Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Jun 1.
Published in final edited form as: J Affect Disord. 2023 Mar 4;330:309–318. doi: 10.1016/j.jad.2023.02.149

Amygdala and Nucleus Accumbens Activation During Reward Anticipation Moderates the Association Between Life Stressor Frequency and Depressive Symptoms

Alyssa N Fassett-Carman 1,*, Amelia D Moser 2, Luka Ruzic 2, Chiara Neilson 2, Jenna Jones 1, Sofia Barnes-Horowitz 3, Christopher D Schneck 4, Roselinde H Kaiser 1,2
PMCID: PMC10695433  NIHMSID: NIHMS1884494  PMID: 36871909

Abstract

Background:

Life stressors confer risk for depressive symptoms, but individuals vary in the extent of their sensitivity to life stressors. One protective factor may be an individual’s level of reward sensitivity, e.g., a stronger neurobiological response to environmental rewards may mitigate emotional responses to stressors. However, the nature of neurobiological reward sensitivity that corresponds with stress resilience is unknown. Further, this model is untested in adolescence, when life stressor frequency and depression increase.

Methods:

We tested the hypothesis that stronger reward-related activation in the left and right nucleus accumbens (NAc), amygdala, and medial prefrontal cortex (mPFC) attenuates the strength of the stress-depression relation. We measured BOLD activation throughout Win and Lose blocks of a monetary reward task, as well as during anticipation and outcome phases of the task. Participants (N=151, ages 13–19) were recruited to be stratified on risk for mood disorders to enhance variance in depressive symptoms.

Results:

Activation during anticipation of rewards in the bilateral amygdala and NAc, but not mPFC, buffered the association between life stressors and depressive symptoms. This buffering effect was not found for reward outcome activation or activation across Win blocks.

Conclusions:

Results highlight the importance of reward anticipation activation of subcortical structures in attenuating the stress-depression link, suggesting that reward motivation may be a cognitive mechanism through which this stress buffering occurs.

Introduction

Depression, a prevalent mood disorder, frequently onsets during adolescence and emerging adulthood (Merikangas et al., 2010; Rohde, Lewinsohn, Klein, Seeley, & Gau, 2013). Exposure to life stressors is a robust risk factor for depression (Grant et al., 2014) that increases during adolescence as individuals transition towards independence, relationships with peers become more complex, and academic pressure rises (Hankin et al., 2016). However, not all adolescents who experience life stressors develop depressive symptoms, thus there must be factors that impact the stress-depression link. One such factor may be neural reward system (RS) function.

It is well known that stressor exposure can blunt RS function (Auerbach, Admon, & Pizzagalli, 2014; Pizzagalli, 2014), which is associated with depressive symptoms (Toenders et al., 2019). One commonly investigated and well-supported model is that life stressor exposure leads to reward system dysfunction, which is associated with anhedonic symptoms (Pizzagalli, 2014). However, the relation between stress and reward systems may be bidirectional. For example, rewarding experiences can attenuate acute stress reactivity (e.g., severity appraisals, physiological and neural reactivity; Christiansen et al., 2011; Dutcher et al., 2020; Speer & Delgado, 2017), and initial studies show that strength of activation in the ventral striatum (VS) in response to rewarding tasks buffers the link between life stressors and depressive symptoms (Corral-Frías et al., 2015; Nikolova, Bogdan, Brigidi, & Hariri, 2012). These findings lend support to the model that a strong, trait-like neural response to rewards may protect against the effects of life stressors in exacerbating depressive symptoms (Dutcher & Creswell, 2018).

However, there are still many gaps in our understanding of this buffering effect. Critically, it is unknown whether reward sensitivity has protective effects during adolescence, a key time of RS development. Across adolescent development, reward sensitivity exhibits a curvilinear relation with age, such that RS activation increases in early adolescence, peaks in mid-adolescence, and then subsequently plateaus or declines (Casey, Getz, & Galvan, 2008; Shulman et al., 2016). This change in neural RS sensitivity is accompanied by concomitant increases and decreases in behavioral reward sensitivity including sensation seeking and risk taking (Shulman et al., 2016), and coincides with a rise in depression prevalence (Rohde et al., 2013). Due to these ongoing developmental changes during adolescence, relations between reward system function, life stressor exposure, and depressive symptoms may differ from those observed in college-age samples (Corral-Frías et al., 2015; Nikolova et al., 2012). If a strong reward system confers protection against stress-related depressive symptoms throughout adolescence, knowledge about the mechanisms underlying this effect could be harnessed for intervention to mitigate depression onset.

Further, the nature of neurobiological reward sensitivity that may correlate with stress-protective traits is poorly understood. Prior research has implicated VS reward sensitivity in stress-buffering (Corral-Frías et al., 2015; Nikolova et al., 2012), but it is unknown whether other regions involved in reward and stress processing, such as the amygdala and mPFC, are involved as well. The mPFC is consistently activated in response to rewards (Haber & Knutson, 2010), and additionally plays a crucial role in stress regulation, with regulatory projections to brainstem and limbic regions, including the amygdala (Arnsten, 2009; Datta & Arnsten, 2019; Maier & Seligman, 2016). Greater mPFC activation to rewards is associated with lower acute and future stress reactivity (Vidal-Ribas et al., 2019), suggesting that the mPFC reward response may reflect a stress-buffering mechanism. However, it is unknown whether the buffering effect extends to longer-term impacts of life stressors such as depressive symptoms. In turn, the amygdala is well known for its role in activating the stress response, but it is additionally interconnected with the VS and mPFC, is active during reward learning, and has a well-documented role in reward-related processing and behavior (Janak & Tye, 2015; Wassum & Izquierdo, 2015). Importantly, studies in rodents have demonstrated that amygdala lesions prevent the buffering effect of food rewards on physiological stress reactivity (Ulrich-Lai et al., 2010), suggesting a key role for the amygdala in buffering acute stress reactivity. Whether or not amygdala activation in humans buffers the life stressor-depression link remains unknown, however. It is important to investigate this broader network of regions beyond the VS to better understand which aspects of RS function protect against the negative consequences of life stressors.

Finally, it is unknown how RS activation during distinct phases of reward processing contribute to attenuating the life stressor-depression link. Previous studies utilized a block design, in which activation is measured across task blocks characterized primarily by either rewards or losses (Corral-Frías et al., 2015; Nikolova et al., 2012). This approach has been important for illustrating that RS activation moderates the stress-depression relation, but it is unable to distinguish whether activation during reward anticipation, reward outcome, or both, drive the buffering effect. Neural responses to reward anticipation versus outcome are associated with different cognitive processes: motivational drive for rewards versus liking of rewards, respectively, (Berridge, Robinson, & Aldridge, 2009; Wang, Smith, & Delgado, 2016). Understanding whether activation during one or both of these reward processes is associated with an attenuated link between life stressors and depressive symptoms will help elucidate possible cognitive mechanisms through which this effect occurs.

