeed implicate subcortical regions such Carlisi and Robinson 3 as the BLA, but they extend the circuitry to encompass medial information flow between these regions in a circuit that drives the prefrontal cortical regions as well. In particular, within the rodent overall output. PFC, subdivisions of the infralimbic (IL) and the prelimbic (PL) Collectively, animal work therefore suggests a putative neural have been posited to play distinct roles in the expression and mechanism of negative bias; one (bi-directional PL–amygdala) extinction of conditioned fear, with the IL supporting fear extinc- circuit may serve to facilitate negative bias, while another (bi- tion as expressed by the amygdala and the PL conversely promot- directional IL–amygdala) circuit may serve to suppress negative ing fear expression as expressed by the amygdala (Klavir et al., bias (Calhoon and Tye, 2015). This simplified heuristic provides 2017; Morgan and LeDoux, 1995; Sierra-Mercado et al., 2011; a framework with which to consider neurobiological research in Vidal-Gonzalez et al., 2006). In monkeys, activity in the dorsal anxious humans. anterior cingulate (dACC) is correlated with the BLA during fear learning and memory acquisition (Klavir et al., 2013; Livneh and Paz, 2012), and in rodents, PL response to cues predicting an Neurobiological basis of negative bias aversive event increases post-fear learning (Burgos-Robles et al., in humans 2009). Using pharmacological or electrical stimulation and optogenetic approaches, it has been shown that, on the other Perhaps unusually for a symptom related to psychiatric disorder, hand, increased activity in the rodent IL predicts fear extinction negative bias in anxiety can be adaptive. For example, when one in the amygdala (Do-Monte et al., 2015; Klavir et al., 2017; is walking home late at night and hears an unexpected noise Milad and Quirk, 2002, 2012). Having said that, recent rodent down a dark alley, an appraisal of this situation as potentially work has suggested that this dissociation may not be as clear-cut threatening raises awareness, preparing the body’s fight or flight as previously thought. For example, the role of the IL in fear response in the event of immediate danger. In other words, nega- extinction has been challenged in optogenetics work showing tive bias in anxiety can promote harm avoidance. However, if this that extinction recall was intact after the silencing of IL neurons heightened anxiety and negative bias does not subside when one and that stimulation of ventromedial prefrontal cortex (vmPFC) is subsequently safe at home, this response becomes maladaptive inputs to the amygdala facilitated extinction memory formation and can impair daily functioning (i.e. it transitions into a patho- but not retrieval (Bukalo et al., 2015; Do-Monte et al., 2015). logical state). Thus, it was suggested by the pioneers of CBT that Moreover, the distinction between PL/dorsal and IL/ventral pre- a biased appraisal of threat (i.e. negative bias) leading to catastro- frontal regions being responsible for fear expression and suppres- phising or excessive worry is a central characteristic of anxiety sion, respectively, has been challenged (Giustino and Maren, disorders (Beck and Clark, 1997). Dorsal regions of the PFC 2015). For instance, rodent work has shown that these regions (dorsomedial PFC (dmPFC) and dACC) seem to be associated have structurally similar projections to the amygdala (Cho et al., with this behavioural response at the neural level; these regions 2013; Gutman et al., 2012; Hübner et al., 2014; Pinard et al., are activated during conscious threat appraisal in healthy indi- 2012), and functionally dichotomous distinctions between these viduals and have been shown to be overactive during threat regions have been shown in the opposite direction than was ini- appraisal in pathological anxiety (for review, see Kalisch and tially postulated (Chang et al., 2010). Gerlicher, 2014). This rodent research nevertheless highlights a key potential Consistent with the animal work highlighted above, patho- mechanism of negative bias; namely that the overall expression logical negative bias in humans may in fact result in part from an of aversive responding may be held in the balance of opposing inability to extinguish conditioned fear responses driven in turn circuitry. Thus, whether negative bias is expressed or dampened by this altered, worry-related PFC responding (Milad et al., 2007; may depend on whether one of these circuits is able to override Rothbaum and Davis, 2003). Healthy subjects show increased the other, with sub-regions of the medial prefrontal cortex activation in the vmPFC during acquisition and retrieval of (mPFC) playing a key role in arbitrating this response. Indeed, extinction (Kalisch et al., 2006; Milad and Rauch, 2007; Phelps across a range of paradigms in rodents, the PFC has been shown et al., 2004), which has led to the suggestion that this region in to play a regulatory role over BLA activation during fear expres- the human brain might be functionally (albeit perhaps not struc- sion, social interaction and anxiety-related behaviours (Bickart turally) homologous to the rodent IL. On the other hand, patients et al., 2014; Bremner, 2004; Davis, 1998; Felix-Ortiz et al., 2013; with anxiety disorders have shown reduced activation in the Janak and Tye, 2015; Kling and Steklis, 1976). vmPFC along with increased activation in the dACC, leading to That said, the implicit implication that the PFC is a ‘top- the suggestion that the dACC might be functionally homologous down’ regulator of the amygdala during fear extinction should be to the rodent PL (Milad et al., 2009). challenged. Optogenetics research in rodents has shown bi-direc- Consistent with this proposition, at the neural level, the dACC tional effects of modulating BLA projections to the IL and PL and adjacent dmPFC have been implicated in the appraisal and during a number of behavioural assays assessing anxiety-like expression of fear (Etkin et al., 2011; Vogt, 2005), as well as the behaviour (Courtin et al., 2014; Felix-Ortiz et al., 2016; Herry anticipation of emotional stimuli (Erk et al., 2006). Moreover, et al., 2008; Laviolette et al., 2005). Specifically, BLA activity Kalisch and Gherlicher (2014) argue that the dACC/dmPFC can projecting up to PL regions is increased during fear conditioning be further subdivided into an anterior part, the rostral dACC/ (Senn et al., 2014), while fear extinction also enhances activity in dmPFC and a posterior part, with the rostral but not posterior part BLA projections up to the IL (Milad and Quirk, 2002; Senn et al., implicated in conscious threat appraisal and worry (Kalisch and 2014). Moreover, recordings from non-human primates further Gerlicher, 2014; Mechias et al., 2010). Thus, activity in dorsal support this bi-directional effect during fear learning (Klavir PFC regions is broadly associated with increased negative bias. et al., 2013). In other words, it is not so much that the PFC ‘regu- The dACC has also been linked to the adaptive control of behav- lates’ the amygdala but rather the reciprocal relationship of iour as well as the risk of development of anxiety disorders 4 Brain and Neuroscience Advances (Cavanagh et al., 2017; Goodkind et al., 2015; McTeague et al., responding (as indexed by a threat-by-valence interaction in reac- 2017; Meyer, 2017). tion times, driven by a valence-specific reduced reaction time to Regarding ventral regions, studies in healthy adults (Bishop, fearful faces under threat vs safe conditions), suggesting a key 2007) as well as adults (Etkin and Wager, 2007; Milad et al., mediating role for dACC/dmPFC–amygdala circuitry in driving 2007; Price and Drevets, 2012) and children and adolescents negative bias. Critically, coupling within this same circuitry was (Guyer et al., 2008; Monk et al., 2006; Strawn et al., 2012) with shown to be elevated at baseline in individuals with an anxiety anxiety disorders have shown abnormal function in orbitofrontal disorder (Robinson et al., 2014; in the absence of induced anxi- cortex (OFC) and ventrolateral PFC (VLPFC). The cause and ety), suggesting that the same circuitry which can be selectively effect of such abnormalities have been studied in non-human pri- engaged and disengaged in healthy controls is more persistently mates through lesions to the anterior OFC and VLPFC (Agustín- engaged in patients with clinical anxiety, thereby providing a Pavón et