The role of reward-related brain activity in response to treatment and later depression severity: data from a randomized controlled trial in early adolescents with anxiety disorders

Abstract

Alterations in reward-related brain activity have been linked to response to psychological treatment in adolescents with anxiety disorders. However, it remains unknown whether these effects are driven by reward anticipation or feedback, which reflect different functional roles in motivated behavior, or whether brain activity changes as a function of treatment response. The current study investigated these questions in the context of a randomized controlled trial of cognitive-behavioral therapy (CBT) for anxiety disorders in adolescents. This study used an fMRI paradigm to investigate reward-related brain activity in youth aged 9–14 with anxiety disorders (ANX; N = 133; 57 female) before and after 16 weeks of CBT or an active comparison (child-centered therapy, CCT). Age- and sex-matched healthy comparison (HC) youth (N = 38; 17 female) completed scans on a similar timeline. A subset of ANX youth completed a 2-year follow-up assessment of depressive symptoms. At pretreatment, ANX compared to HC youth demonstrated reduced brain activity in reward-related regions (e.g. dorsal striatum, thalamus) during reward anticipation, and elevated activity in angular gyrus, PCC and inferior frontal gyrus during reward feedback. Reduced pretreatment activation in the precuneus/cuneus and pre-to-post reductions in left angular gyrus corresponded with treatment response. Finally, pre-to-post increases in posterior cingulate cortex (PCC) corresponded with increased depressive symptoms at 2 years. Our results suggest that reward-related brain activity outside of striatal reward regions, including PCC, precuneus and angular gyrus, plays a role in treatment response in youth with anxiety disorders. Trial registration: ClinicalTrials.gov NCT00774150.

Anxiety disorders are estimated to impact about 5% of children in developed countries, and cause a significant mental health burden [1]. Psychological treatment remains a primary approach to treat adolescent anxiety disorders, with cognitive-behavioral therapy (CBT) considered first-line due to its well-demonstrated efficacy [2]. However, even the best-supported current treatments have only moderate efficacy and relapse is common [3]. Thus, improving treatment response to psychological therapy is a public health urgency. Toward that end, recent investigations have begun to investigate neural correlates of response to psychological treatments [4]. Of these, alterations in reward-related brain functioning have emerged as an especially promising target of study (see reviews in, e.g., [5, 6]).

During the transition into adolescence, there is increased sensitivity to rewards in a collection of brain regions commonly referred to as reward circuitry, which includes ventral and dorsal striatum, anterior insula, anterior cingulate cortex (ACC) and posterior cingulate cortex (PCC) [7]. This increased sensitivity drives increases in exploratory behavior which are thought to promote healthy development of higher-order cognition [8]. Alterations in function of these brain regions during adolescence are related to internalizing (depressive and anxious) [6, 9] as well as to externalizing psychopathology [10], suggesting that abnormal maturation of reward circuitry may contribute to a general psychological vulnerability with specific relevance to anxiety disorders.

Reward-related brain activity can be divided into two phases, which have differing functions and relationships to psychopathology. First, the reward anticipatory phase includes the processes that govern and regulate approach behaviors, including evaluating incentive cues and preparing motor responses [11], which may collectively be termed approach motivation [12]. The effort an individual will expend in pursuit of a reward is associated with amplitude of the cue-P3 EEG waveform, which is associated with striatal activity [13], suggesting that fMRI activity in reward circuitry may serve as an index of approach motivation. Anxiety, particularly social anxiety, is related to anticipatory hyperactivity when rewards or punishments are contingent upon participant behavior and incentives are high in magnitude [6]. Because social anxiety is a risk factor for the development of GAD [14], this suggests that increased sensitivity to high-magnitude incentives (including rewards) may predispose to generalized anxiety. The anticipation period is contrasted with the consummatory period following reward receipt, or feedback. During this phase, dopamine signaling within the striatum contributes to a hedonic response, i.e., the sensation of pleasure at receiving a reward [15], and to learning processes as the brain integrates discrepancies between expected and received rewards (i.e., ‘reward prediction errors’). Neural activity in this phase has been less-consistently related to anxiety disorders [6] but is robustly associated with risk for later depression [9].

In these ways, the two phases of the reward response may confer vulnerability to internalizing symptoms through different mechanisms. Enhanced sensitivity to incentives may lend to pathological avoidance leading to anxiety, while decreased hedonic responding and reward learning may lead an adolescent to expend less energy seeking novel or rewarding experiences and thus become depressed [16,17]. Yet these mechanisms may also represent clinical opportunity. Altered reward processing (particularly approach motivation) could improve responses to exposure-based therapies by promoting an adolescent’s willingness to approach aversive stimuli [18]; similarly, altering hedonic responding and reward learning in the feedback period may lessen anxiety symptoms and vulnerability to later depression (although there has been debate about the degree to which alterations in reward-related brain activity is truly causal with regards to depression; for a discussion, see [19]).

Prior research has demonstrated that pretreatment reward-related brain activity is associated with response to psychological treatment in anxiety disorders. Sequeira et al. [20], in a study derived from the same randomized, controlled trial of CBT for pediatric anxiety disorders as the present report [21], found that higher levels of pretreatment reward-related activity in a cluster in striatum and subgenual anterior cingulate cortex (sgACC) corresponded to decreased anxiety symptomatology after 16 weeks of treatment in both CBT and active therapy control groups. This study extended prior findings relating elevated brain activity in reward-related regions (striatum and sgACC) to post-CBT reductions in anxiety symptoms [6] and suggested this effect may not be specific to CBT but apply more generally to psychological treatments. However, Sequeira’s analyses did not differentiate reward anticipation from feedback, leaving questions about the specific mechanisms relating reward-related brain activity to treatment response.

Another remaining question is the extent to which psychological treatments alter brain function in reward circuitry in adolescents. Such effects could help explain the long-term protective effect that can be seen with psychological treatment during this critical developmental period [22, 23]. Recent meta-analyses of adult studies, and a systematic review of the smaller number of adolescent studies, indicate that psychological treatments alter function in cognitive and emotional networks [24, 25–26]. However, the studies included used fMRI tasks focused on negative valence systems such as threat rather than reward. In the largest such study to date, Haller et al. [27] found that unmedicated youth aged 8–17 with anxiety disorders had abnormally elevated activity in amygdala, dorsal striatum, and frontoparietal activity to a dot-probe task with threat stimuli (angry faces) prior to treatment with CBT. Amygdala and striatal activity did not change with treatment but frontoparietal activity normalized to the level of healthy controls, suggesting that psychological treatment altered not automatic processing (associated with amygdala and subcortical regions), but deliberate cognitive processing (associated with frontoparietal network [28]). These alterations did not relate to treatment response, limiting interpretation of the role they play in reducing anxiety symptoms. Nevertheless, they provide evidence that psychological treatments, including CBT, can normalize function in brain regions relevant to anxiety disorders.