The current study addresses these gaps by testing the hypothesis that RS activation during acute rewarding experiences in a priori ROIs, the NAc, amygdala, and mPFC, is associated with an attenuated relation between life stressor frequency and depressive symptoms in a sample of adolescents ages 13–19. We used an ROI approach for targeted hypothesis testing of these specific regions given their established RS roles and anatomical and functional connections to stress-processing neural systems, but we additionally conducted a whole-brain exploratory analysis to test whether other reward-responsive regions may be important in buffering the relation between life stressor frequency and depression symptoms. All information relating to the whole-brain analysis can be found in Supplemental Materials. To increase our ability to capture links between risk factors and depressive symptoms and increase the generalizability of our results across levels of risk for psychopathology, participants were recruited at low- and high-risk for mood disorders based on family history. We measured BOLD responses in ROIs across task blocks characterized primarily by reward versus loss, as well as used an event-related design to measure BOLD responses to anticipation and outcome of rewards. We tested whether RS activation across highly rewarded blocks and activation specifically during reward anticipation and reward outcome interacts with life stressor frequency to predict depressive symptoms. Analyses were preregistered on Open Science Framework1.

Method

Participants

Participants were 151 adolescents (see Table 1 for sample demographic information) recruited from the Boulder/Denver metropolitan area. Participants were either high- or low-risk for mood disorders. High-risk recruitment was operationalized as having a first-degree family member (i.e., biological parent or sibling) with a diagnosed lifetime history of a mood disorder, including unipolar disorders (n=64), bipolar disorders (n=31) or disruptive mood dysregulation disorder (n=1). Low-risk recruitment (n=55) was defined by individuals with no first- or second-degree family members (i.e., biological parents, siblings, or grandparents) with a history of a mood disorder. Recruitment methods included participant portals for faculty, students, staff, and others at the University of Colorado Boulder, University email newsletter, ResearchMatch, social media, flyers in public places and on the University campus, community outreach and direct mail, and recruitment of (high-risk) participants whose family members were patients at Boulder/Denver area clinics. Individuals who were identified as potentially eligible by clinics, or the parent/guardian of potentially eligible minors, were informed that the decision of whether to be screened for and participate in the study would not influence their medical care. Inclusion criteria were being in the 13–192 age range, fluent in English, right-handed, able to pass the standard MRI safety screen at the neuroimaging facility, and not currently in a major depressive episode, manic episode, or hypomanic episode. Exclusion criteria included any of the following: having a lifetime full criteria diagnosis of any psychotic disorder, moderate-to-severe substance use disorders, or eating disorders, currently experiencing suicidal ideation to an extent deemed unsafe by the study clinician, history of a seizure disorder, head trauma, or physical disorders that affect brain functioning or blood flow, neurological abnormalities, substance use within 24 hours prior to the neuroimaging session, use of stimulant medication or medication with cardiac effects within 48 hours prior to the neuroimaging session, antipsychotic medications or changes to any ongoing prescribed medications within six weeks prior to the neuroimaging session, and colorblindness.

Table 1:

Sample Demographics

Sex Assigned at Birth
 60% Female
 40% Male
Age
 Mean = 16.48
 SD = 1.91
Gender
 56% Women
 39% Men
 3% Nonbinary
 1% Transgender
 1% Did not report
Race
 84% White
 10% Multiracial
 4% Asian
 1% Black or African American
 1% Other
Ethnicity
 90% Not Hispanic or Latino/a/x
 10% Hispanic or Latino/a/x
Family Income
 37% Greater than $100,000
 21% $75,000–100,000
 18% $50,000–75,000
 14% $25,000–50,000
 8% $10,000–25,000
 2% Less than $10,000
Open in a new tab

Procedure

Participants over 18 or a parent/guardian of participants under 18 provided informed consent, and participants younger than 18 provided informed assent. Participation in the study involved two visits to the lab one year apart and a follow-up clinical evaluation two years after the initial visit. During lab visits, participants completed a neuroimaging session, cognitive tasks, a clinical interview, and questionnaires. In between lab visits, participants periodically responded to daily diary prompts, completed longer questionnaires, and participated in phone interviews. The longitudinal study is not yet complete. The current study uses data from Time Point 1 collected before December 31, 2021.

Questionnaire and Interview Measures

Participants completed a demographics survey in which they indicated their sex assigned at birth as “male” or “female” and reported their age.

Participants reported current medication use during the clinical interview. Dichotomized variables were created to indicate use of stimulants, beta blockers, selective serotonin reuptake inhibitors (SSRIs), norepinephrine and dopamine reuptake inhibitors, atypical antipsychotics, anticonvulsants, benzodiazepines, anxiolytics, lithium, tricyclics, anticholinergics, and unknown medications.

The Life Events Interview (LEI; Safford et al., 2007) is an interview form of the Life Events Scale (LES; Francis-Raniere, Alloy, & Abramson, 2006; Needles & Abramson, 1990) that assesses the frequency and perceived intensity of life events over the past year. The current study uses a subset of LES items selected based on expert consultation with an author of the LES to represent domains particularly salient to adolescents (e.g., school, relationships, friendships) and piloting on a separate sample of participants to select a set of items that showed coverage across the sample (i.e., each item was endorsed by at least one participant).

A trained interviewer asked participants about 24 negative life events (e.g., “Did poorly on or failed an exam or major paper or project in an important class”) and 22 positive events (e.g., “Received an ‘A’ on an exam or major project in an important class”), and participants indicated the number of times each event occurred for them over the past year. Event frequency was coded as 0=this event did not happen, 1=happened once, 2=happened twice, 3=happened three times, 4=happened four or more times. Frequency scores were calculated by summing frequency codes for each participant. Participants additionally rated how intense their endorsed events felt to them on a scale from 0 (“Not intense”) to 9 (“Extremely intense”). Only frequency scores for negative items were used in the current analyses. The LEI shows good reliability and validity (Safford et al., 2007).