al., 2012; Izquierdo and Murray, 2005; Kalin et al., route by which adaptive anxiety can transition into a maladaptive 2007; Machado and Bachevalier, 2008), with findings largely state. Across both studies, however, the correlation between the showing increased anxiety during fear conditioning paradigms amygdala and dorsal cortical regions was positive. In other words, when these regions are lesioned. Thus, broadly speaking, activity activity in the dorsal cortical regions increases as activity increases in ventral cortical regions is associated with reduced negative in the amygdala and vice versa. The role that this circuit seems to bias (although it should be noted that this association may not be play in threat responding therefore appears somewhat analogous as consistent as previously thought (cf. Shackman et al., 2011). to the role of the PL in rodents. Thus, a human functional homo- The hippocampus is another structure that has been hypothe- logue of the rodent PL–amygdala circuit may drive increased sised to play a critical role in the pathophysiology of anxiety. threat responding and negative affective biases in anxiety Specifically, this region is a key mediator of the acquisition and disorders. expression of learned fear, as demonstrated by a number of early However, rodent work has also highlighted the contrasting lesion studies showing that hippocampal lesions dampened fear role of the inhibitory IL circuit (Kim et al., 2011a). To this end, response to previous learned associations (Kim and Fanselow, another study in humans (Vytal et al., 2014) expanded this puta- 1992; Phillips and LeDoux, 1992; Selden et al., 1991). Studies in tive circuitry to encompass a reciprocal inhibitory circuit. both human and rodents suggest that this region integrates con- Specifically, inducing anxiety during an adapted resting-state textual information during fear conditioning and may regulate scan replicated positive dmPFC–amygdala coupling, but at the context-dependent recall after extinction (Giustino and Maren, same time enhanced negative coupling between a ventral medial 2015). The rodent PL and IL receive excitatory inputs from both prefrontal region and the amygdala. In other words, while the dorsal and ventral hippocampus (Little and Carter, 2013), and increased dorsal activation was associated with increased amyg- it has been suggested that, similar to the amygdala, these projec- dala activation, increased ventral activation was associated with tions may inhibit downstream mPFC outputs (Sotres-Bayon decreased amygdala activity (Vytal et al., 2014). Earlier positron et al., 2012). In humans, Linnman et al. (2011) demonstrated that emission tomography (PET) studies in humans have shown simi- fear (elicited by electric shock expectation) was associated with larly contrasting relationships between prefrontal and subcortical increased connectivity between the hippocampus and the vmPFC, regions (Linnman et al., 2012a, 2012b). For example, during and decreased connectivity between the hippocampus and the red extinction training, resting amygdala metabolism positively pre- nucleus midbrain region, suggesting that the hippocampus may dicted vmPFC activation and negatively predicted dACC activa- facilitate a switch between what they term a ‘fear’ network and a tion, but during extinction recall, these relationships were in the ‘resting’ network. opposite direction (Linnman et al., 2012a). In a rodent study However, as highlighted by the rodent literature above, nega- investigating the impact of early-life environmental stress, tive bias is not so much driven by regions acting in isolation. Johnson et al. (2018) showed that stress was related to increased Rather, it is the cortical–subcortical circuitry that is important for amygdala–PFC and amygdala–hippocampus coupling and that anxiety response. For example, Kalin et al. (2016) used a viral this connectivity was related to anxiety-like behaviours in a vector approach in primates to demonstrate a relationship between translational model of early-life stress. Thus, these studies dem- overexpression of corticotropin-releasing hormone (CRH) in the onstrate that the relationship between distinct mPFC regions may dorsal amygdala and increased defensive behaviour during expo- have opposing effects on aversive responding. Another study of sure to threat. Moreover, this link between metabolism and behav- resting-state functional connectivity in healthy humans showed iour has also been observed in rodents and was associated with that those individuals who reported high levels of anxiety were functional connectivity between the dorsal amygdala and OFC characterised by negatively correlated amygdala–vmPFC con- (Regev et al., 2012). To this end, connectivity between the dACC/ nectivity, while this connectivity was positively correlated in dmPFC and the amygdala has been implicated in the pathophysi- those reporting low levels of anxiety (Kim et al., 2011b). ology of anxiety in humans. Structurally, the integrity of white Moreover, amygdala–dmPFC connectivity was negatively cor- matter tracts between the amygdala and the PFC has been shown related only in those reporting low anxiety. More dorsal regions to predict individual differences in trait anxiety (Kim and Whalen, of the PFC (like the PL in rodents) may increase aversive 2009). Functionally, Robinson et al. (2012a) studied the role of responding, while more ventral regions (like the IL in rodents) these regions in negative bias during induced anxiety in healthy may reduce aversive responding. Moreover, the nature of these individuals. They found that connectivity increased during the functional imaging connectivity analyses means that they are processing of threatening stimuli (fearful faces) selectively in the non-directional. In other words, it is not possible to say whether context of induced anxiety. Moreover, the strength of this connec- one region is driving the other – it is simply a correlation. Given tivity was positively correlated with participants’ subjective rat- the bi-directional nature of the rodent work highlighted above, ings of anxiety, as well as the extent of negative bias in behavioural these circuits therefore should not be considered ‘top-down’ or Carlisi and Robinson 5 ‘bottom-up’; rather, the overall reciprocal cortical–subcortical One influential idea common to the human and rodent devel- interaction likely drives the ultimate behavioural expression. opmental literature is that learned fear associations (i.e. memo- Together, these findings highlight the value of translational ries) from early life are important contributing factors to the research. A model of cortical–subcortical interactions during subsequent development of anxiety disorders (Britton et al., negative bias inspired by rodent work provides a framework 2011; Glenn et al., 2012; Jacobs and Nadel, 1985). Cross-species within which to consider the role of neural circuitry in negative animal work (Harlow and Harlow, 1965; Hess et al., 1962) has bias in humans. shown that fear learning is characterised by approach behaviour (such as maternal attachment or odour approach) in infants, but is characterised by almost diametrically opposed avoidance behav- Development iour (such as maternal or odour avoidance or avoidance of ele- vated/open areas in typical rodent paradigms) in adults (for an The work reviewed above therefore suggests that medial prefron- extensive review, see Ganella and Kim, 2014). This also suggests tal–amygdala interactions may drive the negative bias symptoms that at some point during development, there is a change in the that are a core feature of anxiety disorders. However, how these underlying neurobiology promoting this behaviour. mechanisms develop and persist across the lifespan remains Integrating this within the circuitry framework of the pre- unclear. If we want to target these symptoms and intervene early, sent review, Chan et al. (2011) inactivated PL in juvenile, pre- it is important to determine when and how alterations to these adolescent and adult rats and found that PL inactivation circuits emerge. significantly reduced freezing behaviour, as would be pre- Within the general population, pathological anxiety com- dicted by the above reviewed evidence, but that it only did so monly emerges during childhood or adolescence and reflects a in adolescent and adult rats, suggesting that the role of differ- combination of genetic factors and early-life experiences (Pine, ent