It is unclear whether CBT might impact reward processing in similar ways. However, it is notable that frontoparietal network and other associative cortical regions are implicated in reward processing in adolescents. In particular, lateral prefrontal cortex and angular gyrus, regions associated with cognitive control [29] and conceptual thought [30], are more active in adolescents during both reward anticipant and receipt than adults [7]. This has been interpreted to suggest that adolescents engage more off-task processing, perhaps self-referential or conceptual thoughts about the meaning of rewards or their relationship to the adolescent’s self-concept [7]. It is not difficult to imagine how such types of thought could interfere with reward processing; indeed, one form of negative self-referential processing, rumination, has been related to both anticipatory and consummatory reward-related brain activity [31, 32], and activity in the default-mode network (DMN), which is associated with rumination, is elevated during reward processing in social anxious individuals [33]. Thus, impacts of psychological treatments on cognitive processes instantiated in non-reward related regions, such as frontoparietal and default-mode networks, could potentially impact reward processing in adolescents with anxiety disorders. Early adolescence represents a crucial time to understand the impact of such interventions as it is a time of important developmental changes in motivation and underlying neural substrates [22].

The aim of the present pre-registered study was to expand upon the existing literature in several ways. First, building upon Sequeira et al. [20], we examined whether associations with treatment response in early adolescents with anxiety disorders differed between anticipation and feedback task periods, in order to elucidate how different components of the reward response (i.e., approach motivation vs. hedonic responding) are impacted by psychological treatment. Second, we examined whether reward-related brain activity within those phases changed as a function of treatment response. Following Haller et al. [27], we compared reward-related brain activity between anxious and healthy comparison (HC) youth to identify patterns of disrupted functioning at baseline (pretreatment), and examined to what extent functioning within this circuitry normalized toward levels seen in HC youth after treatment. Third, the presence of an active therapy control group provided the opportunity to test whether these effects apply to both psychological treatments, as in Sequeira et al. [20], or were specific to CBT. And finally, although not originally pre-registered, we examined how reward-related brain activity related to depressive symptoms in a subset of participants two years later, to examine whether these brain activity differences had implications for differences in risk for later depression previously described in this sample [23].

We hypothesized that both reward anticipation and feedback would engage brain regions including amygdala, ventral and dorsal striatum, anterior insula, dorsal anterior cingulate, thalamus, hippocampus, posterior cingulate, medial PFC and subgenual ACC, which have previously been described [34]. Specifically, we hypothesized that early adolescents with anxiety disorders (ANX) would exhibit heightened pretreatment activity compared to HC youth, based on previously-reported differences in reward reactivity in anxiety [6]. Further, we hypothesized that youth who responded to treatment would show heightened response to reward anticipation and feedback prior to treatment, and post-treatment normalization of activity to the level seen in HC youth. Youth who did not respond to treatment were predicted to continue to have elevated levels of reward-related activity especially during reward anticipation. Our primary analyses examined both therapy groups (CBT and child centered therapy (CCT)) together, both to keep analyses parallel with Sequeira et al. [20], and to maximize power given our sample size, and as such our hypotheses relate to common treatment factors. However, we conducted exploratory analyses of group differences for all pretreatment and pre-to-post hypotheses, and additionally, following Sequeira et al. [20], examined measures speaking to mechanisms common between both therapies (therapeutic alliance) and specific to CBT (number and therapist-reported quality of exposures); these are reported in the Supplement. Additional exploratory analyses were associations with depressive symptoms at two-year follow-up [23] and trait rumination and worry (reported in the Supplement).

Materials and methods

The present work derives from a randomized controlled trial of cognitive-behavioral therapy (CBT) compared to an active control, child-centered therapy (CCT) and a sample without psychiatric diagnoses (healthy comparison, HC) (clinicaltrials.gov #NCT00774150). Primary study endpoints have been previously published [21].

Participants

Participants were treatment-seeking youth aged 9–14 meeting DSM-IV diagnostic criteria for generalized anxiety disorder, separation anxiety disorder and/or social phobia, residing in or around the Pittsburgh metropolitan area. Exclusion criteria included IQ < 70, current use of psychoactive medication (excluding stimulants, which could not be taken on day of MRI scan); current diagnoses of major depressive disorder, obsessive-compulsive disorder, or hyperactive/combined-type attention-deficit hyperactivity disorder; lifetime history of psychosis, autism spectrum disorder, or bipolar disorder.

A total of 133 youth were randomized into CBT or CCT treatment conditions, of which 97 (67 CBT, 30 CCT) had usable pretreatment and 73 (51 CBT, 22 CCT) had usable post-treatment fMRI data for the present task. The HC group comprised 46 youth who were age, sex, and IQ-matched to the anxious participants. Of these, 38 had usable pre- and post-treatment data. 80 ANX youth completed a two-year followup assessment of depression [23]; of these, 52 had usable pre- and post-treatment fMRI data. Only participants with complete data were analyzed for pre-to-post difference analyses (see Supplement for details). No differences were found between CBT, CCT or HC in age, sex at birth, or family income level (Table 1).

Table 1. Demographic and clinical characteristics.

	CBT		CCT		HC
	Mean	SD	Mean	SD	Mean	SD
Age (years)	11.13	1.37	11.37	1.70	11.63	1.68
	No.	%	No.	%	No.	&
Sex assigned at birth
Female	42	62.69	15	50.00	17	44.74
Male	25	37.31	15	50.00	21	55.26
Ethnicity
Not Hispanic or Latino	65	97.01	30	100.00	36	94.74
Hispanic or Latino	2	2.99	0	0	2	5.26
Race
White	60	89.55	27	90.00	27	71.05
Black/African-American	1	1.49	2	6.67	6	15.79
Biracial	5	7.46	1	3.33	3	7.89
Hispanic	1	1.49	0	0	2	5.26
Parent Highest Level of Education1
Some high school	1	1.52	0	0.00	0	0.00
High school graduate	10	15.15	2	6.67	6	15.79
Some college	12	18.18	10	33.33	7	18.42
College degree	29	43.93	7	23.33	19	50.00
Graduate professional training	14	21.21	11	36.67	6	15.79
Total family income per year
$0–30,000	7	10.93	2	8.00	7	18.92
$30,000–60,000	18	28.13	8	32.00	11	29.73
$60,000–80,000	12	18.75	3	12.00	7	18.92
$80,000–100,000	10	15.63	6	24.00	4	10.81
>$100,000	17	26.56	6	24.00	8	21.62
Anxiety disorder diagnosis2
Generalized anxiety disorder	46	68.66	22	73.33	0	0.00
Separation anxiety disorder	14	20.90	5	16.67	0	0.00
Social anxiety disorder	16	23.88	8	26.67	0	0.00
Specific phobia	8	11.94	3	10.00	0	0.00
PTSD	2	2.99	0	0.00	0	0.00
Panic disorder	1	1.49	1	3.33	0	0.00
Other comorbid diagnoses
Major depressive disorder	1	1.49	1	3.33	0	0.00
ADHD, inattentive type or NOS	2	2.99	3	10.00	0	0.00
Oppositional defiant disorder	2	2.99	1	3.33	0	0.00
Enuresis	4	5.97	1	3.33	0	0.00
Tic disorder	1	1.49	3	10.00	0	0.00
PARS 6-item Total Score
	21.41	5.00	20.40	4.92	1.34	2.47

Study procedure

Full recruitment and study procedures are published elsewhere [20, 21]. Briefly, participants were screened by phone then underwent a clinical intake including a structured diagnostic interview (DSM-IV K-SADS-PL [35]). Participants who qualified were invited for a pre-treatment MRI visit, randomized to 16 sessions of CBT or CCT, and finally returned for a second assessment and a second MRI visit. HC youth did not receive treatment but completed screening procedures and MRI visits on approximately the same timeline.