The Mood and Anxiety Symptom Questionnaire (MASQ; Watson et al., 1995a, b) is a self-report measure that assesses depression and anxiety symptoms. The current study included subscales assessing anxious arousal (e.g., “Startled easily”) and loss of interest (e.g., “Felt like there isn’t anything interesting or fun to do”). Participants rated how often they experienced each symptom over the past two weeks on a 5-point scale, 1=Not at all, 2=A little bit, 3=Moderately, 4=Quite a bit, and 5=Extremely. The MASQ shows good reliability and validity (Watson et al., 1995a, b). Only the loss of interest subscale was used in the current analyses.

fMRI Tasks

The Dice Task (Figure 1) is a modified version of the Human Connectome Project (HCP) Lifespan Gambling Task (Delgado et al., 2000; Barch et al., 2013). In each ~10 second trial, participants guess whether the values of two rolled dice will sum to greater or less than seven. Trials are either reward trials, in which correct guesses win money and incorrect guesses yield neutral outcomes (no money won or lost), or loss trials, in which incorrect guesses lose money and correct guesses yield neutral outcomes.

Figure 1. Dice Task Example Trial.

Figure 1

Open in a new tab

Note. When presented with the Guess stimulus, participants have three seconds to indicate their response. Their response (higher or lower) is highlighted on the screen until the full three seconds have elapsed. The Dice Roll animation that follows corresponds to the Anticipation period, and the Monetary Outcome slide corresponds to the Outcome event. A jittered fixation cross separates trials. The example trial shown depicts a reward trial from the Win block resulting in a monetary reward of $10. Alternative stimuli used for Lose blocks (loss trials) are shown in the Additional Stimuli box, along with the neutral outcome stimulus. Win blocks are characterized by a 2:1 likelihood of winning (versus receiving a neutral outcome, no money won) across trials; Lose blocks are characterized by a 2:1 likelihood of losing (versus receiving a neutral outcome, no money lost) across trials. Outcomes are standardized across participants.

Participants complete one run consisting of four Win blocks and four Lose blocks. Win blocks consist of six reward trials in which participants have a 2:1 likelihood of winning money versus receiving a neutral outcome. Lose blocks consist of six loss trials in which participants have a 2:1 likelihood of losing money versus receiving a neutral outcome. Outcomes are standardized across participants. The principles of this task are the same as the HCP Lifespan Gambling Task, but the task presentation was modified based on piloting to be more engaging for participants, using a rolling dice animation during anticipation of the outcome rather than static images. For event-related design analyses, the Anticipation periods were defined as the 3s rolling dice animations, and the Outcome periods were defined as the 3s monetary outcome screen.

MRI Acquisition Parameters

Imaging data were collected using a Siemens 3-Tesla MAGNETOM Prisma scanner and a 32-channel head coil at the Intermountain Neuroimaging Consortium. We acquired a high-resolution T1-weighted anatomical scan (TR=2400ms, TE=2.07ms, flip angle=8 degrees, 224 slices 0.8mm thick, field of view=256mm, voxel size=0.8×0.8×0.8mm) and echo-planar imaging (EPI) functional task and rest scans (identical to the Adolescent Brain Cognitive Development [ABCD] study sequences [Casey et al., 2018], TR=800ms, TE=30ms, flip angle=52 degrees, 60 slices 2.4mm thick, field of view=216mm, voxel size=2.4×2.4×2.4mm, multiband acceleration factor=6). Out of the functional scans, only the Dice Task scan (volumes=1010, ~14 minutes) was used in the current analyses.

fMRI Analyses

Preprocessing and Functional Models

We followed the HCP minimal preprocessing pipeline for functional data (Glasser et al., 2013) and used FMRIB’s ICA-based X-noiseifier (FIX; Griffanti et al., 2014; Salimi-Khorshidi et al., 2014) to denoise the data.

First-level models were run in FSL (Jenkinson, Beckmann, Behrens, Woolrich, & Smith, 2012; Smith et al., 2004) to estimate mean reward-related activation across voxels for each ROI. Data were spatially smoothed with a 6mm kernel and high-pass filtered with a cutoff of 200s. For block design analyses, the Win blocks and Lose blocks of the task were modeled and convolved with the canonical double-gamma hemodynamic response function (HRF) to create regressors for the GLMs predicting single-voxel time series. For event-related design analyses, Dice Task events (i.e., Reward Anticipation, Loss Anticipation, Reward Outcome, Loss Outcome, and Neutral Outcome) were modeled and convolved with the double-gamma (HRF) to create regressors for the GLMs. The twelve motion parameters calculated by the HCP minimum preprocessing pipeline (i.e., translation and rotation in the x, y, and z directions for each volume using the first volume as the reference image, and their derivatives) were included as regressors as well. We used FSL’s prewhitening function to account for autocorrelation of residuals in the data.

Contrasts were used to estimate functional activation associated with particular aspects of reward experiences. In the block design, a Win>Lose contrast compared activation during Win versus Lose blocks. In the event-related design, reward outcome activation was estimated using both Reward Outcome>Loss Outcome and Reward Outcome> Neutral Outcome contrasts due to differences in the information they capture. Reward Outcome>Loss Outcome is thought to hold salience of events relatively constant to measure activation associated uniquely with reward. In the Reward Outcome>Neutral Outcome contrast, salience is not held constant but estimates of activation are valence-specific. Anticipatory reward activation was estimated using a Reward Anticipation>Loss Anticipation contrast. Contrasts of parameter estimates (cope values) were extracted from all voxels within each ROI and averaged to get a mean activation estimate for each ROI for each participant.

ROI Parcellation (Figure S1)

Amygdala ROIs were parcellated according to the Automated Anatomical Atlas (AAL; Tzourio-Mazoyer et al., 2002) and the NAc ROIs were parcellated using the Individual Brain Atlases using Statistical Parametric Mapping tool (IBASPM; Alemán-Gómez et al., 2006). The mPFC ROIs were made by combining Gordon atlas mPFC parcels (Left: 29, 116, 117, 152; Right: 183, 184, 278, 279; Gordon et al., 2016) that largely overlapped with mPFC activation identified in meta-analyses of reward-related activation, including Neurosynth (Diekhof, Kaps, Falkai, & Gruber, 2012; Jauhar et al., 2021; Oldham et al., 2018; Yarkoni, Poldrack, Nichols, Van Essen, & Wager, 2011).

Group-level Hypothesis Testing

Linear regressions tested whether ROI activation moderated the link between life stressor frequency and depressive symptoms, controlling for age and sex. Three regressions were run for each ROI, testing the interactions between life stressor frequency and: (1) block design reward activation (Win vs. Lose block contrast), (2) Reward Anticipation (Reward Anticipation vs. Loss Anticipation event contrast) and Reward Outcome (Reward Outcome vs. Loss Outcome event contrast), and (3) Reward Anticipation (Reward Anticipation vs. Loss Anticipation event contrast) and Reward Outcome (Reward Outcome vs. Neutral Outcome event contrast). FDR correction was used to correct for multiple comparisons. Follow-up analyses tested whether results were consistent when controlling for familial risk status and medication use. We further tested whether the RS stress buffering results differed between adolescents with or without a family history of mood disorders via analyses including a three-way interaction between familial risk, ROI activation, and stressor frequency.