medial prefrontal regions in negative bias changes across 2007). ‘Anxious temperament’ is considered to be a stable trait development. In other words, differences in fear responding, across time, and those with extreme levels of such traits are at a mediated by amygdala–medial prefrontal pathways, may par- higher risk for developing clinical or pathological anxiety tially be a result of a more protracted course of development (Arnaudova et al., 2013; Jones, 2013; Nugent et al., 2011). and reorganisation in these cortical–subcortical pathways Similarly, the stable traits of ‘behavioural inhibition’, a tempera- (Arruda-Carvalho et al., 2017; Ganella and Kim, 2014; ment characterised by a tendency to withdraw from new situa- LeDoux, 2000; Pattwell et al., 2016). Similar developmental tions (Kagan et al., 1987; Svihra and Katzman, 2004) and, more changes in prefrontal–subcortical negative bias circuitry are broadly, ‘dispositional negativity’ (Shackman et al., 2016) are also seen in humans; in typically developing humans, mPFC– thought to be early phenotypes of anxiety disorders. There is evi- amygdala connections are immature during childhood and dence that anxiety-related amygdala abnormalities and affected strengthen to adult levels during adolescence (Gee et al., top-down prefrontal regulation originate early in development 2013a, 2013b), and structural changes in white matter have (Clauss and Blackford, 2012; Kalin, 2017). Moreover, it has been been shown to mediate amygdala function in adolescents estimated that 50% of children showing increased behavioural (Swartz et al., 2014). Moreover, early perturbations in medial inhibition in childhood will later develop stress-related psycho- prefrontal circuitry have been implicated in the development of pathology (Clauss and Blackford, 2012). This is paralleled by anxiety and depression. For example, a preliminary study in findings of reduced amygdala–dorsolateral prefrontal cortex adolescents with depression found that patients had decreased (dlPFC) coupling in preadolescent children diagnosed with an functional connectivity in a subgenual (ventral) anterior cingu- anxiety disorder as well as in young non-human primates with late (ACC)-based network compared to healthy adolescents elevated levels of traits related to anxious temperament (includ- (Cullen et al., 2009). Moreover, negative coupling within ing heightened behavioural inhibition; Birn et al., 2014). vmPFC–amygdala circuitry during fear extinction was recently Although more longitudinal studies are needed to confirm this, shown only in adults and not adolescents (Ganella et al., 2017). this evidence suggests that rapid changes in the mPFC and the These ventral regions may reflect overlapping human homo- later maturation of amygdala–cortical connections during adoles- logues of the rodent IL. Thus, a developmental delay in the cence, a period recently suggested to encompass 10–24 years of ability to engage the circuitry that can dampen negative bias age (Sawyer et al., 2018), may contribute to the emergence of might explain the emergence of anxiety disorders during ado- anxiety during a specific developmental window (Andersen, lescence. This work is in its infancy, but the concept of recipro- 2003; Casey et al., 2008). Indeed, prospective studies in humans cal cortical–subcortical circuits again provides a framework (Giedd et al., 1999; Jones et al., 2017; Kalin, 2017; Swartz and with which to consider the emergence of anxiety and negative Monk, 2013) as well as rodent studies (Arruda-Carvalho et al., bias during development. 2017; Cohen et al., 2013; Gee et al., 2013a; Pattwell et al., 2016) have shown that this period constitutes a window of heightened risk for the development of anxiety. However, vast structural Treatment brain changes have also been observed during childhood, sug- gesting that children are subject to a heightened vulnerability to If prefrontal–subcortical circuitry is critical in the development environmental impacts which may influence the development of and manifestation of negative bias, then modulation of this cir- anxiety even before the onset of adolescence. Indeed, behaviour- cuitry should serve to modify negative affective biases and hence ally inhibited temperament has been observed in young children treat symptoms. The first-line treatments for clinical anxiety are who later develop anxiety, with similar neural circuitry altera- serotonergic medication and psychological therapy. Emerging tions linking these phenotypes (Buzzell et al., 2017; Gold et al., evidence suggests that successful response to both types of treat- 2016; Sylvester et al., 2016). ment may also depend on this prefrontal–subcortical circuitry. 6 Brain and Neuroscience Advances assessed whether CBT combined with attention bias modifica- The role of serotonin in pharmacological tion therapy (ABMT) was more clinically effective than CBT treatment alone and whether this treatment response could be predicted Serotonin (5-hydroxytryptamine (5-HT)) has long been impli- through pre-treatment amygdala-based functional connectivity cated in the neuropsychopharmacology of anxiety (Dayan and (White et al., 2017). This study found that patients differed from Huys, 2009; Harmer et al., 2009, 2011), largely because selective controls in amygdala–insula connectivity on a threat attention serotonin reuptake inhibitors (SSRIs) are the most common and task. Moreover, while both CBT groups showed clinical improve- effective pharmacological treatment for anxiety disorders ment, the combined CBT + ABMT group showed the greatest (Harmer et al., 2009, 2011). It is thought that serotonin plays a reduction in symptoms and that baseline amygdala functional particular role in maintaining the balance between the processing connectivity differentially predicted the level of treatment of appetitive and aversive information (Cools et al., 2008; response in patients. Crockett et al., 2009; Robinson et al., 2012b) and more precisely However, whether these changes in cortical–subcortical cir- in the inhibition of PFC-linked neural circuitry important for cuits are driven by CBT, or whether they simply reflect reduced driving negative bias (Crockett et al., 2009; Robinson et al., overall anxiety and negative bias per se, is unclear. To this end, 2013b). basic research has attempted to determine causality. Specifically, The impact of serotonin in healthy humans can be studied by it has been shown that in healthy individuals, simple attentional acute tryptophan depletion – a dietary manipulation that tempo- instruction can alter the engagement of affective-bias-related rarily reduces serotonin levels (Crockett et al., 2012). Reduced dmPFC–amygdala circuitry (Robinson et al., 2016). When sub- serotonin has been shown to increase positive coupling within jects are instructed to pay attention to neutral aspects of com- the same circuit shown to be elevated by induced anxiety (Vytal pound cues (rather than the affectively salient components of the et al., 2014) and at baseline in individuals with an anxiety disor- same cues), anxiety-induced amygdala–dmPFC coupling (as der (Robinson et al., 2014), suggesting that serotonergic drugs seen above; Robinson et al., 2012a) is down-regulated. This sug- (which putatively elevate serotonin availability) may work by gests that psychological treatments such as CBT may reduce reducing activity within this dorsal prefrontal circuit (Robinson negative bias by down-regulating the dorsal PFC–amygdala cir- et al., 2013b), thus reducing negative bias. By contrast, a study cuitry that promotes negative bias. In the context of threat pro- using a different paradigm showed that tryptophan depletion can cessing, there has been limited work showing whole-brain also decrease coupling between the amygdala and a more ventral increased ventrolateral prefrontal activation in anxious youth prefrontal region (Passamonti et al., 2012). Moreover, direct who underwent CBT relative to controls (Maslowsky et al., reductions in ventrally located orbitofrontal serotonin in the mar- 2010), as well as reduced dorsomedial prefrontal activation post- moset can increase negative bias (Rygula et al., 2015). Within the CBT relative to pre-CBT in individuals with social anxiety framework described above, this suggests that serotonin can also (Klumpp et al., 2013). Nevertheless, the role that CBT plays in serve to promote ventral PFC circuits that inhibit aversive pro- dorsal versus ventral prefrontal–amygdala