Study materials

The materials below are a subset of instruments and procedures from [21] relevant to the present study.

Clinical measures

Each participant was interviewed by a trained evaluator blinded to treatment condition using the Schedule for Affective Disorders and Schizophrenia for School-Age Children–Present and Lifetime Version (K-SADS-PL) [35]. Anxiety severity was established using the Pediatric Anxiety Rating Scale (PARS) (RUPP Study Group, [36]), administered by independent evaluators. Treatment response was defined as a 35% reduction in PARS score at post-treatment. This binarized outcome measure was used to conform with the methods of the parent study [21] such that our results could be interpreted in the context of previously-published clinical and imaging results [20, 21] and published guidelines [37]. Participants also completed the Mood and Feelings Questionnaire [38] for depressive symptoms. A subset (N = 80) of anxious youth completed the MFQ at 2-year follow-up [23].

Treatment

Both treatments were administered over 16 weekly sessions, with parent sessions at weeks 4 and 9. 1. CBT was administered using Coping Cat [39], a structured, manual-based and validated treatment for children with anxiety disorders. The first 8 sessions comprise anxiety-management skills such as learning to identify somatic symptoms of anxiety and learning relaxation and problem-solving strategies to manage anxiety. The second half of treatment employs graded exposure through increasingly anxiety-provoking situations. 2. CCT [40] is a manual-based and empirically-supported treatment which is nondirective and based on principles such as unconditional positive regard, empathy, and therapeutic alliance. In contrast to CBT, CCT is designed to mirror typical supportive therapies that youth might experience in the community. In each session the therapist encouraged the child to discuss feelings and used skills such as supportive listening, reflection, and accurate empathy.

fMRI task and instructions

A previously-published approach-avoid task [41] was modified to include approach, avoid, and ambiguous trials (divided into ambiguous-approach and ambiguous-avoid based on participant behavior) (see Fig. 1). Only approach and ambiguous-approach trials were analyzed because the avoid manipulation did not appear sufficiently strong (see Supplement for details). Approach and ambiguous-approach trials were combined with the rationale that, although ambiguous trials theoretically contained some competing avoidance drive, choosing only the trials in which participants chose to approach would minimize its impact.

Fig. 1 Approach-avoidance task design. [Images not available. See PDF.]

A) The Approach-Avoidance task. Participants completed 2 runs, each comprising 8 randomized trials of 4 conditions: approach reward (“approach”), avoid loss (“avoid”), ambiguous (“ambiguous”), and control trials. Participants were familiarized with the task and contingencies in a practice run outside of the scanner. Specifically, they were instructed that there was a “bank” off-screen to the right of the display and that they should attempt to maximize their winnings by moving different cues into the bank or away from the bank. The cues differed by trial type: the approach (“$”) cue appeared on the left of the screen and moved slowly toward the bank; the avoid cue, referred to in instructions as a “money-eating snake” (“ö”) appeared on the left of the screen and moved quickly toward the bank; the ambiguous (“?”) cue appeared on the left of the screen and moved slowly toward the bank; and the control (“X”) cue appeared in the middle of the screen and did not move. Moving the approach cue to the bank resulted in a fixed winning of 50 points; moving the avoid (“snake”) cue into the bank resulted in a fixed loss of 50 points while moving it away from the bank avoided this loss; moving the ambiguous cue into the bank would result in a 50 point win 75% of the time and a 50 point loss 25% of the time; and moving the control cue did not result in wins or losses but was necessary to end the trial. Participants thus learned to move the approach and ambiguous cues toward the bank and the avoid cue away from the bank. Six consecutive button-presses to the right or left were required to move a cue either into or away from the bank and end the trial (except in the control condition, where any six button presses were sufficient, and failure to move the cue would prompt feedback encouraging participation); thus, the task maintained a fixed-ratio 6 reinforcement schedule throughout. Total time for all 6 button presses varied from 1.2–7.6 s, after which the cue disappeared and the “Calculating…” screen displayed. The reward displayed on the screen changed approximately halfway through the study, such that the first half of participants were shown points (which were exchanged for money after the scan), while the second half of participants were shown the actual monetary value they earned. Both groups were included in analyses. Only approach and ambiguous-approach trials were included for analyses of Reward Anticipation, and only trials where participants received rewards were included for analyses of Feedback (see Supplement). B) Task regressors and timing. Each color denotes a different regressor, which were further divided among task conditions. The first 2 s of the response period and the feedback period were regressors of interest; the post-response period was modeled but not included in analyses, and the fixation period was left unmodeled as implicit baseline. Variance inflation factor calculated for a selection of subjects were within acceptable ranges (<5) for the regressors of interest.

Consistent with prior published analyses of this task [41], we constructed a regressor comprising the first 2 s of the beginning of each trial beginning with cue presentation (“Anticipation”). Only the first 2 s were modeled to capture brain activity related to approach motivation with minimal interference from approach behavior (motor activity, i.e., button presses), although this could not be excluded entirely. A “Feedback” regressor comprised the full 3–12 s feedback period prior to the inter-trial fixation cross, and included only trials where participants received rewards (i.e., not ambiguous-loss trials).

fMRI data collection and analysis

Data collection and preprocessing

Imaging was completed on a Siemens 3T Trio MRI scanner. Two runs of 6 m24 s (230 volumes) were collected at both time points. Preprocessing was conducted using a standard fMRIPrep version 21.0.1 [42] pipeline including skull-stripping, co-registration of structural images from both timepoints and normalization to standard (MNI) space; and motion correction, registration, resampling to MNI space, smoothing and bandpass filtering of functional images. Full details of image acquisition and preprocessing are found in the Supplement.

Wholebrain analyses: first and second levels

fMRI data were analyzed using a general linear model implemented in FSL [43]. Regressors (visualized in Fig. 1) were constructed at the first level by convolving boxcar functions with a double-gamma canonical hemodynamic response function. We included regressors to model standard and extended motion parameters (6 rigid-body regressors comprising translation and rotation in 3 dimensions, their derivatives, and the squares of these derivates for 24 total) and exclude any TRs with framewise displacement > 0.9 mm (0–93 volumes per run, average = 20). Runs with > 25% of TRs censored were excluded from analyses (78 runs, <10% of sample).

Contrasts were constructed at the first level by subtracting regressors between the condition of interest and the Control condition (e.g., Reward > Control Anticipation). Contrast estimates were averaged across runs using a fixed effects linear model (second level). For pre-to-post analyses, a paired t-test was first conducted between brain activity from the 2 pre- and 2 post-treatment runs, which was carried forward to the third level. Group-level regressions on individual contrast estimates and pre-to-post paired t-test maps were computed at the third level using FSL’s FLAME1.