Results

Analyses were completed using RStudio version 2021.9.1.372 (RStudio Team, 2021; See Supplemental Materials for packages used).

Of the 151 eligible participants enrolled in the study before December 31, 2021, two participants did not complete the neuroimaging session and four participants moved excessively during the reward task (framewise displacement greater than 0.5mm for over 15% of volumes) and were excluded from analysis. Six participants were missing LES data (n=5 did not participate in the interview due to time constraints during the study session, n=1 data were lost). Thus, the number of participants retained in regression analyses was 140.

Descriptive Statistics

Negative life event frequency was positively skewed (>1; Table S1) so a square-root transformation of the variable was used in all subsequent analyses. Expected significant correlations between life stressor frequency and depressive symptoms, and BOLD signal elicited by the reward task among ROIs, were present (Table S2).

Life Stressor-Depression Relation

Stressor frequency was significantly associated with depressive symptom severity while controlling for age and sex (β=0.328, p<.001).

Reward-Related BOLD Activation in ROIs

Significant reward-related BOLD activation (percent signal change) was observed in ROIs (ts>9.236, ps<.001; Table S3).

Block Design Stress-Buffering Analyses (Table S4)

No stressor frequency by ROI reward activation interactions passed FDR correction in their association with depression (qs>.264).

Event-Related Design Stress-Buffering Analyses

Reward>Loss Outcome Contrast, Reward>Loss Anticipation Contrast (Table 2).

Table 2:

Event-Related Design ROI Reward Anticipation (Reward Anticipation>Loss Anticipation) and Outcome (Reward Outcome>Loss Outcome) Activation, Life Stress, and their Interaction1 Predicting Depressive Symptoms

Measure b β SE Est./SE p q
Left Amygdala Stress Frequency** 1.164 0.341 0.280 4.156 <.001
Outcome Activation 0.004 0.013 0.021 0.167 .868
Anticipation Activation 0.024 0.086 0.022 1.086 .279
Sex* 0.973 0.165 0.468 2.080 .039
Age* −0.852 −0.277 0.241 −3.534 .001
Stress Frequency x Outcome Activation 0.012 0.075 0.013 0.959 .340 .603
Stress Frequency x Anticipation Activation * −0.043 −0.265 0.013 −3.368 .001 .006
Right Amygdala Stress Frequency** 1.127 0.330 0.285 3.953 <.001
Outcome Activation −0.014 −0.050 0.023 −0.639 .524
Anticipation Activation 0.028 0.098 0.023 1.202 .232
Sex 0.859 0.145 0.481 1.787 .076
Age* −0.849 −0.276 0.245 −3.459 .001
Stress Frequency x Outcome Activation 0.004 0.024 0.013 0.307 .759 .759
Stress Frequency x Anticipation Activation * −0.038 −0.230 0.013 −2.847 .005 .015
Left NAc Stress Frequency** 1.144 0.335 0.282 4.052 <.001
Outcome Activation 0.018 0.070 0.020 0.904 .368
Anticipation Activation 0.001 0.006 0.020 0.074 .941
Sex 0.931 0.157 0.473 1.968 .051
Age* −0.810 −0.263 0.242 −3.342 .001
Stress Frequency x Outcome Activation −0.009 −0.065 0.010 −0.842 .402 .603
Stress Frequency x Anticipation Activation * −0.030 −0.195 0.012 −2.503 .014 .027
R NAc Stress Frequency** 1.152 0.337 0.279 4.128 <.001
Outcome Activation 0.016 0.059 0.022 0.763 .447
Anticipation Activation −0.004 −0.016 0.021 −0.194 .847
Sex* 0.992 0.168 0.485 2.044 .043
Age* −0.813 −0.264 0.243 −3.347 .001
Stress Frequency x Outcome Activation −0.006 −0.035 0.013 −0.463 .644 .759
Stress Frequency x Anticipation Activation * −0.031 −0.190 0.013 −2.387 .018 .027
Left mPFC Stress Frequency** 1.150 0.337 0.280 4.103 <.001
Outcome Activation 0.027 0.117 0.018 1.475 .142
Anticipation Activation 0.008 0.036 0.017 0.456 .649
Sex 0.790 0.134 0.489 1.618 .108
Age* −0.821 −0.267 0.244 −3.372 .001
Stress Frequency x Outcome Activation 0.011 0.091 0.010 1.133 .259 .603
Stress Frequency x Anticipation Activation −0.020 −0.151 0.011 −1.812 .072 .072
Right mPFC Stress Frequency** 1.156 0.339 0.280 4.126 <.001
Outcome Activation 0.024 0.106 0.018 1.323 .188
Anticipation Activation 0.007 0.035 0.017 0.433 .666
Sex 0.879 0.149 0.480 1.829 .070
Age* −0.821 −0.267 0.244 −3.359 .001
Stress Frequency x Outcome Activation 0.010 0.079 0.010 0.942 .348 .603
Stress Frequency x Anticipation Activation* −0.022 −0.172 0.011 −1.993 .048 .058
Open in a new tab

Note.

*

p<.05

**

p<.001. Bold=q<.05 across ROIs within the family of anticipation or outcome analyses. Sex coded 1=female, −1=male.

No stressor frequency by reward Outcome activation interactions passed FDR correction (qs>.603). Stressor frequency by reward Anticipation activation interactions were significantly associated with depressive symptoms after FDR correction for the bilateral amygdala (Left: β=−0.265, p=.001, q=.006; Right: β=−0.230, p=.005, q=.015) and bilateral NAc (Left: β=−0.195, p=.014, q=.027; Right: β=−0.190, p=.018, q=.027; Figures 2 & 3). The stressor frequency by mPFC reward Anticipation activation interactions were marginally associated with depressive symptoms after FDR correction (Left: β=−0.151, p=.072, q=.072; Right: β=−0.172, p=.048, q=.058).

Figure 2. Reward Anticipation Activation Stress Buffering in the Amygdala and NAc.

Figure 2

Open in a new tab

Note. Right and left hemisphere amygdala, NAc, and mPFC ROIs are displayed. Statistics describe results for interactions between life stress and ROI activation during reward anticipation predicting depressive symptoms. Reward Anticipation contrast was always Reward Anticipation>Loss Anticipation. Two sets of regressions are displayed: one controlling for Reward Outcome calculated using a Reward Outcome>Loss Outcome contrast and one controlling for Reward Outcome calculated using a Reward Outcome>Neutral Outcome contrast. Bold=passed FDR correction; non-bold black=FDR correction was marginal (.05<q<.1); gray=FDR correction was not significant (q>.1).