circuitry in humans cessing while inhibiting dorsal PFC circuits that promote aver- has not been systematically studied. Moreover, recovery rates of sive responding. As such, pharmacological treatments may work patients with anxiety undergoing psychological treatment are less by restoring the balance between the circuits that, respectively, than 50% (Community and Mental Health team, 2014), so it is promote and inhibit negative bias. Recent work also suggests that plausible that these mechanisms are again only relevant in a sub- the influence of serotonergic drugs on this circuitry might be set of patients. mediated by genetic factors (Perna et al., 2005; Santangelo et al., 2016), which may in turn explain why such medications only work for a subset of anxious patients. Conclusions and future directions In this review, we have outlined evidence across animals and humans suggesting that bi-directional prefrontal–subcortical cir- Psychological treatment cuits and their interactions may drive elevated aversive process- CBT is the most common psychological intervention used to treat ing, or negative bias, in anxiety. Specifically, we point to anxiety and is based on the premise that negative biases in converging evidence within the Negative Valance Systems thoughts and actions can be shifted through cognitive reappraisal domain of the RDoC which suggests that ventral PFC–subcortical and emotion regulation strategies (Beck and Clark, 1997). There circuitry in humans may be associated with reduced negative have been numerous studies (see review by Brooks and Stein, bias, while more dorsal PFC–subcortical circuitry may be associ- 2015) which suggest that CBT modulates prefrontal–subcortical ated with increased negative bias. Moreover, we provide evi- interactions. Indeed, baseline medial prefrontal and amygdala dence suggesting that the emergence of anxiety in adolescence activity might even predict treatment response to CBT in anxiety may be a result of differential developmental trajectories of these (Klumpp et al., 2017). For instance, Shou et al. (2017) showed circuits and that both pharmacological and psychological inter- that functional connectivity between the amygdala and the ventions might be effective by modulating the overall balance of fronto-parietal network increased in patients with major depres- these circuits in driving negative affective bias. These findings sive disorder (MDD) or PTSD who underwent a course of CBT are summarised in Figure 1. compared to controls, supporting a mechanism by which this cir- Nevertheless, it is still unclear exactly how we bridge the gap cuitry may interact with psychological intervention (although it between brain and behaviour. Although we can associate these should be noted that this study did not include a patient group that circuits with negative bias, we cannot yet say how exactly the did not undergo CBT, so the specificity of these results is difficult underlying neuronal activity is translated into observable behav- to quantify). Similarly, a study of adolescents with anxiety iour. One particularly promising avenue in this regard is the Carlisi and Robinson 7 data (Thompson et al., 2014). Moreover, the UK Biobank (Sudlow et al., 2015) is a consortium across 22 research centres in the United Kingdom with genetic and longitudinal physical health and behavioural data on over 500,000 participants, all of which has been made open access. These are early efforts, par- ticularly in the field of anxiety disorders, but promising mega- and meta-analyses have already come out of such efforts in other fields of psychiatry such as obsessive–compulsive disorder and schizophrenia (Boedhoe et al., 2016; De Wit et al., 2014; Van Erp et al., 2016). Finally, it is important to investigate how these effects change over time. That is, are these mechanisms stable, or do they change across development to influence symptom onset and persistence? Longitudinal studies are critical for understanding these ques- tions. There have been longitudinal studies investigating brain changes over time in adolescents (e.g. the IMAGEN study; Schumann et al., 2010), but this investigation needs to be scaled up to larger populations and multiple time points and age ranges if we are to truly understand the developmental changes that occur across the life course of anxiety disorders. One promising example of this work currently underway is the Adolescent Brain Cognitive Development study (ABCD; https://abcdstudy.org/ index.html), which is the largest long-term longitudinal study of brain development in the United States, currently in the process of collecting biological and behavioural data on over 10,000 chil- dren aged 9–10. Similarly, to gain an understanding of the under- lying genetic contributions of anxiety, it is important to investigate Figure 1. Schematic summarising findings and proposed simplified the extent to which certain features and symptoms are heritable. model of negative affective bias in anxiety. Bi-directional excitatory This can be achieved through longitudinal twin studies (e.g. the connections between dorsal regions of the mPFC/ACC and the amygdala Twins Early Development Study (TEDS; Oliver and Plomin, promote negative bias, while inhibitory connections between ventral 2007) and the Tennessee Twin Study (Lahey et al., 2008)), but regions coupled with the amygdala inhibit negative bias. The ventral many of the existing studies do not focus on brain imaging due to inhibitory circuit may only fully develop in adulthood, meaning limited time and resources and the high cost involved in neuroim- that adolescence is a period of high vulnerability to negative bias. aging research. Regardless, observational population-based stud- Successful treatments (SSRIs and CBT) may be effective via promotion ies are an important complimentary approach to the small-scale of the ventral circuit and inhibition of the dorsal circuit. case–control designs more frequently implemented in neuroim- ACC: anterior cingulate cortex; mPFC: medial prefrontal cortex; CBT: cognitive- aging research on anxiety. behavioural therapy; SSRI: selective serotonin reuptake inhibitor. In conclusion, work has begun to delineate overlapping neu- ral networks involving the PFC and subcortical regions includ- ing the amygdala that may drive aversive responding and nascent field of computational psychiatry, which attempts to negative bias in both animals and humans. There is also promis- bridge the gap between brain activation and observable symp- ing evidence that pharmacological and psychological interven- toms (Huys et al., 2016). Specifically, using mathematical theo- tions can shape this circuitry and hence ameliorate negative ries of cognition and human behaviour, computational psychiatry affective bias. Future research should expand these findings to aims to objectively quantify the calculations generated by neu- larger populations and investigate how these neural underpin- rons which shape behaviour (Huys et al., 2016). This work is in nings arise in childhood/adolescence and change over time to its infancy but has begun to delineate the computational basis of shape behaviour. common symptoms in anxiety linked to negative bias, such as avoidance (Mkrtchian et al., 2017), risk aversion (Charpentier et al., 2017) and goal-directed behaviour (Carlisi et al., 2017; Declaration of conflicting interests Gillan et al., 2014). The author(s) declared no potential conflicts of interest with respect to Furthermore, if we are to understand current findings in a the research, authorship and/or publication of this article. truly generalisable context, it is critical to investigate these mech- anisms in large-scale populations. Cohort studies are an ideal Funding way to examine these questions at the population level, often C.O.C. is supported by a Wellcome Trust Sir Henry Wellcome Postdoctoral sampling from a diverse community of individuals. Data sharing Fellowship (206459/Z/17/Z) and O.J.R. is supported by a Medical efforts have attempted to address this. For example, the ENIGMA Research Council Career Development Award (MR/K024280/1). consortium is an international collaboration of research centres which aims to combine neuroimaging and genetic datasets from ORCID iD sites around the world in an attempt to amass sample sizes large enough to detect very small effects in brain imaging and genetic Christina O. Carlisi https://orcid.org/0000-0002-0942-8586 8 Brain and Neuroscience