Wholebrain analyses: third (group) level

Main effects of task conditions were tested with one-sample t-tests of pretreatment brain data. Group differences at pretreatment were tested with two-sample t-tests between groups (ANX vs HC and treatment responders vs. nonresponders) on pretreatment brain data. Group differences in pre-to-post changes in brain activity were tested using two-sample t-tests on the statistical maps from pre-to-posttreatment paired t-tests computed at the second level (see above). Correlations with individual difference measures, including two-year follow-up MFQ scores and exploratory measures, were computed voxelwise on pretreatment and pre-to-posttreatment data. Multiple comparison correction was done using cluster-based random field theory with a cluster forming threshold of z = 3.1 (p < 0.001) and a corrected cluster p-value threshold of 0.05 [44]. All analyses were corrected for age and sex assigned at birth.

ROI analyses

In addition to whole-brain analyses, brain activity was averaged across regions-of-interest (ROIs) and entered into statistical analyses for all hypotheses. To limit the total number of comparisons, eight ROIs were selected based on theoretical interest and prior literature. These were defined either anatomically or by placing spheres in anatomical regions of interest in MNI space. For 7 of these, no significant results for any hypothesis survived multiple comparison correction, so we report only on activity for one test in midline PCC and additional results are discussed in the Supplement. This ROI was defined as a 3 mm radius sphere placed at MNI coordinates (0, −54, 26) due to this location’s involvement in the default-mode network (DMN) (confirmed using an automated meta-analysis via Neurosynth [45]). Pre-to-post treatment hypotheses were tested using an ANCOVA model with post-treatment data as outcome, and pre-treatment data, age and sex as covariates; group differences were examined by including group as the predictor of interest (e.g., post-treatment PCC fMRI ~ Treatment response + pre-treatment PCC fMRI + Age + Sex). Variables in this model were mean-centered. To correct for multiple comparisons, p-values for each model were adjusted using the Holm method [46].

Reliability analyses

Split-half reliabilities of task conditions of interest were calculated in the above ROIs. Full methods and results are reported in the Supplement.

Sensitivity analyses

To test whether differences in baseline anxiety severity and depressive symptoms affected results, sensitivity analyses were conducted with these variables as covariates for all analyses above. These are reported only for significant effects.

Data analytic checks

Data were visually inspected for normality and to check model assumptions. All reported statistical tests are two-tailed.

Location of code

Analysis code and template files are available at the following link: https://github.com/ceciwestbrook/RewardAndAnxiety.

Results

Participants

Therapy completion

Of the youth with usable pre- and post-treatment data, 61 (83.56%) attended all 16 sessions. Of the remainder, 4 completed 15 sessions, 4 completed 14 sessions, 1 each completed 10 and 11 sessions, and 1 each completed extra sessions (18 and 19).

Treatment responders vs. non-responders

No differences were found between treatment responders and nonresponders in age, sex at birth, family income level, or baseline PARS or MFQ scores. A full table of demographic and clinical characteristics can be found in Table S1 in the supplement.

Youth with two-year followup depression data

Adolescents with and without two-year followup depression data did not differ significantly in age, sex at birth, race, or baseline PARS or MFQ scores. There was a significant difference in family income level such that youth who returned for two-year followup had overall higher income levels (U = 574, p < 0.05). Of the youth who returned for followup assessment, 42 (81%) maintained treatment response as defined by a 35% reduction in PARS score from pre-treatment.

Behavioral performance

Accuracy rates approached 100% and there were no differences in response rates between ANX and HC participants or therapy groups. See the Supplement for full statistics.

Main effects of task condition (fMRI)

Full results, summarized below, are visualized in Fig. 1 and tabulated in Table S2 in the Supplement.

Reward > control anticipation

Expected activations were seen in dorsal anterior cingulate (ACC), medial posterior cingulate cortex (PCC), dorsal striatum, ventral striatum and anterior insula. Expected activations were not seen in amygdala, thalamus, hippocampus, posterior cingulate, or subgenual ACC. Unexpected activations and deactivations were seen in occipital cortex.

Reward > control feedback

Expected activations were seen in anterior insula, dorsal ACC, medial PCC, dorsal striatum, ventral striatum, thalamus, and posterior cingulate. Expected activations were not seen in amygdala, hippocampus, and subgenual ACC. Unexpected activations were seen in dorsolateral prefrontal cortex (dlPFC), lateral parietal/angular gyri, and occipital cortex, and unexpected deactivations were seen in occipital cortex.

Primary hypotheses

Pre-treatment group differences: ANX vs HC

Full results, summarized below, are visualized in Fig. 2 and tabulated in Table 2. There were no significant results seen for ROI analyses of any of the below contrasts.

Fig. 2 Task main effects and baseline activity differences between early adolescents with anxiety disorders and healthy comparison youth. [Images not available. See PDF.]

Main effects visualized: a Reward > Control Anticipation and b Reward > Control Feedback contrasts. Group differences between ANX and HC youth visualized: c Reward > Control Anticipation and d Reward > Control Feedback contrasts. Blue = reduced activity in early adolescents with anxiety disorders (ANX) compared to healthy comparison (HC) youth; orange/yellow = heightened activity in ANX compared to HC youth. Accompanying color bar indicates the statistical value. Contrasts visualized have been adjusted for age and sex. Multiple comparison correction was done using cluster-based random field theory with a cluster forming threshold of Z = 3.1 and a corrected cluster p-value threshold of 0.05. Abbreviations: dlPFC dorsolateral prefrontal cortex, OFC orbitofrontal cortex, NAc nucleus accumbens, IFG inferior frontal gyrus, PCC posterior cingulate cortex.

Table 2. Significant fMRI activations for pairwise contrasts between baseline brain activity in adolescents with anxiety disorders vs. healthy comparison youth.

Peak voxel location	BA	Side	Cluster size (k)	Peak activation (MNI)			t
Reward > Control Approach Motivation
None
Control > Reward Anticipation
Orbitofrontal cortex	11	L	95	−10	24	−12	4.22
Thalamus	–	R	80	12	−10	8	4.45
Premotor/Supplementary motor	6	L	76	−52	10	28	4.18
Reward > Control Feedback
Visuomotor cortex	7	L	427	−6	−54	52	5.04
Angular gyrus	39	R	206	64	−54	16	4.53
Supramarginal gyrus	40	L	71	−52	−40	36	4.46
Inferior frontal gyrus	44	R	71	58	22	10	3.85
Control > Reward Feedback
None

Reward > control anticipation

In whole-brain analyses, compared with HC youth, ANX youth showed reduced activity in reward-related regions, namely right thalamus extending to caudate and left orbitofrontal cortex extending to caudate and ventral striatum. Activity was also reduced in a cluster in dlPFC.

Reward > control feedback

In whole-brain analyses, compared to HC youth, ANX youth had elevated activity to reward vs. control in right inferior frontal gyrus, bilateral angular gyrus, and precuneus.