Figure 3. Loop Plots of ROI Reward Anticipation Activation x Stress Interactions.

Figure 3

Open in a new tab

(a) from regressions controlling for reward outcome activation calculated using a Reward Outcome>Loss Outcome contrast

(b) from regressions controlling for reward outcome activation calculated using a Reward Outcome>Neutral Outcome contrast

Note. NAc=nucleus accumbens; mPFC=medial prefrontal cortex; cope=contrast of parameter estimates; b=unstandardized coefficient; 95% confidence intervals shown. Reward Anticipation contrast was always Reward Anticipation>Loss Anticipation. ROI anticipation activation cope values are mean centered. Blue dotted line indicates the point at which the stress-depression relation is no longer significant.

Reward>Neutral Outcome Contrast, Reward>Loss Anticipation Contrast (Table 3).

Table 3:

Event-Related Design ROI Reward Anticipation (Reward Anticipation>Loss Anticipation) and Outcome (Reward Outcome>Neutral Outcome) Activation, Life Stress, and their Interaction Predicting1 Depressive Symptoms

Measure b β SE Est./SE p q
Left Amygdala Stress Frequency** 1.138 0.333 0.273 4.174 <.001
Outcome Activation <0.001 <0.001 0.023 −0.001 .999
Anticipation Activation 0.021 0.073 0.023 0.921 .358
Sex* 0.938 0.159 0.467 2.008 .047
Age* −0.813 −0.264 0.240 −3.392 .001
Stress Frequency x Outcome Activation −0.012 −0.063 0.014 −0.808 .420 .504
Stress Frequency x Anticipation Activation * −0.042 −0.260 0.013 −3.288 .001 .006
Right Amygdala Stress Frequency** 1.129 0.331 0.275 4.114 <.001
Outcome Activation −0.024 −0.076 0.025 −0.953 .342
Anticipation Activation 0.024 0.086 0.023 1.058 .292
Sex 0.831 0.141 0.477 1.741 .084
Age* −0.826 −0.268 0.241 −3.426 .001
Stress Frequency x Outcome Activation −0.013 −0.077 0.013 −0.977 .331 .497
Stress Frequency x Anticipation Activation * −0.041 −0.249 0.013 −3.109 .002 .006
Left NAc Stress Frequency** 1.108 0.324 0.276 4.011 <.001
Outcome Activation 0.024 0.087 0.022 1.091 .277
Anticipation Activation −0.003 −0.011 0.020 −0.144 .886
Sex 0.906 0.153 0.470 1.926 .056
Age* −0.804 −0.261 0.240 −3.348 .001
Stress Frequency x Outcome Activation −0.017 −0.132 0.011 −1.630 .106 .497
Stress Frequency x Anticipation Activation * −0.033 −0.212 0.013 −2.593 .011 .017
Right NAc Stress Frequency** 1.115 0.327 0.277 4.022 <.001
Outcome Activation 0.011 0.036 0.023 0.468 .640
Anticipation Activation −0.005 −0.020 0.021 −0.252 .801
Sex 0.959 0.162 0.484 1.982 .050
Age* −0.811 −0.264 0.242 −3.347 .001
Stress Frequency x Outcome Activation −0.015 −0.090 0.013 −1.152 .251 .497
Stress Frequency x Anticipation Activation * −0.034 −0.211 0.013 −2.583 .011 .017
Left mPFC Stress Frequency** 1.116 0.327 0.282 3.963 <.001
Outcome Activation 0.009 0.035 0.020 0.447 .656
Anticipation Activation 0.009 0.040 0.017 0.501 .617
Sex 0.926 0.157 0.487 1.901 .060
Age* −0.796 −0.259 0.247 −3.227 .002
Stress Frequency x Outcome Activation −0.005 −0.031 0.012 −0.393 .695 .695
Stress Frequency x Anticipation Activation −0.016 −0.120 0.011 −1.484 .140 .140
Right mPFC Stress Frequency** 1.086 0.318 0.279 3.889 <.001
Outcome Activation 0.008 0.029 0.021 0.370 .712
Anticipation Activation 0.006 0.030 0.017 0.366 .715
Sex* 0.981 0.166 0.480 2.042 .043
Age* −0.789 −0.256 0.244 −3.227 .002
Stress Frequency x Outcome Activation −0.015 −0.101 0.011 −1.295 .197 .497
Stress Frequency x Anticipation Activation −0.017 −0.135 0.010 −1.689 .094 .113
Open in a new tab

Note.

*

p<.05

**

p<.001. Bold=q<.05 across ROIs within the family of anticipation or outcome analyses. Sex coded 1=female, −1=male.

Stressor frequency by ROI reward Outcome activation interactions did not pass FDR correction (qs>.497). Stressor frequency by reward Anticipation activation interactions were significantly associated with depressive symptoms after FDR correction for the bilateral amygdala (Left: β=− 0.260, p=.001, q=.006; Right: β=−0.249, p=.002, q=.006) and NAc (Left: β=−0.212, p=.011, q=.017; Right: β=−0.211, p=.011, q=.017; Figures 2 & 3). Stress frequency by mPFC reward Anticipation activation interactions did not pass FDR correction (qs>.113).

Follow-Up Analyses: Familial Risk and Medication Use

Medications present in our sample included stimulants, beta blockers, SSRIs, and unknown medications. The pattern of stress-buffering results across ROIs did not change when controlling for familial risk or medication use (Tables S5–S7). Familial risk did not moderate the associations between ROI activation x stressor frequency interaction terms and depressive symptoms (Tables S8–S10).

Discussion

The current study tested whether stronger reward-related activation in the NAc, amygdala, and mPFC was associated with an attenuated association between life stressors and depressive symptoms in adolescence. As hypothesized, greater reward-related activation in the NAc and amygdala was associated with a weaker association between life stressors and depressive symptoms, consistent with a stress-buffering model. Effects that passed FDR correction were specific to subcortical regions during reward Anticipation, although right and left mPFC activation during reward Anticipation marginally buffered the stress-depression link after FDR correction when controlling for reward Outcome activation as calculated using the Reward Outcome>Loss Outcome contrast. These results build on previous evidence to support a model in which greater neural reward sensitivity is associated with stress-buffering effects (Corral-Frías et al., 2015; Nikolova et al., 2012).