Advances Buzzell GA, Troller-Renfree SV, Barker TV, et al. (2017) A neurobe- References havioral mechanism linking behaviorally inhibited temperament and Abend R, de Voogd L, Salemink E, et al. (2017) Association between later adolescent social anxiety. Journal of the American Academy of attention bias to threat and anxiety symptoms in children and adoles- Child & Adolescent Psychiatry 56(12): 1097–1105. cents. Depression and Anxiety 35(3): 229–238. Calhoon GG and Tye KM (2015) Resolving the neural circuits of anxiety. Agustín-Pavón C, Braesicke K, Shiba Y, et al. (2012) Lesions of ventro- Nature Neuroscience 18: 1394–1404. lateral prefrontal or anterior orbitofrontal cortex in primates heighten Campese VD, Gonzaga R, Moscarello JM, et al. (2015) Modulation of negative emotion. Biological Psychiatry 72(4): 266–272. instrumental responding by a conditioned threat stimulus requires Andersen SL (2003) Trajectories of brain development: Point of vulner- lateral and central amygdala. Frontiers in Behavioral Neuroscience ability or window of opportunity? Neuroscience & Biobehavioral 9: 293. Reviews 27(1–2): 3–18. Campese VD, Soroeta JM, Vazey EM, et al. (2017) Noradrenergic regu- Arnaudova I, Krypotos A-M, Effting M, et al. (2013) Individual differ- lation of central amygdala in aversive pavlovian-to-instrumental ences in discriminatory fear learning under conditions of ambiguity: transfer. eNeuro 4(5): pii. A vulnerability factor for anxiety disorders? Frontiers in Psychology Carlisi CO, Norman L, Murphy CM, et al. (2017) Shared and disorder- 4: 298. specific neurocomputational mechanisms of decision-making in Arruda-Carvalho M, Wu W-C, Cummings KA, et al. (2017) Optogenetic autism spectrum disorder and obsessive-compulsive disorder. Cere- examination of prefrontal-amygdala synaptic development. Journal bral Cortex 27(12): 5804–5816. of Neuroscience 37(11): 2976–2985. Casey B, Jones RM and Hare TA (2008) The adolescent brain. Annals of Aue T and Okon-Singer H (2015) Expectancy biases in fear and anxiety the New York Academy of Sciences 1124(1): 111–126. and their link to biases in attention. Clinical Psychology Review 42: Caspi A, Houts RM, Belsky DW, et al. (2014) The p factor: One gen- 83–95. eral psychopathology factor in the structure of psychiatric disorders? Aylward J and Robinson OJ (2017) Towards an emotional ‘stress test’: Clinical Psychological Science 2(2): 119–137. A reliable, non-subjective cognitive measure of anxious responding. Cavanagh JF, Meyer A and Hajcak G (2017) Error-specific cognitive Scientific Reports 7: 40094. control alterations in generalized anxiety disorder. Biological Psy- Beck AT (1967) Depression: Clinical, experimental, and theoretical chiatry 2(5): 413–420. aspects. Philadelphia: University of Pennsylvania Press. Chan T, Kyere K, Davis BR, et al. (2011) The role of the medial prefron- Beck AT and Clark DA (1997) An information processing model of tal cortex in innate fear regulation in infants, juveniles, and adoles- anxiety: Automatic and strategic processes. Behaviour Research and cents. Journal of Neuroscience 31(13): 4991–4999. Therapy 35(1): 49–58. Chang C-H, Berke JD and Maren S (2010) Single-unit activity in the Beesdo K, Knappe S and Pine DS (2009) Anxiety and anxiety disorders medial prefrontal cortex during immediate and delayed extinction of in children and adolescents: Developmental issues and implications fear in rats. PLoS ONE 5(8): e11971. for DSM-V. Psychiatric Clinics of North America 32(3): 483–524. Charpentier CJ, Aylward J, Roiser JP, et al. (2017) Enhanced risk aver- Bickart KC, Dickerson BC and Barrett LF (2014) The amygdala as a sion, but not loss aversion, in unmedicated pathological anxiety. Bio- hub in brain networks that support social life. Neuropsychologia 63: logical Psychiatry 81(12): 1014–1022. 235–248. Cho J-H, Deisseroth K and Bolshakov VY (2013) Synaptic encod- Birn RM, Shackman AJ, Oler JA, et al. (2014) Evolutionarily conserved ing of fear extinction in mPFC-amygdala circuits. Neuron 80(6): prefrontal-amygdalar dysfunction in early-life anxiety. Molecular 1491–1507. Psychiatry 19(8): 915–922. Cisler JM and Koster EH (2010) Mechanisms of attentional biases Bishop SJ (2007) Neurocognitive mechanisms of anxiety: An integrative towards threat in anxiety disorders: An integrative review. Clinical account. Trends in Cognitive Sciences 11(7): 307–316. Psychology Review 30(2): 203–216. Boedhoe PS, Schmaal L, Abe Y, et al. (2016) Distinct subcortical volume Clark LA, Cuthbert B, Lewis-Fernández R, et al. (2017) Three approaches alterations in pediatric and adult OCD: A worldwide meta-and mega- to understanding and classifying mental disorder: ICD-11, DSM-5, analysis. American Journal of Psychiatry 174(1): 60–69. and the National Institute of Mental Health’s Research Domain Cri- Boeke EA, Moscarello JM, LeDoux JE, et al. (2017) Active avoid- teria (RDoC). Psychological Science in the Public Interest 18(2): ance: Neural mechanisms and attenuation of pavlovian conditioned 72–145. responding. Journal of Neuroscience 37(18): 4808–4818. Clauss JA and Blackford JU (2012) Behavioral inhibition and risk for Bradford DE, Starr MJ, Shackman AJ, et al. (2015) Empirically based developing social anxiety disorder: A meta-analytic study. Journal comparisons of the reliability and validity of common quantification of the American Academy of Child & Adolescent Psychiatry 51(10): approaches for eyeblink startle potentiation in humans. Psychophysi- 1066–1075. ology 52(12): 1669–1681. Cohen MM, Tottenham N and Casey BJ (2013) Translational develop- Bremner JD (2004) Brain imaging in anxiety disorders. Expert Review of mental studies of stress on brain and behavior: Implications for ado- Neurotherapeutics 4(2): 275–284. lescent mental health and illness? Neuroscience 249: 53–62. Britton JC, Lissek S, Grillon C, et al. (2011) Development of anxiety: Community and Mental Health team (2014) Psychological therapies, The role of threat appraisal and fear learning. Depression and Anxi- annual report on the use of IAPT services: England– 2013/14. Avail- ety 28(1): 5–17. able at: https://digital.nhs.uk/catalogue/PUB14899 Brooks SJ and Stein DJ (2015) A systematic review of the neural bases of Cools R, Roberts AC and Robbins TW (2008) Serotoninergic regulation psychotherapy for anxiety and related disorders. Dialogues in Clini- of emotional and behavioural control processes. Trends in Cognitive cal Neuroscience 17(3): 261–279. Sciences 12(1): 31–40. Brown R, Galanter E, Hess EH, et al. (1962). New directions in psychology. Copeland WE, Angold A, Shanahan L, et al. (2014) Longitudinal patterns Oxford: Holt, Rinehart, & Winston. of anxiety from childhood to adulthood: The great smoky mountains Bukalo O, Pinard CR, Silverstein S, et al. (2015) Prefrontal inputs to study. Journal of the American Academy of Child & Adolescent Psy- the amygdala instruct fear extinction memory formation. Science chiatry 53(1): 21–33. Advances 1(6): e1500251. Courtin J, Chaudun F, Rozeske RR, et al. (2014) Prefrontal parvalbumin Burgos-Robles A, Vidal-Gonzalez I and Quirk GJ (2009) Sustained con- interneurons shape neuronal activity to drive fear expression. Nature ditioned responses in prelimbic prefrontal neurons are correlated 505(7481): 92–96. with fear expression and extinction failure. Journal of Neuroscience Craske MG, Stein MB, Eley TC, et al. (2017) Anxiety disorders. Nature 29(26): 8474–8482. Reviews Disease Primers 3: 17024. Carlisi and Robinson 9 Crockett M, Clark L, Roiser J, et al. (2012) Converging evidence for Gee DG, Humphreys KL, Flannery J, et al. (2013b) A developmental central 5-HT effects in acute tryptophan depletion. Molecular Psy- shift from positive to negative connectivity in human amygdala–pre- chiatry 17(2): 121–123. frontal circuitry. Journal of Neuroscience 33(10): 4584–4593. Crockett MJ, Clark L and Robbins TW (2009) Reconciling the role of Giedd JN, Blumenthal J, Jeffries NO, et al. (1999) Brain development serotonin in behavioral inhibition and aversion: Acute tryptophan during childhood and adolescence: A longitudinal MRI study. depletion abolishes punishment-induced inhibition in humans. Jour- Nature Neuroscience 2(10): 861–863. nal of Neuroscience 29(38): 11993–11999. Gillan CM, Morein-Zamir S, Kaser M, et al. (2014) Counterfactual Cullen KR, Gee DG, Klimes-Dougan B, et al. (2009) A preliminary study processing of economic action-outcome alternatives in obsessive- of functional connectivity in comorbid adolescent depression. Neu- compulsive