Pre-treatment neural activity corresponding to treatment response

Reward > control anticipation

In whole-brain analyses, compared to non-responders, treatment responders showed reduced activation in two clusters in precuneus/cuneus (225 voxels, peak MNI coordinates −2, −78, 34, peak t = 4.46) and occipital pole (78 voxels, peak MNI coordinates −26, −66, 2, peak t = 3.96). These are visualized in Fig. 3.

Fig. 3 Brain activity differing between treatment responders and nonresponders. [Images not available. See PDF.]

Accompanying color bar indicates the statistical value for panels b) and c); blue = reduced activity in early adolescents who responded to treatment compared to those who did not respond to treatment. Subfigure c) presents clusters irrespective of statistical significance to visualize their relative locations; here, yellow denotes the ANX > HC pretreatment difference, and blue denotes the pre-to-post treatment change in activity. a) Reduced pretreatment brain activity for responders compared to nonresponders during Reward > Control Anticipation. b) Larger decerase in pre-to-post brain activity for responders compared to nonresponders during Reward > Control Feedback. c) Overlay of panel b) with the pretreatment ANX > HC activity difference illustrating that ANX participants had heightened activity in a region very nearby to the region which showed greater increase in treatment responders. Contrasts have been adjusted for age and sex. Multiple comparison correction was done using cluster-based random field theory with a cluster forming threshold of z = 3.1 and a corrected cluster p-value threshold of 0.05. Abbreviations: PCC posterior cingulate cortex.

Reward > control feedback

Whole brain analyses yielded no significant group differences in pre-treatment activity between youth who responded or did not respond to treatment.

Pre-to-post changes in neural activity and treatment response

Reward > control anticipation

Whole-brain analyses yielded no significant group differences in pre-to-post activity changes between youth who responded or did not respond to treatment.

Reward > control feedback

In whole-brain analyses, compared to non-responders, treatment responders showed decreased pre-to-post brain activity in left dorsal angular gyrus (97 voxels, peak MNI coordinates −60, −50, 44, peak t = 4.41). This result was no longer significant after controlling for baseline depression scores. This cluster was adjacent to but did not directly overlap with a cluster from the pretreatment ANX vs. HC contrast for this epoch described above (see Fig. 3).

Post-treatment normalization of function

In whole-brain and ROI analyses, compared to HC youth, no significant differences were seen in pre-to-post treatment brain activity for either all ANX youth or the subset of treatment responders in either task phase. There were no significant differences in whole-brain analyses between ANX and HC groups at post-treatment.

Exploratory analyses

Therapy type

Neither whole-brain nor ROI analyses significant differences between CBT and CCT for either task period.

Depressive symptoms at 2-year follow-up

Whole-brain analyses did not yield significant correlations.

Reward > control anticipation

In ROI analyses, controlling for pre-treatment activity, pre-to-post brain activity change in a PCC sphere (β = 17.98, t(48) = 2.94, p < 0.01) was positively correlated with MFQ score at two-year follow-up such that reductions in PCC activity corresponded with lower levels of depressive symptoms (Fig. 4). This result survived multiple comparison correction (adjusted p < .05) and remained significant when baseline depressive symptoms were controlled in sensitivity analyses.

Fig. 4 Scatterplots of relationships with pre-to-post changes in anticipation-related ROI data from posterior cingulate cortex and two-year follow-up depressive symptoms. [Images not available. See PDF.]

Y-axis represents absolute scores on the adolescent self-report measure, the Mood and Feelings Questionnaire (MFQ; see text for reference). X-axis represents individual parameter estimates for the Reward > Control Anticipation contrast calculated on the within-subject paired t-test maps (i.e., Post – Pre brain activity difference). Shaded bar represents the 95% confidence interval for predictions from the linear model. Red sphere indicates the location of the region of interest in the brain when registered to Montreal Neurological Institute (MNI) space. Results were robust to removal of one outlier with brain activity > 1000 (pictured).

Discussion

This study was among the largest to date to examine relationships between reward-related brain activity and response to psychological treatment in early adolescents with anxiety disorders Our findings partially support our a priori hypotheses, with unexpected results having novel implications for reward processing and its relationship to treatment response and later risk for depression in youth with anxiety disorders.

Consistent with prior findings, youth with anxiety disorders showed significant differences in pre-treatment activity in reward-related brain regions (e.g., ventral striatum, caudate) during reward anticipation. These are not in the expected direction of hyper-activation most commonly seen in prior studies [6], but rather hypo-activation. Methodological differences may explain this discrepancy: the majority of studies reviewed by Sequeira et al. [6] found that anxiety related to brain activity for high-magnitude compared to low-magnitude incentives. The present task did not contain incentives of varying magnitude, and thus may not have been able to elicit the sensitivity to high-magnitude incentives. Furthermore, we included both approach and ambiguous-approach trials, and as such there was an element of approach/avoid conflict within our Reward > Control Anticipation regressor. This decision was based on evidence that anxiety is not robustly related to aversion of monetary loss [47], and the low avoidance to loss cues in this task. Nevertheless, we did see elevated brain activity in the approach versus ambiguous-approach trials in regions associated with reward as well as cognitive control and salience (see Supplement), suggesting that these trials were less motivating. Slower reaction times in ambiguous trials supports this interpretation. Furthermore, although no differences were seen between ANX and HC youth in reaction times, intolerance of uncertainty is a known aspect of anxiety in adolescents [48], so it is possible that ANX youth processed these trials differentially. Overall, our results suggest that reward-related hypoactivity to small/moderate and/or ambiguous incentives may be a feature of anxiety disorders in adolescents.

However, ANX youth did show hyper-activation during reward feedback in one hypothesized region, the posterior cingulate (PCC), as well as inferior frontal gyrus and angular gyrus. The PCC is strongly associated with reward receipt in meta-analyses of reward-related activations [7, 49]. Furthermore, in a recent meta-analysis, all three of these regions (PCC, lateral PFC and angular gyrus) were found to be more reactive to reward receipt in adolescents than adults [49]. Notably both the PCC and the angular gyrus are nodes of the default-mode network (DMN) which is implicated in self-referential processes [50] and rumination [51]. We hypothesized that this heightened activity could be a result of rumination, which is known to interfere with reward processing in both anticipation and feedback phases [31, 32–33]. However, whole-brain relationships with self-reported rumination and worry were not significant (see Supplement). Thus, these activations may represent other processes that we did not directly measure. The PCC, in particular, is thought to integrate value information in order to optimize strategy in reward tasks [52]. Alternatively, given that the angular gyrus is a DMN region associated with the manipulation of conceptual knowledge and mental representations [30], ANX youth engage in more elaborative thought about the meaning of rewards.

Regarding the relationship of pre-treatment brain activity to treatment response, we did not observe feedback-elicited striatal/sgACC activity predicting treatment response, as reported in Sequeira’s findings from the same sample [20]. This could be a result of differences in task design between the two studies, as that task did not differentiate anticipation and feedback; it could also be a result of poor reliability in striatum and ACC. Instead, we found that treatment response corresponded to reduced activation in the posterior precuneus and occipital cortex, two regions that were recently reported as the main treatment effects in a meta-analysis of CBT across multiple disorders [25]. Furthermore, at least one study has found that reduction in cuneus/precuneus was associated with treatment response to CBT for depression in adults [53]. Although the function of these brain regions is unclear, there is precedent for it in the literature. The precuneus is associated with self-referential processing as part of the DMN [50], while occipital cortex is associated with visual processing and attention [54]; again, this suggests cognitive/attentional or self-referential processing.