These results were consistent with past work demonstrating that the strength of VS activity to acute rewards was associated with an attenuated link between life stress and depression (Corral-Frías et al., 2015; Nikolova et al., 2012). Our results extended this work by demonstrating that activation during anticipation of rewards, rather than reward outcomes, was key to this protective effect. Activation during reward anticipation is thought to be associated with motivational drive to gain rewards (Berridge et al., 2009; Wang et al., 2016), giving rise to possible mechanisms through which stronger activation may be associated with a weaker stress-depression link. First, greater NAc and amygdala activation during reward anticipation may be associated with greater motivational drive to bring about rewarding experiences, and the consequent exposure to positive events attenuates the stress-depression relation. This interpretation is consistent with evidence that the frequency of positive events can buffer the link between negative events and depressive symptoms (Riskind, Kleiman, & Schafer, 2013; Shahar & Priel, 2002). While there is limited research linking reward activation in the lab to positive events in participants’ lives, one PET study found that greater dopamine release in the right VS and caudate to rewards in the lab is associated with the extent to which enjoyment of active engagement increases the probability of future active engagement (Kasanova et al., 2017). Although the PET study did not investigate dopaminergic activity specifically during reward anticipation, it provides initial evidence that RS activity in the lab is associated with real-world motivational drive. Understanding how anticipatory reward activation in the NAc and amygdala relate to real world behaviors and events will be important to determining the mechanisms of the stress-buffering effects identified in the current study.

Second, it is possible that anticipatory reward activation in the NAc and amygdala is associated not only with motivation for rewards, but also motivation to cope when faced with stressors. Animal research suggests that similar neurobiological mechanisms underlie approach motivation and defensive coping (Berridge, 2019). For example, dopamine release in the NAc is associated with motivation for rewards as well as active coping behaviors when stressors are present (Cabib & Puglisi-Allegra, 2012). Further, manipulating glutamatergic and dopaminergic activity in the NAc, and CRF activity in the amygdala, can promote either approach or defensive behaviors depending on context (Berridge, 2019). If similar neurobiological mechanisms bring about motivated behavior towards rewards and stressful experiences, individual differences in these types of behaviors may be related, such that an individual with stronger approach motivation to rewards will similarly exhibit stronger motivation to cope with stressors when they arise. However, to our knowledge, this model has not been tested and is a direction for future work.

Although stress-buffering effects in the mPFC did not significantly pass FDR correction, interactions between life stressor frequency and bilateral mPFC reward anticipation activation were marginally associated with depression after FDR correction when controlling for mPFC reward outcome activation using the Reward Outcome>Loss Outcome contrast. It is possible that mPFC activation to reward anticipation does play a role in protecting against depressive symptoms when faced with life stressors, but perhaps follows a different pattern than subcortical structures. For example, there is evidence that although the mPFC is active during reward processing, high levels of mPFC activity are associated with inhibition of the NAc, which may play a role in blunting responses to rewards as a result of chronic stress (Ironside, Kumar, Kang, & Pizzagalli, 2018). Perhaps there is an optimal level of mPFC activation to rewards that reflects protection from life stressors, while high levels of mPFC activation to rewards reflect a vulnerability in the link between life stressors and depressive symptoms. Further, some research suggests that the mPFC is more likely to show activation during reward outcome, thought to reflect processing of hedonic aspects of rewards, rather than reward anticipation (Oldham et al., 2018). In this model, the role of mPFC may be less relevant to protection against the negative impact of stressors via proposed approach motivation mechanisms described above. These possibilities should be investigated in the future.

Limitations and Future Directions

The current study has limitations which will be important avenues for future research. First, this study used an adolescent sample, which is beneficial as it captures a period of increased life stressor exposure, emergence of depression, and reward system development, but may not be generalizable to other stages of development. Second, the presented analyses are cross-sectional, preventing us from investigating directionality of effects, although the retrospective reporting on life stress captured a period that preceded the measurement of current symptom severity. Longitudinal analyses can be conducted once Time Point 2 data collection is complete.

Another limitation is that using a Reward>Loss contrast for calculating RS anticipation activation, it is unclear whether greater activation to reward anticipation, reduced activation to loss anticipation, or both, are driving the moderation of the stress-depression link. We chose a task design in which the primary anticipation contrast aims to hold constant the salience of outcomes that are anticipated (i.e., anticipating the loss or gain of monetary rewards of similar magnitude and likelihood). However, another approach to monetary reward task design is to compare anticipation of reward with anticipation of neutral outcomes. A recent meta-analysis demonstrated that anticipation of rewards and losses compared to anticipation of neutral outcomes both activate VS and amygdala, suggesting that they may be involved in valence-general motivation system and/or involved in processes such as task engagement that are likely greater during conditions with a reward or loss compared to neutral outcome (Oldham et al., 2018). In the current study, NAc and amygdala activation for anticipating rewards relative to anticipating losses was significantly greater than zero (Table S3), suggesting that we did capture activation unique to anticipation of rewards over losses. However, a future direction will be to test the stress-buffering model with a task paradigm that allows for Reward>Neutral and Loss>Neutral anticipation contrasts to better determine which individual differences in anticipatory neural activation are driving the NAc and VS stress buffering identified in the current study.

A strength of the current study is the robust, hypothesis-driven use of a priori ROIs to test the question of whether reward-related function in canonical RS regions is associated with an attenuated stress-depression link during adolescence. Given that these ROIs, the NAc, amygdala, and mPFC, are structurally and functionally connected, it will be important in the future to investigate whether individual differences in coordinated neural activity between these regions is important for buffering the link between life stressors and depressive symptoms.

Conclusion

Adolescents who showed stronger activation in the NAc and amygdala during anticipation of rewards exhibited an attenuated association between life stressor frequency and depressive symptoms. These findings elucidate a protective neurobiological process at a time of development in which depression tends to emerge. Future work should test cognitive and behavioral mechanisms through which this buffering occurs, which may be targets for treatment or prevention of depression.

Supplementary Material

1

Highlights.

  • Life stressor exposure and depressive symptoms are strongly linked

  • Strong neural response to rewards buffers stress-depression link in adolescents

  • Stress buffering is specific to neural activity during reward anticipation

  • Reward-related neural activity may be protective against stress-related depression

Acknowledgements:

We thank Amy Hegarty for support with data analysis and Teryn Wilkes, the head MRI technologist at the Intermountain Neuroimaging Consortium for supporting the collection of high quality brain imaging data. We additionally thank all others who helped with data collection.

Financial Support:

This work is supported by the National Institute of Mental Health (R01 MH117131).