disorder: Further evidence of impaired goal-directed roscience Letters 460(3): 227–231. behavior. Biological Psychiatry 75(8): 639–646. Cuthbert BN and Insel TR (2013) Toward the future of psychiatric diag- Giustino TF and Maren S (2015) The role of the medial prefrontal cortex nosis: The seven pillars of RDoC. BMC Medicine 11: 126–126. in the conditioning and extinction of fear. Frontiers in Behavioral Davis M (1998) Are different parts of the extended amygdala involved in Neuroscience 9: 298. fear versus anxiety? Biological Psychiatry 44(12): 1239–1247. Glenn CR, Klein DN, Lissek S, et al. (2012) The development of fear Davis M (2000) The role of the amygdala in conditioned and uncon- learning and generalization in 8–13 year-olds. Developmental Psy- ditioned fear and anxiety. In: Aggleton JP (ed.) The Amygdala: chobiology 54(7): 675–684. A Functional Analysis. New York: Oxford University Press, Gold AL, Brotman MA, Adleman NE, et al. (2016) Comparing brain pp. 213–288. morphometry across multiple childhood psychiatric disorders. Jour- Davis M (2001) Fear-Potentiated Startle in Rats: Current Protocols in nal of the American Academy of Child & Adolescent Psychiatry Neuroscience. Hoboken, NJ: John Wiley & Sons. 55(12): 1027–1037. Davis M, Walker DL, Miles L, et al. (2010) Phasic vs sustained fear in Goodkind M, Eickhoff SB, Oathes DJ, et al. (2015) Identification of a rats and humans: Role of the extended amygdala in fear vs anxiety. common neurobiological substrate for mental illness. JAMA Psy- Neuropsychopharmacology 35(1): 105–135. chiatry 72(4): 305–315. Dayan P and Huys QJM (2009) Serotonin in affective control. Annual Griebel G and Holmes A (2013) 50 years of hurdles and hope in anxio- Review of Neuroscience 32(2): 95–126. lytic drug discovery. Nature Reviews Drug Discovery 12(9): 667– De Wit SJ, Alonso P, Schweren L, et al. (2014) Multicenter voxel-based 687. morphometry mega-analysis of structural brain scans in obsessive- Grillon C, Ameli R, Woods SW, et al. (1991) Fear-potentiated startle in compulsive disorder. American Journal of Psychiatry 171(3): humans: Effects of anticipatory anxiety on the acoustic blink reflex. 340–349. Psychophysiology 28(5): 588–595. DiLuca M and Olesen J (2014) The cost of brain diseases: A burden or a Grillon C and Davis M (2007) Effects of stress and shock anticipation challenge? Neuron 82(6): 1205–1208. on prepulse inhibition of the startle reflex. Psychophysiology 34(5): Dodd HF, Vogt J, Turkileri N, et al. (2017) Task relevance of emotional 511–517. information affects anxiety-linked attention bias in visual search. Grupe DW and Nitschke JB (2011) Uncertainty is associated with biased Biological Psychology 122: 13–20. expectancies and heightened responses to aversion. Emotion 11(2): Do-Monte FH, Manzano-Nieves G, Quiñones-Laracuente K, et al. (2015) 413–424. Revisiting the role of infralimbic cortex in fear extinction with opto- Grupe DW and Nitschke JB (2013) Uncertainty and anticipation in anxi- genetics. Journal of Neuroscience 35(8): 3607–3615. ety: An integrated neurobiological and psychological perspective. Erk S, Abler B and Walter H (2006) Cognitive modulation of emotion Nature Reviews Neuroscience 14(7): 488–501. anticipation. European Journal of Neuroscience 24(4): 1227–1236. Gungor NZ and Paré D (2016) Functional heterogeneity in the bed Etkin A and Wager TD (2007) Functional neuroimaging of anxiety: A nucleus of the Stria Terminalis. Journal of Neuroscience 36(31): meta-analysis of emotional processing in PTSD, social anxiety dis- 8038–8049. order, and specific phobia. American Journal of Psychiatry 164(10): Gutman DA, Keifer OP Jr, Magnuson ME, et al. (2012) A DTI tractogra- 1476–1488. phy analysis of infralimbic and prelimbic connectivity in the mouse Etkin A, Egner T and Kalisch R (2011) Emotional processing in anterior using high-throughput MRI. Neuroimage 63(2): 800–811. cingulate and medial prefrontal cortex. Trends in Cognitive Sciences Guyer AE, Monk CS, McClure-Tone EB, et al. (2008) A developmental 15(2): 85–93. examination of amygdala response to facial expressions. Journal of Felix-Ortiz A, Burgos-Robles A, Bhagat N, et al. (2016) Bidirectional Cognitive Neuroscience 20(9): 1565–1582. modulation of anxiety-related and social behaviors by amygdala Hakamata Y, Lissek S, Bar-Haim Y, et al. (2010) Attention bias modifi- projections to the medial prefrontal cortex. Neuroscience 321: cation treatment: A meta-analysis toward the establishment of novel 197–209. treatment for anxiety. Biological Psychiatry 68(11): 982–990. Felix-Ortiz AC, Beyeler A, Seo C, et al. (2013) BLA to vHPC inputs Harlow HF and Harlow MK (1965) Chapter 8: The affectional sys- modulate anxiety-related behaviors. Neuron 79(4): 658–664. tems. In: Schrier AM, Stollnitz F and Harlow HF (eds) Behavior Fox AS and Shackman AJ (2017) The central extended amygdala in fear of Nonhuman Primates. Cambridge, MA: Academic Press, pp. and anxiety: Closing the gap between mechanistic and neuroimaging 287–334. research. Neuroscience Letters. Epub ahead of print 30 November. Harmer CJ, Cowen PJ and Goodwin GM (2011) Efficacy markers in DOI: 10.1016/j.neulet.2017.11.056. depression. Journal of Psychopharmacology 25(9): 1148–1158. Ganella DE and Kim JH (2014) Developmental rodent models of fear Harmer CJ, Goodwin GM and Cowen PJ (2009) Why do antidepres- and anxiety: From neurobiology to pharmacology. British Journal of sants take so long to work? A cognitive neuropsychological model Pharmacology 171(20): 4556–4574. of antidepressant drug action. British Journal of Psychiatry 195(2): Ganella DE, Barendse MEA, Kim JH, et al. (2017) Prefrontal-amygdala 102–108. connectivity and state anxiety during fear extinction recall in adoles- Herry C, Ciocchi S, Senn V, et al. (2008) Switching on and off fear by cents. Frontiers in Human Neuroscience 11: 584. distinct neuronal circuits. Nature 454(1): 600–606. Gee DG, Gabard-Durnam LJ, Flannery J, et al. (2013a) Early develop- Hess EH, Brown R, Galanter E, et al. (1962) New directions in psychology. mental emergence of human amygdala–prefrontal connectivity after Hettema JM, Prescott CA, Myers JM, et al. (2005) The structure of maternal deprivation. Proceedings of the National Academy of Sci- genetic and environmental risk factors for anxiety disorders in men ences 110(39): 15638–15643. and women. Archives of General Psychiatry 62(2): 182–189. 10 Brain and Neuroscience Advances Hübner C, Bosch D, Gall A, et al. (2014) Ex vivo dissection of optoge- Kim MJ, Loucks RA, Palmer AL, et al. (2011b) The structural and func- netically activated mPFC and hippocampal inputs to neurons in the tional connectivity of the amygdala: From normal emotion to patho- basolateral amygdala: Implications for fear and emotional memory. logical anxiety. Behavioural Brain Research 223(2): 403–410. Frontiers in Behavioral Neuroscience 8: 64. Klavir O, Genud-Gabai R and Paz R (2013) Functional connectivity Huys QJM, Maia TV and Frank MJ (2016) Computational psychiatry as between amygdala and cingulate cortex for adaptive aversive learn- a bridge from neuroscience to clinical applications. Nature Neurosci- ing. Neuron 80(5): 1290–1300. ence 19(3): 404–413. Klavir O, Prigge M, Sarel A, et al. (2017) Manipulating fear associations Insel T, Cuthbert B, Garvey M, et al. (2010) Research domain crite- via optogenetic modulation of amygdala inputs to prefrontal cortex. ria (RDoC): Toward a new classification framework for research Nature Neuroscience 20(6): 836–844. on mental disorders. American Journal of Psychiatry 167(7): Kling A and Steklis H (1976) A neural substrate for affiliative behav- 748–751. ior in nonhuman primates. Brain, Behavior and Evolution 13(2–3): Izquierdo A and Murray EA (2005) Opposing effects of amygdala and 216–238. orbital prefrontal cortex lesions on the extinction of instrumental Klumpp H, Fitzgerald DA and Phan KL (2013) Neural predictors and responding in macaque monkeys. European Journal of Neuroscience mechanisms of cognitive behavioral therapy on threat processing in 22(9): 2341–2346. social anxiety disorder. Progress in Neuro-psychopharmacology & Jacobs W and Nadel L (1985) Stress-induced recovery of fears and pho- Biological Psychiatry 45: 83–91. bias. Psychological Review 92(4): 512–531. Klumpp H, Fitzgerald JM, Kinney KL, et al. (2017) Predicting cognitive Janak PH and Tye KM (2015) From circuits to behaviour in the amyg- behavioral therapy response in social anxiety disorder with anterior dala. Nature 517(7534): 284–292. cingulate cortex and amygdala during emotion regulation. Neuroim- Johnson FK, Delpech J-C, Thompson GJ, et al. (2018) Amygdala hyper- age Clinical 15: 25–34. connectivity in a mouse model of unpredictable early life stress. Kotov R, Krueger RF, Watson D, et al. (2017) The hierarchical tax- Translational Psychiatry 8(1): 49. onomy of psychopathology (HiTOP): A dimensional alternative Jones PB (2013) Adult mental health disorders and their age at onset. to traditional nosologies. Journal of Abnormal Psychology 126(4): British Journal of Psychiatry 202(1): s5–s10. 