Finally, we did not see any significant pre-to-post changes in early adolescents with anxiety disorders compared to healthy comparison youth, nor did we see any changes when examining only the adolescents who responded to treatment. Thus, we do not find evidence of normalization of function when compared to the healthy comparison group. This may be due to limited power and low reliability of results. Compared to treatment nonresponders, however, we found that those who responded to treatment showed pre-to-post activity decreases in left angular gyrus. Sensitivity analysis revealed that such changes were accounted for by pretreatment depressive symptoms, despite the fact that there were no differences between groups in pretreatment MFQ scores. A possible interpretation is that the impact of successful treatment on brain activity differed depending on the degree of depression prior to treatment. However, as above, it remains as unclear why activity in the angular gyrus would relate to pretreatment depressive symptoms, although it has previously been associated with symptom reduction in mood and anxiety disorders [55].

Because findings were observed across both CBT and CCT, pre-to-post treatment effects must be interpreted as representing common effects between therapeutic approaches. Nevertheless, exploratory analyses may provide clues about potential mechanistic differences between these therapies. Notably, the number of exposures completed and clinician-rated degree to which fears were confronted, which are unique to CBT, modulated activity in reward-related regions (e.g., dorsal striatum, dorsal ACC, PCC). Conversely, therapeutic alliance—a nonspecific therapy variable—related to more ventral striatal activity but also increases in PCC and angular gyrus. Overall, these results suggest impacts both of specific aspects of CBT as well as shared mechanisms between CBT and CCT. However, in the absence of significant differences between therapies, this interpretation remains speculative.

Taken together, our results suggest that, during reward processing, youth with anxiety disorders engage in differential processing utilizing PCC and angular gyrus compared to non-axnious peers. This may represent some form of conceptual processing, or expending excessing resources on reward-related processes, such as integrating value knowledge to decide on future actions [52]. Such an account would require additional experimental work to confirm, but if so, would suggest that an intervention to normalized cognition during the feedback period could be a useful adjunct to anxiety-focused therapies. Such an approach may be particularly relevant during early adolescence when self-referential and reward processes are increasing [56].

Interpretation of findings should consider strengths and limitations. First, our task had some methodological limitations; although designed to engage approach-avoid motivation, brain activity to avoid trials was not detected. We suspect this was due to low aversiveness of the loss, as it was very easy to avoid, and participants did so with nearly perfect accuracy. Nevertheless, we did observe expected activity to both anticipation and feedback. The time window for anticipation was also short (2 s), which may have limited sensitivity, and could not exclude fast motor responses. Second, our study had significant attrition, both at post-treatment MRI and at the follow-up timepoint when depression was assessed, and attrition related to parental education level. Out pre-to-post results must thus be interpreted considering this confound, but pretreatment results are not affected. Third, split-half reliability varied between task conditions, although the coefficients were well within the range of those previously reported [57]. Reliability coefficients were especially poor for Reward > Control contrasts in striatum and dACC; thus, our lack of results in these regions may not reflect absence of effect, but rather the poor internal consistency of the subtraction-based approach. Future work should consider alternative approaches such as residualized individual difference scores. Finally, we note that the lack of a waitlist control condition means that we cannot definitively attribute pre-to-post changes in anxious youth to treatment rather than spontaneous remission. The inclusion of a healthy control group, however, mitigates the likelihood of results occurring purely due to regression to the mean.

Notable strengths of this study include a large sample (N = 97) diagnosed with a gold-standard diagnostic interview (the K-SADS-PL) and assessed by independent evaluators blind to group and treatment status, and inclusion of a healthy comparison group. Inclusion of two active therapy groups means our results are more generalizable to psychological treatments other than CBT. Hypotheses were preregistered, and analyses were based on a modified version of a task that had previously been validated [41].

Overall, our findings provide new insights into the role of reward processing in psychological treatment for anxiety disorders in adolescents. Specifically, our results suggest that brain activity associated with self-referential or cognitive processes during reward processing may impact symptomatology and treatment in youth with anxiety disorders. Furthermore, our work expands on prior research showing impacts of psychological treatment on cognitive-emotional brain circuitry [27] to demonstrate brain activity corresponding with treatment success. Hence, targeting cognition during reward processing could represent an avenue to enhance treatment, particularly during early adolescence.

Acknowledgements

Supported by National Institute of Mental Health (NIMH) grant P50 MH080215 (PI: Ryan). Support for research participant recruitment was also provided by the Clinical and Translational Science Institute at the University of Pittsburgh (NIH/NCRR/CTSA grant UL1 RR024153). This research was supported in part by the University of Pittsburgh Center for Research Computing and Data, RRID:SCR_022735, through the resources provided. Specifically, this work used the HTC cluster, which is supported by NIH award number S10OD028483, and the H2P cluster, which is supported by NSF award number OAC-2117681. The authors would like to thank the participants and their families for participation in this study and Jeanette Mumford, PhD, for guidance on developing fMRI models.

Author contributions

CS: Formal analysis, software, visualization, writing (original draft, review and editing); CDL: conceptualization, supervision, review, editing. MS: conceptualization, methodology, supervision, review, editing. JSS, EEF, NDR, RED, DLM, PCK, AM: conceptualization, supervision, review. Previous presentations: Portions of this work were presented in a poster at the American College of Neuropsychopharmacology’s annual meeting (Phoenix, AZ; December 4–7, 2022).

Data availability

At the time of data collection, participants did not consent to public distribution of data, but did consent to release of de-identified data upon request. Therefore, data will be made available upon request.

Competing interests

Dr. Kendall receives royalties from the sale of materials related to the treatment of anxiety in youth. Dr. Forbes has received consulting fees from Brown University, George Mason University, University of California Berkeley, University of California Los Angeles, and Vanderbilt University. Dr. Ladouceur receives consulting fees from Boys Town National Research Hospital Institute for Human Neuroscience. The other authors report no financial relationships with commercial interests.