Footnotes

Conflicts of Interest: None

Ethical Standards: The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

Author Statement

Alyssa N. Fassett-Carman contributed to study conceptualization, analysis, and writing, reviewing, and editing the manuscript. Amelia D. Moser contributed to reviewing and editing the manuscript. Luka Ruzic contributed to data analysis. Chiara Neilson contributed to project management and data curation. Jenna Jones contributed to project management, data curation, and reviewing and editing the manuscript. Sofia Barnes-Horowitz contributed to data presentation and reviewing and editing the manuscript. Christopher D. Schneck contributed to obtaining funding and reviewing and editing the manuscript. Roselinde H. Kaiser contributed to obtaining funding, study conceptualization, reviewing and editing the manuscript.

All authors have approved the submitted manuscript.

2

One participant who was 19 at the time of recruitment turned 20 prior to the neuroimaging session.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Alemán-Gómez Y, Melie-Garcia L, & Valdés-Hernández P. (2006). IBASPM: toolbox for automatic parcellation of brain structures [CD]. In: Proceedings of the 12th Annual Meeting of the Organization for Human Brain Mapping, June 11–15, Florence, Italy. NeuroImage. 27. [Google Scholar]
  2. Arnsten AFT (2009). Stress signalling pathways that impair prefrontal cortex structure and function. Nature Reviews Neuroscience, 10(6), 410–422. 10.1038/nrn2648 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Auerbach RP, Admon R, & Pizzagalli DA (2014). Adolescent Depression: Stress and Reward Dysfunction. Harvard Review of Psychiatry, 22(3), 139–148. 10.1097/HRP.0000000000000034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Barch DM, Burgess GC, Harms MP, Petersen SE, Schlaggar BL, Corbetta M, … Van Essen DC (2013). Function in the human connectome: Task-fMRI and individual differences in behavior. NeuroImage, 80, 169–189. 10.1016/j.neuroimage.2013.05.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Berridge KC (2019). Affective valence in the brain: Modules or modes? 25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Berridge KC, Robinson TE, & Aldridge JW (2009). Dissecting components of reward: ‘Liking’, ‘wanting’, and learning. Current Opinion in Pharmacology, 9(1), 65–73. 10.1016/j.coph.2008.12.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bray S, Shinsuke S, & O’Doherty JP (2010). Human Medial Orbitofrontal Cortex Is Recruited During Experience of Imagined and Real Rewards. Journal of Neurophysiology, 103(5), 2506–2512. 10.1152/jn.01030.2009 [DOI] [PubMed] [Google Scholar]
  8. Cabib S, & Puglisi-Allegra S. (2012). The mesoaccumbens dopamine in coping with stress. Neuroscience & Biobehavioral Reviews, 36(1), 79–89. 10.1016/j.neubiorev.2011.04.012 [DOI] [PubMed] [Google Scholar]
  9. Casey BJ, Cannonier T, Conley MI, Cohen AO, Barch DM, Heitzeg MM, … Dale AM (2018). The Adolescent Brain Cognitive Development (ABCD) study: Imaging acquisition across 21 sites. Developmental Cognitive Neuroscience, 32, 43–54. 10.1016/j.dcn.2018.03.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Casey BJ, Getz S, & Galvan A. (2008). The adolescent brain. Developmental Review, 28(1), 62–77. 10.1016/j.dr.2007.08.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Christiansen AM, DeKloet AD, Ulrich-Lai YM, & Herman JP (2011). “Snacking” causes long term attenuation of HPA axis stress responses and enhancement of brain FosB/deltaFosB expression in rats. Physiology & Behavior, 103(1), 111–116. 10.1016/j.physbeh.2011.01.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Corral-Frías NS, Nikolova YS, Michalski LJ, Baranger DAA, Hariri AR, & Bogdan R. (2015). Stress-related anhedonia is associated with ventral striatum reactivity to reward and transdiagnostic psychiatric symptomatology. Psychological Medicine, 45(12), 2605–2617. 10.1017/S0033291715000525 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Datta D, & Arnsten A. (2019). Loss of Prefrontal Cortical Higher Cognition with Uncontrollable Stress: Molecular Mechanisms, Changes with Age, and Relevance to Treatment. Brain Sciences, 9(5), 113. 10.3390/brainsci9050113 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Diekhof EK, Kaps L, Falkai P, & Gruber O. (2012). The role of the human ventral striatum and the medial orbitofrontal cortex in the representation of reward magnitude – An activation likelihood estimation meta-analysis of neuroimaging studies of passive reward expectancy and outcome processing. Neuropsychologia, 50(7), 1252–1266. 10.1016/j.neuropsychologia.2012.02.007 [DOI] [PubMed] [Google Scholar]
  15. Dutcher JM, & Creswell JD (2018). The role of brain reward pathways in stress resilience and health. Neuroscience & Biobehavioral Reviews, 95, 559–567. 10.1016/j.neubiorev.2018.10.014 [DOI] [PubMed] [Google Scholar]
  16. Dutcher JM, Eisenberger NI, Woo H, Klein WMP, Harris PR, Levine JM, & Creswell JD (2020). Neural mechanisms of self-affirmation’s stress buffering effects. Social Cognitive and Affective Neuroscience, 15(10), 1086–1096. 10.1093/scan/nsaa042 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Francis-Raniere EL, Alloy LB, & Abramson LY (2006). Depressive personality styles and bipolar spectrum disorders: Prospective tests of the event congruency hypothesis. Bipolar Disorders, 8(4), 382–399. 10.1111/j.1399-5618.2006.00337.x [DOI] [PubMed] [Google Scholar]
  18. Glasser MF, Sotiropoulos SN, Wilson JA, Coalson TS, Fischl B, Andersson JL, … Jenkinson M. (2013). The minimal preprocessing pipelines for the Human Connectome Project. NeuroImage, 80, 105–124. 10.1016/j.neuroimage.2013.04.127 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gordon EM, Laumann TO, Adeyemo B, Huckins JF, Kelley WM, & Petersen SE (2016). Generation and Evaluation of a Cortical Area Parcellation from Resting-State Correlations. Cerebral Cortex, 26(1), 288–303. 10.1093/cercor/bhu239 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Grant KE, McMahon SD, Carter JS, Carleton RA, Adam EK, & Chen E. (2014). The Influence of Stressors on the Development of Psychopathology. In Lewis M. & Rudolph KD (Eds.), Handbook of Developmental Psychopathology (pp. 205–223). Boston, MA: Springer US. 10.1007/978-1-4614-9608-3_11 [DOI] [Google Scholar]