454–477. Jones SA, Morales AM, Lavine JB, et al. (2017) Convergent neurobio- Krueger RF and Eaton NR (2015) Transdiagnostic factors of mental dis- logical predictors of emergent psychopathology during adolescence. orders. World Psychiatry 14(1): 27–29. Birth Defects Research 109(20): 1613–1622. Lahey BB, Applegate B, Hakes JK, et al. (2012) Is there a general factor Kaczkurkin A, Moore T, Calkins M, et al. (2017) Common and dissociable of prevalent psychopathology during adulthood? Journal of Abnor- regional cerebral blood flow differences associate with dimensions of mal Psychology 121(4): 971–977. psychopathology across categorical diagnoses. Molecular Psychiatry. Lahey BB, Rathouz PJ, Van Hulle C, et al. (2008) Testing structural mod- Epub ahead of print 19 September. DOI: 10.1038/mp.2017.174. els of DSM-IV symptoms of common forms of child and adolescent Kagan J, Reznick JS and Snidman N (1987) The physiology and psychol- psychopathology. Journal of Abnormal Child Psychology 36(2): ogy of behavioral inhibition in children. Child Development 58(6): 187–206. 1459–1473. Lahey BB, Zald DH, Perkins SF, et al. (2017) Measuring the hierarchical Kalin NH (2017) Mechanisms underlying the early risk to develop anxi- general factor model of psychopathology in young adults. Interna- ety and depression: A translational approach. European Neuropsy- tional Journal of Methods in Psychiatric Research 27(1). chopharmacology 27(6): 543–553. Laviolette SR, Lipski WJ and Grace AA (2005) A subpopulation of neu- Kalin NH, Fox AS, Kovner R, et al. (2016) Overexpressing corticotro- rons in the medial prefrontal cortex encodes emotional learning with pin-releasing factor in the primate amygdala increases anxious tem- burst and frequency codes through a dopamine D4 receptor-depen- perament and alters its neural circuit. Biological Psychiatry 80(5): dent basolateral amygdala input. Journal of Neuroscience 25(26): 345–355. 6066–6075. Kalin NH, Shelton SE and Davidson RJ (2007) Role of the primate orbi- LeDoux JE (2000) Emotion circuits in the brain. Annual Review of Neu- tofrontal cortex in mediating anxious temperament. Biological Psy- roscience 23(1): 155–184. chiatry 62(10): 1134–1139. Linnman C, Rougemont-Bücking A, Beucke JC, et al. (2011) Uncon- Kalisch R and Gerlicher AMV (2014) Making a mountain out of a mole- ditioned responses and functional fear networks in human classical hill: On the role of the rostral dorsal anterior cingulate and dorsome- conditioning. Behavioural Brain Research 221(1): 237–245. dial prefrontal cortex in conscious threat appraisal, catastrophizing, Linnman C, Zeidan MA, Furtak SC, et al. (2012a) Resting amygdala and worrying. Neuroscience & Biobehavioral Reviews 42: 1–8. and medial prefrontal metabolism predicts functional activation of Kalisch R, Wiech K, Critchley HD, et al. (2006) Levels of appraisal: A the fear extinction circuit. American Journal of Psychiatry 169(4): medial prefrontal role in high-level appraisal of emotional material. 415–423. Neuroimage 30(4): 1458–1466. Linnman C, Zeidan MA, Pitman RK, et al. (2012b) Resting cerebral Kessler R, Avenevoli S, Costello E, et al. (2012) Prevalence, persis- metabolism correlates with skin conductance and functional brain tence, and sociodemographic correlates of DSM-IV disorders in activation during fear conditioning. Biological Psychology 89(2): the national comorbidity survey replication adolescent supplement. 450–459. Archives of General Psychiatry 69(4): 372–380. Little JP and Carter AG (2013) Synaptic mechanisms underlying strong Kessler R, Demler R, Olfson F, et al. (2005) Prevalence and treatment of reciprocal connectivity between the medial prefrontal cortex and baso- mental disorders, 1990 to 2003. New England Journal of Medicine lateral amygdala. Journal of Neuroscience 33(39): 15333–15342. 352(24): 2515–2523. Livneh U and Paz R (2012) Amygdala-prefrontal synchronization under- Kim JJ and Fanselow MS (1992) Modality-specific retrograde amnesia of lies resistance to extinction of aversive memories. Neuron 75(1): fear. Science 256(5057): 675–677. 133–142. Kim MJ and Whalen PJ (2009) The structural integrity of an amygdala– Machado CJ and Bachevalier J (2008) Behavioral and hormonal prefrontal pathway predicts trait anxiety. Journal of Neuroscience reactivity to threat: Effects of selective amygdala, hippocampal 29(37): 11614–11618. or orbital frontal lesions in monkeys. Psychoneuroendocrinology Kim MJ, Gee DG, Loucks RA, et al. (2011a) Anxiety dissociates dorsal 33(7): 926–941. and ventral medial prefrontal cortex functional connectivity with the McTeague LM, Huemer J, Carreon DM, et al. (2017) Identification amygdala at rest. Cerebral Cortex 21(7): 1667–1673. of common neural circuit disruptions in cognitive control across Carlisi and Robinson 11 psychiatric disorders. American Journal of Psychiatry 174(7): Pine DS (2007) Research review: A neuroscience framework for pedi- 676–685. atric anxiety disorders. Journal of Child Psychology and Psychiatry Maslowsky J, Mogg K, Bradley BP, et al. (2010) A preliminary investiga- 48(7): 631–648. tion of neural correlates of treatment in adolescents with generalized Price JL and Drevets WC (2012) Neural circuits underlying the patho- anxiety disorder. Journal of Child and Adolescent Psychopharma- physiology of mood disorders. Trends in Cognitive Sciences 16(1): cology 20(2): 105–111. 61–71. Mechias M-L, Etkin A and Kalisch R (2010) A meta-analysis of Regev L, Tsoory M, Gil S, et al. (2012) Site-specific genetic manipula- instructed fear studies: Implications for conscious appraisal of threat. tion of amygdala corticotropin-releasing factor reveals its imperative Neuroimage 49(2): 1760–1768. role in mediating behavioral response to challenge. Biological Psy- Meyer A (2017) A biomarker of anxiety in children and adolescents: A chiatry 71(4): 317–326. review focusing on the error-related negativity (ERN) and anxiety Robinson OJ, Charney DR, Overstreet C, et al. (2012a) The adaptive across development. Developmental Cognitive Neuroscience 27(1): threat bias in anxiety: Amygdala–dorsomedial prefrontal cortex cou- 58–68. pling and aversive amplification. Neuroimage 60(1): 523–529. Milad MR and Quirk GJ (2002) Neurons in medial prefrontal cortex sig- Robinson OJ, Cools R and Sahakian BJ (2012b) Tryptophan depletion nal memory for fear extinction. Nature 420(6911): 70–74. disinhibits punishment but not reward prediction: Implications for Milad MR and Quirk GJ (2012) Fear extinction as a model for transla- resilience. Psychopharmacology 219(2): 599–605. tional neuroscience: Ten years of progress. Annual Review of Psy- Robinson OJ, Krimsky M, Lieberman L, et al. (2014) The dorsal medial chology 63: 129–151. prefrontal (anterior cingulate) cortex–amygdala aversive amplifica- Milad MR and Rauch SL (2007) The role of the orbitofrontal cortex tion circuit in unmedicated generalised and social anxiety disorders: in anxiety disorders. Annals of the New York Academy of Sciences An observational study. Lancet Psychiatry 1(4): 294–302. 