Ethics approval and consent to participate

All study procedures were approved by the University of Pittsburgh Institutional Review Board (IRB#PRO07110273) and all methods were performed in accordance with relevant guidelines and regulations. All parents provided written informed consent and all youth provided written assent after receiving a complete description of the study at the clinical intake visit.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41398-025-03388-2.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Barican, JL; Yung, D; Schwartz, C; Zheng, Y; Georgiades, K; Waddell, C. Prevalence of childhood mental disorders in high-income countries: a systematic review and meta-analysis to inform policymaking. Evid Based Ment Health; 2022; 25, pp. 36-44. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34281985]

2. Carpenter, JK; Andrews, LA; Witcraft, SM; Powers, MB; Smits, JAJ; Hofmann, SG. Cognitive behavioral therapy for anxiety and related disorders: a meta-analysis of randomized placebo-controlled trials. Depress Anxiety; 2018; 35, pp. 502-14. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29451967]

3. Bandelow, B; Reitt, M; Röver, C; Michaelis, S; Görlich, Y; Wedekind, D. Efficacy of treatments for anxiety disorders: a meta-analysis. Int Clin Psychopharmacol; 2015; 30, pp. 183-92. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25932596]

4. Marwood, L; Wise, T; Perkins, AM; Cleare, AJ. Meta-analyses of the neural mechanisms and predictors of response to psychotherapy in depression and anxiety. Neurosci Biobehav Rev; 2018; 95, pp. 61-72. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30278195][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6267850]

5. Auerbach, RP; Pagliaccio, D; Hubbard, NA; Frosch, I; Kremens, R; Cosby, E et al. Reward-related neural circuitry in depressed and anxious adolescents: a human connectome project. J Am Acad Child Adolesc Psychiatry; 2022; 61, pp. 308-20. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33965516]

6. Sequeira, SL; Forbes, EE; Hanson, JL; Silk, JS. Positive valence systems in youth anxiety development: a scoping review. J Anxiety Disord; 2022; 89, 102588. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35691120]

7. Silverman, MH; Jedd, K; Luciana, M. Neural networks involved in adolescent reward processing: an activation likelihood estimation meta-analysis of functional neuroimaging studies. Neuroimage; 2015; 122, pp. 427-39. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26254587]

8. Larsen, B; Luna, B. Adolescence as a neurobiological critical period for the development of higher-order cognition. Neurosci Biobehav Rev; 2018; 94, pp. 179-95. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30201220][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6526538]

9. Mackin DM, Nelson BD, Klein DN Chapter 40 - reward processing and depression: current findings and future directions. In: Martin CR, Hunter LA, Patel VB, Preedy VR, Rajendram R, editors. The neuroscience of depression [Internet]. London, UK: Academic Press; 2021 [cited 2024 Mar 13]. p. 425–33. Available from: https://www.sciencedirect.com/science/article/pii/B9780128179352000519.

10. Crone, EA; van Duijvenvoorde, ACK; Peper, JS. Annual research review: neural contributions to risk-taking in adolescence – developmental changes and individual differences. J Child Psychol Psychiatry; 2016; 57, pp. 353-68. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26889896]

11. Berridge, KC; Robinson, TE; Aldridge, JW. Dissecting components of reward: ‘liking’, ‘wanting’, and learning. Curr Opin Pharmacol; 2009; 9, pp. 65-73. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19162544][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2756052]

12. Development and Definitions of the RDoC Domains and Constructs - National Institute of Mental Health (NIMH) [Internet]. [cited 2024 Sep 20]. Available from: https://www.nimh.nih.gov/research/research-funded-by-nimh/rdoc/development-and-definitions-of-the-rdoc-domains-and-constructs.

13. Zhang, Y; Li, Q; Wang, Z; Liu, X; Zheng, Y. Temporal dynamics of reward anticipation in the human brain. Biol Psychol; 2017; 128, pp. 89-97. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28735969]

14. Newman, MG; Shin, KE; Zuellig, AR. Developmental risk factors in generalized anxiety disorder and panic disorder. J Affect Disord; 2016; 206, pp. 94-102. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27466747][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5077703]

15. Admon, R; Pizzagalli, DA. Dysfunctional reward processing in depression. Curr Opin Psychol; 2015; 4, pp. 114-8. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26258159]

16. Weinberg, A; Kujawa, A; Riesel, A. Understanding trajectories to anxiety and depression: neural responses to errors and rewards as indices of susceptibility to stressful life events. Curr Dir Psychol Sci; 2022; 31, pp. 115-23.

17. Silk, JS; Davis, S; McMakin, DL; Dahl, RE; Forbes, EE. Why do anxious children become depressed teenagers? The role of social evaluative threat and reward processing. Psychol Med; 2012; 42, pp. 2095-107. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22340187][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3360132]

18. Zbozinek, TD; Craske, MG. The role of positive affect in enhancing extinction learning and exposure therapy for anxiety disorders. J Exp Psychopathol; 2017; 8, pp. 13-39.

19. Nielson, DM; Keren, H; O’Callaghan, G; Jackson, SM; Douka, I; Vidal-Ribas, P et al. Great expectations: a critical review of and suggestions for the study of reward processing as a cause and predictor of depression. Biol Psychiatry; 2021; 89, pp. 134-43. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32797941]

20. Sequeira, SL; Silk, JS; Ladouceur, CD; Hanson, JL; Ryan, ND; Morgan, JK et al. Association of neural reward circuitry function with response to psychotherapy in youths with anxiety disorders. Am J Psychiatry; 2021; 178, pp. 343-51. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33472390][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8016705]

21. Silk, JS; Tan, PZ; Ladouceur, CD; Meller, S; Siegle, GJ; McMakin, DL et al. A randomized clinical trial comparing individual cognitive behavioral therapy and child-centered therapy for child anxiety disorders. J Clin Child Adolesc Psychol; 2018; 47, pp. 542-54. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26983904]

22. Ladouceur, CD. Can cognitive-behavioral therapy normalize neural function in youths with pediatric anxiety disorders? A developmental neuroscience perspective. Am J Psychiatry; 2024; 181, pp. 175-7. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38425260]

23. Silk, JS; Price, RB; Rosen, D; Ryan, ND; Forbes, EE; Siegle, GJ et al. A longitudinal follow-up study examining adolescent depressive symptoms as a function of prior anxiety treatment. J Am Acad Child Adolesc Psychiatry; 2019; 58, pp. 359-67. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30768411]

24. Schrammen, E; Roesmann, K; Rosenbaum, D; Redlich, R; Harenbrock, J; Dannlowski, U et al. Functional neural changes associated with psychotherapy in anxiety disorders – a meta-analysis of longitudinal fMRI studies. Neurosci Biobehav Rev; 2022; 142, 104895. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36179918]

25. Yuan, S; Wu, H; Wu, Y; Xu, H; Yu, J; Zhong, Y et al. Neural effects of cognitive behavioral therapy in psychiatric disorders: a systematic review and activation likelihood estimation meta-analysis. Front Psychol; 2022; 13, 853804. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35592157][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9112423]https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2022.853804

26. La Buissonniere-Ariza, V; Fitzgerald, K; Meoded, A; Williams, LL; Liu, G; Goodman, WK et al. Neural correlates of cognitive behavioral therapy response in youth with negative valence disorders: a systematic review of the literature. J Affect Disord; 2021; 282, pp. 1288-307. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33601708]

27. Haller, SP; Linke, JO; Grassie, HL; Jones, EL; Pagliaccio, D; Harrewijn, A et al. Normalization of fronto-parietal activation by cognitive-behavioral therapy in unmedicated pediatric patients with anxiety disorders. Am J Psychiatry; 2024; 181, pp. 201-12. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38263879][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11423803]

28. LeDoux, J; Daw, ND. Surviving threats: neural circuit and computational implications of a new taxonomy of defensive behaviour. Nat Rev Neurosci; 2018; 19, pp. 269-82. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29593300]

29. Vincent, JL; Kahn, I; Snyder, AZ; Raichle, ME; Buckner, RL. Evidence for a frontoparietal control system revealed by intrinsic functional connectivity. J Neurophysiol; 2008; 100, pp. 3328-42. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18799601][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2604839]

30. Seghier, ML. The angular gyrus: multiple functions and multiple subdivisions. Neuroscientist; 2013; 19, pp. 43-61. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22547530][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4107834]

31. Kocsel, N; Szabó, E; Galambos, A; Édes, A; Pap, D; Elliott, R et al. Trait rumination influences neural correlates of the anticipation but not the consumption phase of reward processing. Front Behav Neurosci; 2017; 11, 85. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28539875][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5423966]https://www.frontiersin.org/journals/behavioral-neuroscience/articles/10.3389/fnbeh.2017.00085/full

32. Erdman, A; Abend, R; Jalon, I; Artzi, M; Gazit, T; Avirame, K et al. Ruminative tendency relates to ventral striatum functionality: evidence from task and resting-state fMRI. Front Psychiatry; 2020; 11, 67. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32153443][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7044182]https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2020.00067/full

33. Maresh, EL; Allen, JP; Coan, JA. Increased default mode network activity in socially anxious individuals during reward processing. Biol Mood Anxiety Disord; 2014; 4, 7. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25075275][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4114426]

34. Aupperle, RL; Melrose, AJ; Francisco, A; Paulus, MP; Stein, MB. Neural substrates of approach-avoidance conflict decision-making. Hum Brain Mapp; 2015; 36, pp. 449-62. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25224633]

35. Bergman H, Maayan N, Kirkham AJ, Adams CE, Soares‐Weiser K. Schedule for Affective Disorders and Schizophrenia for School‐Age Children (K‐SADS) for diagnosing schizophrenia in children and adolescents with psychotic symptoms. Cochrane Database of Systematic Reviews 2015, Issue 6. Art. No.: CD011733. https://doi.org/10.1002/14651858.CD011733. Accessed 12 May 2025.

36. The research units on pediatric psychopharmacology anxiety study. The pediatric anxiety rating scale (PARS): development and psychometric properties. J Am Acad Child Adolesc Psychiatry; 2002; 41, pp. 1061-9.

37. Caporino, NE; Brodman, DM; Kendall, PC; Albano, AM; Sherrill, J; Piacentini, J et al. Defining treatment response and remission in child anxiety: signal detection analysis using the pediatric anxiety rating scale. J Am Acad Child Adolesc Psychiatry; 2013; 52, pp. 57-67. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23265634]

38. Angold, A; Costello, EJ. Mood and feelings questionnaire (MFQ); 1987; Durham, Developmental Epidemiology Program, Duke University:

39. Kendall PC, Hedtke KA The coping cat program workbook. Ardmore, PA: Workbook Publishing; 2006.

40. Cohen, JA; Deblinger, E; Mannarino, AP; Steer, RA. A multisite, randomized controlled trial for children with sexual abuse–related PSTD symptoms. J Am Acad Child Adolesc Psychiatry; 2004; 43, pp. 393-402. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15187799][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1201422]

41. Schlund, MW; Siegle, GJ; Ladouceur, CD; Silk, JS; Cataldo, MF; Forbes, EE et al. Nothing to fear? Neural systems supporting avoidance behavior in healthy youths. Neuroimage; 2010; 52, pp. 710-9. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20430103]

42. Esteban, O; Markiewicz, CJ; Blair, RW; Moodie, CA; Isik, AI; Erramuzpe, A et al. fMRIPrep: a robust preprocessing pipeline for functional MRI. Nat Methods; 2019; 16, pp. 111-6. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30532080]

43. FSL - FslWiki [Internet]. [cited 2022 Sep 14]. Available from: https://fsl.fmrib.ox.ac.uk/fsl/fslwiki.

44. Eklund, A; Nichols, TE; Knutsson, H. Cluster failure: why fMRI inferences for spatial extent have inflated false-positive rates. Proc Natl Acad Sci USA; 2016; 113, pp. 7900-5. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27357684][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4948312]

45. Yarkoni, T; Poldrack, RA; Nichols, TE; Van Essen, DC; Wager, TD. Large-scale automated synthesis of human functional neuroimaging data. Nat Methods; 2011; 8, pp. 665-70. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21706013][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3146590]

46. Holm, S. A simple sequentially rejective multiple test procedure. Scand J Stat; 1979; 6, pp. 65-70.

47. Charpentier, CJ; Aylward, J; Roiser, JP; Robinson, OJ. Enhanced risk aversion, but not loss aversion, in unmedicated pathological anxiety. Biol Psychiatry; 2017; 81, pp. 1014-22. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28126210][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5466268]

48. Dugas, MJ; Laugesen, N; Bukowski, WM. Intolerance of uncertainty, fear of anxiety, and adolescent worry. J Abnorm Child Psychol; 2012; 40, pp. 863-70. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22302481]

49. Jauhar, S; Fortea, L; Solanes, A; Albajes-Eizagirre, A; McKenna, PJ; Radua, J. Brain activations associated with anticipation and delivery of monetary reward: a systematic review and meta-analysis of fMRI studies. PLoS ONE; 2021; 16, e0255292. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34351957][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8341642]

50. Raichle, ME. The brain’s default mode network. Annu Rev Neurosci; 2015; 38, pp. 433-47. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25938726]

51. Berman, MG; Peltier, S; Nee, DE; Kross, E; Deldin, PJ; Jonides, J. Depression, rumination and the default network. Soc Cogn Affect Neurosci; 2011; 6, pp. 548-55. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20855296]

52. Dohmatob, E; Dumas, G; Bzdok, D. Dark control: the default mode network as a reinforcement learning agent. Hum Brain Mapp; 2020; 41, pp. 3318-41. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32500968][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7375062]

53. Rubin-Falcone, H; Weber, J; Kishon, R; Ochsner, K; Delaparte, L; Doré, B et al. Neural predictors and effects of cognitive behavioral therapy for depression: the role of emotional reactivity and regulation. Psychol Med; 2020; 50, pp. 146-60. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30739618]

54. Rosen, ML; Sheridan, MA; Sambrook, KA; Peverill, MR; Meltzoff, AN; McLaughlin, KA. The role of visual association cortex in associative memory formation across development. J Cogn Neurosci; 2018; 30, pp. 365-80. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29064341]

55. Liu, Q; Ely, BA; Schwartz, JJ; Alonso, CM; Stern, ER; Gabbay, V. Reward function as an outcome predictor in youth with mood and anxiety symptoms. J Affect Disord; 2021; 278, pp. 433-42. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33010568]

56. Crone, EA; Dahl, RE. Understanding adolescence as a period of social–affective engagement and goal flexibility. Nat Rev Neurosci; 2012; 13, pp. 636-50. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22903221]

57. Luking, KR; Nelson, BD; Infantolino, ZP; Sauder, CL; Hajcak, G. Internal consistency of functional magnetic resonance imaging and electroencephalography measures of reward in late childhood and early adolescence. Biol Psychiatry Cogn Neurosci Neuroimaging; 2017; 2, pp. 289-97. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29057369]