1121: 546–561. Robinson OJ, Krimsky M, Lieberman L, et al. (2016) Anxiety-poten- Milad MR, Pitman RK, Ellis CB, et al. (2009) Neurobiological basis of tiated amygdala-medial frontal coupling and attentional control. failure to recall extinction memory in posttraumatic stress disorder. Translational Psychiatry 6(6): e833. Biological Psychiatry 66(12): 1075–1082. Robinson OJ, Overstreet C, Allen PS, et al. (2013b) The role of serotonin Milad MR, Quirk GJ, Pitman RK, et al. (2007) A role for the human dor- in the neurocircuitry of negative affective bias: Serotonergic modu- sal anterior cingulate cortex in fear expression. Biological Psychiatry lation of the dorsal medial prefrontal-amygdala ‘aversive amplifica- 62(10): 1191–1194. tion’ circuit. Neuroimage 78(1): 217–223. Mkrtchian A, Aylward J, Dayan P, et al. (2017) Modeling avoidance in Robinson OJ, Vytal K, Cornwell B, et al. (2013a) The impact of anxiety mood and anxiety disorders using reinforcement learning. Biological upon cognition: Perspectives from human threat of shock studies. Psychiatry 82(7): 532–539. Frontiers in Human Neuroscience 7: 203. Monk C, Nelson E, McClure E, et al. (2006) Ventrolateral prefrontal Rothbaum BO and Davis M (2003) Applying learning principles to the cortex activation and attentional bias in response to angry faces in treatment of post-trauma reactions. Annals of the New York Academy adolescents with generalized anxiety disorder. American Journal of of Sciences 1008: 112–121. Psychiatry 163(6): 1091–1097. Roy AK, Vasa RA, Bruck M, et al. (2008) Attention bias toward threat Morgan MA and LeDoux JE (1995) Differential contribution of dorsal in pediatric anxiety disorders. Journal of the American Academy of and ventral medial prefrontal cortex to the acquisition and extinc- Child & Adolescent Psychiatry 47(10): 1189–1196. tion of conditioned fear in rats. Behavioral Neuroscience 109(4): Roy-Byrne P (2015) Treatment-refractory anxiety; definition, risk fac- 681–688. tors, and treatment challenges. Dialogues in Clinical Neuroscience Nugent NR, Tyrka AR, Carpenter LL, et al. (2011) Gene–environment 17(2): 191–206. interactions: Early life stress and risk for depressive and anxiety dis- Rygula R, Clarke HF, Cardinal RN, et al. (2015) Role of central sero- orders. Psychopharmacology 214(1): 175–196. tonin in anticipation of rewarding and punishing outcomes: Effects Okon-Singer H and Aue T (2017) Neurocognitive mechanisms modulat- of selective amygdala or orbitofrontal 5-HT depletion. Cerebral ing attention bias in anxiety: Current perspectives. Biological Psy- Cortex 25(9): 3064–3076. chology: 122: 1–3. Santangelo AM, Roberts AC, Ferguson-Smith AC, et al. (2016) Novel Oliver BR and Plomin R (2007) Twins’ early development study (TEDS): primate model of serotonin transporter genetic polymorphisms asso- A multivariate, longitudinal genetic investigation of language, cog- ciated with gene expression, anxiety and sensitivity to antidepres- nition and behavior problems from childhood through adolescence. sants. Neuropsychopharmacology 41(9): 2366–2376. Twin Research and Human Genetics 10(1): 96–105. Sawyer SM, Azzopardi PS, Wickremarathne D, et al. (2018) The age of Passamonti L, Crockett MJ, Apergis-Schoute AM, et al. (2012) Effects adolescence. Lancet Child & Adolescent Health 2(9): 223–228. of acute tryptophan depletion on prefrontal-amygdala connectivity Schumann G, Loth E, Banaschewski T, et al. (2010) The IMAGEN while viewing facial signals of aggression. Biological Psychiatry study: Reinforcement-related behaviour in normal brain function and 71(1): 36–43. psychopathology. Molecular Psychiatry 15(12): 1128–1139. Pattwell SS, Liston C, Jing D, et al. (2016) Dynamic changes in neural Selden N, Everitt B, Jarrard L, et al. (1991) Complementary roles for the circuitry during adolescence are associated with persistent attenua- amygdala and hippocampus in aversive conditioning to explicit and tion of fear memories. Nature Communications 7: 11475. contextual cues. Neuroscience 42(2): 335–350. Perna G, Favaron E, Di Bella D, et al. (2005) Antipanic efficacy of parox- Sengupta A, Yau JO, Bressel PJ-RD, et al. (2017) Basolateral amygdala etine and polymorphism within the promoter of the serotonin trans- neurons maintain aversive emotional salience. Journal of Neurosci- porter gene. Neuropsychopharmacology 30(12): 2230–2235. ence 38(12): 2460–2417. Phelps EA, Delgado MR, Nearing KI, et al. (2004) Extinction learning in Senn V, Wolff SB, Herry C, et al. (2014) Long-range connectivity defines humans: Role of the amygdala and vmPFC. Neuron 43(6): 897–905. behavioral specificity of amygdala neurons. Neuron 81(2): 428–437. Phillips R and LeDoux J (1992) Differential contribution of amygdala Shackman AJ and Fox AS (2016) Contributions of the central extended and hippocampus to cued and contextual fear conditioning. Behav- amygdala to fear and anxiety. Journal of Neuroscience 36(31): ioral Neuroscience 106(2): 274–285. 8050–8063. Pinard CR, Mascagni F and McDonald AJ (2012) Medial prefrontal cor- Shackman AJ, Salomons TV, Slagter HA, et al. (2011) The integration tical innervation of the intercalated nuclear region of the amygdala. of negative affect, pain and cognitive control in the cingulate cortex. Neuroscience 205: 112–124. Nature Reviews Neuroscience 12(3): 154–167. 12 Brain and Neuroscience Advances Shackman AJ, Tromp DP, Stockbridge MD, et al. (2016) Dispositional Sylvester CM, Barch DM, Harms MP, et al. (2016) Early childhood negativity: An integrative psychological and neurobiological per- behavioral inhibition predicts cortical thickness in adulthood. spective. Psychol Bull 142(12): 1275–1314. Journal of the American Academy of Child & Adolescent Psychiatry Shin LM and Liberzon I (2009) The neurocircuitry of fear, stress, and 55(2): 122–129. anxiety disorders. Neuropsychopharmacology 35(1): 169–191. Terburg D, Morgan BE, Montoya ER, et al. (2012) Hypervigilance for Shou H, Yang Z, Satterthwaite TD, et al. (2017) Cognitive behavioral fear after basolateral amygdala damage in humans. Translational therapy increases amygdala connectivity with the cognitive con- Psychiatry 2(5): e115. trol network in both MDD and PTSD. Neuroimage Clinical 14(1): Thompson PM, Stein JL, Medland SE, et al. (2014) The ENIGMA con- 464–470. sortium: Large-scale collaborative analyses of neuroimaging and Sierra-Mercado D, Padilla-Coreano N and Quirk GJ (2011) Dissociable genetic data. Brain Imaging and Behavior 8(2): 153–182. roles of prelimbic and infralimbic cortices, ventral hippocampus, and Tovote P, Fadok JP and Lüthi A (2015) Neuronal circuits for fear and basolateral amygdala in the expression and extinction of conditioned anxiety. Nature Reviews Neuroscience 16(10): 317. fear. Neuropsychopharmacology 36(2): 529–538. Van Erp TG, Hibar DP, Rasmussen JM, et al. (2016) Subcortical brain Sotres-Bayon F, Sierra-Mercado D, Pardilla-Delgado E, et al. (2012) volume abnormalities in 2028 individuals with schizophrenia and Gating of fear in prelimbic cortex by hippocampal and amygdala 2540 healthy controls via the ENIGMA consortium. Molecular Psy- inputs. Neuron 76(4): 804–812. chiatry 21(4): 547–553. Strawn JR, Bitter SM, Weber WA, et al. (2012) Neurocircuitry of gen- Vidal-Gonzalez I, Vidal-Gonzalez B, Rauch SL, et al. (2006) Micro- eralized anxiety disorder in adolescents: A pilot functional neuro- stimulation reveals opposing influences of prelimbic and infralimbic imaging and functional connectivity study. Depression and Anxiety cortex on the expression of conditioned fear. Learning & Memory 29(11): 939–947. 13(6): 728–733. Sudlow C, Gallacher J, Allen N, et al. (2015) UK biobank: An open Vogt BA (2005) Pain and emotion interactions in subregions of the cin- access resource for identifying the causes of a wide range of gulate gyrus. Nature Reviews Neuroscience 6(7): 533–544. complex diseases of middle and old age. PLoS Medicine 12(3): Vos T, Allen C, Arora M, et al. (2016) Global, regional, and national e1001779. incidence, prevalence, and years lived with disability for 310 dis- Svihra M and Katzman MA (2004) Behavioural inhibition: A predictor of eases and injuries, 1990–2015: A systematic analysis for the Global anxiety. Paediatrics & Child Health 9(8): 547–550. Burden of Disease Study 2015. Lancet 388(10053): 1545–1602. Swartz JR and Monk CS (2013) The role of corticolimbic circuitry in Vytal KE, Overstreet C, Charney DR, et al. (2014) Sustained anxiety the development of anxiety disorders in children and adolescents. increases amygdala–dorsomedial prefrontal coupling: A mechanism In: Andersen SL and Pine DS (eds) The Neurobiology of Childhood. for maintaining an anxious state in healthy adults. Journal of Psy- Berlin: Springer, pp. 133–148. chiatry & Neuroscience 39(5): 321–329. Swartz JR, Carrasco M, Wiggins JL, et al. (2014) Age-related changes in White LK, Sequeira S, Britton JC, et al. (2017) Complementary features the structure and function of prefrontal cortex–amygdala circuitry in of attention bias modification therapy and cognitive-behavioral ther- children and adolescents: A multi-modal imaging approach. Neuro- apy in pediatric anxiety disorders. American Journal of Psychiatry image 86(2): 212–220. 174(8): 775–784.

Journal

Brain and Neuroscience AdvancesSAGE

Published: May 8, 2018

Keywords: Anxiety; circuit; negative bias; prefrontal cortex

There are no references for this article.