  21. Griffanti L, Salimi-Khorshidi G, Beckmann CF, Auerbach EJ, Douaud G, Sexton CE, … Smith SM (2014). ICA-based artefact removal and accelerated fMRI acquisition for improved resting state network imaging. NeuroImage, 95, 232–247. 10.1016/j.neuroimage.2014.03.034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Haber SN, & Knutson B. (2010). The Reward Circuit: Linking Primate Anatomy and Human Imaging. Neuropsychopharmacology, 35(1), 4–26. 10.1038/npp.2009.129 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hankin BL, Snyder HR, Gulley LD, Schweizer TH, Bijttebier P, Nelis S, … Vasey MW (2016). Understanding comorbidity among internalizing problems: Integrating latent structural models of psychopathology and risk mechanisms. Development and Psychopathology, 28(4pt1), 987–1012. 10.1017/S0954579416000663 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Ironside M, Kumar P, Kang M-S, & Pizzagalli DA (2018). Brain mechanisms mediating effects of stress on reward sensitivity. Current Opinion in Behavioral Sciences, 22, 106–113. 10.1016/j.cobeha.2018.01.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Janak PH, & Tye KM (2015). From circuits to behaviour in the amygdala. Nature, 517(7534), 284–292. 10.1038/nature14188 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Jauhar S, Fortea L, Solanes A, Albajes-Eizagirre A, McKenna PJ, & Radua J. (2021). Brain activations associated with anticipation and delivery of monetary reward: A systematic review and meta-analysis of fMRI studies. PLOS ONE, 16(8), e0255292. 10.1371/journal.pone.0255292 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jenkinson M, Beckmann CF, Behrens TEJ, Woolrich MW, & Smith SM (2012). FSL. NeuroImage, 62(2), 782–790. 10.1016/j.neuroimage.2011.09.015 [DOI] [PubMed] [Google Scholar]
  28. Kasanova Z, Ceccarini J, Frank MJ, Amelsvoort T. van, Booij J, Heinzel A, … Myin-Germeys I. (2017). Striatal dopaminergic modulation of reinforcement learning predicts reward—Oriented behavior in daily life. Biological Psychology, 127, 1–9. 10.1016/j.biopsycho.2017.04.014 [DOI] [PubMed] [Google Scholar]
  29. Maier SF, & Seligman MEP (2016). Learned helplessness at fifty: Insights from neuroscience. Psychological Review, 123(4), 349–367. 10.1037/rev0000033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Merikangas KR, He J, Burstein M, Swanson SA, Avenevoli S, Cui L, … Swendsen J. (2010). Lifetime Prevalence of Mental Disorders in U.S. Adolescents: Results from the National Comorbidity Survey Replication–Adolescent Supplement (NCS-A). Journal of the American Academy of Child & Adolescent Psychiatry, 49(10), 980–989. 10.1016/j.jaac.2010.05.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Needles DJ, & Abramson LY (1990). Positive Life Events, Attributional Style, and Hopefulness: Testing a Model of Recovery From Depression [Data set]. Journal of Abnormal Psychology, Vol. 99, pp. 156–165. 10.1037/e334482004-007 [DOI] [PubMed] [Google Scholar]
  32. Nikolova YS, Bogdan R, Brigidi BD, & Hariri AR (2012). Ventral Striatum Reactivity to Reward and Recent Life Stress Interact to Predict Positive Affect. Biological Psychiatry, 72(2), 157–163. 10.1016/j.biopsych.2012.03.014 [DOI] [PubMed] [Google Scholar]
  33. Oldham S, Murawski C, Fornito A, Youssef G, Yücel M, & Lorenzetti V. (2018). The anticipation and outcome phases of reward and loss processing: A neuroimaging meta-analysis of the monetary incentive delay task. Human Brain Mapping, 39(8), 3398–3418. 10.1002/hbm.24184 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Pizzagalli DA (2014). Depression, Stress, and Anhedonia: Toward a Synthesis and Integrated Model. Annual Review of Clinical Psychology, 10(1), 393–423. 10.1146/annurev-clinpsy-050212-185606 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Riskind JH, Kleiman EM, & Schafer KE (2013). “Undoing” Effects of Positive Affect: Does it Buffer the Effects of Negative Affect in Predicting Changes in Depression? Journal of Social and Clinical Psychology, 32(4), 363–380. 10.1521/jscp.2013.32.4.363 [DOI] [Google Scholar]
  36. Rohde P, Lewinsohn PM, Klein DN, Seeley JR, & Gau JM (2013). Key Characteristics of Major Depressive Disorder Occurring in Childhood, Adolescence, Emerging Adulthood, and Adulthood. Clinical Psychological Science, 1(1), 41–53. 10.1177/2167702612457599 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. RStudio Team (2021). RStudio: Integrated Development Environment for R. RStudio, PBC, Boston, MA: URL http://www.rstudio.com/. [Google Scholar]
  38. Safford SM, Alloy LB, Abramson LY, & Crossfield AG (2007). Negative cognitive style as a predictor of negative life events in depression-prone individuals: A test of the stress generation hypothesis. Journal of Affective Disorders, 99(1–3), 147–154. 10.1016/j.jad.2006.09.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Salimi-Khorshidi G, Douaud G, Beckmann CF, Glasser MF, Griffanti L, & Smith SM (2014). Automatic denoising of functional MRI data: Combining independent component analysis and hierarchical fusion of classifiers. NeuroImage, 90, 449–468. 10.1016/j.neuroimage.2013.11.046 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Shahar G, & Priel B. (2002). Positive Life Events and Adolescent Emotional Distress: In Search of Protective-Interactive Processes. Journal of Social and Clinical Psychology, 21(6), 645–668. 10.1521/jscp.21.6.645.22798 [DOI] [Google Scholar]
  41. Shulman EP, Smith AR, Silva K, Icenogle G, Duell N, Chein J, & Steinberg L. (2016). The dual systems model: Review, reappraisal, and reaffirmation. Developmental Cognitive Neuroscience, 17, 103–117. 10.1016/j.dcn.2015.12.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ, Johansen-Berg H, … Matthews PM (2004). Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage, 23, S208–S219. 10.1016/j.neuroimage.2004.07.051 [DOI] [PubMed] [Google Scholar]
  43. Speer ME, & Delgado MR (2017). Reminiscing about positive memories buffers acute stress responses. Nature Human Behaviour, 1(5), 0093. 10.1038/s41562-017-0093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Toenders YJ, van Velzen LS, Heideman IZ, Harrison BJ, Davey CG, & Schmaal L. (2019). Neuroimaging predictors of onset and course of depression in childhood and adolescence: A systematic review of longitudinal studies. Developmental Cognitive Neuroscience, 39, 100700. 10.1016/j.dcn.2019.100700 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, … Joliot M. (2002). Automated Anatomical Labeling of Activations in SPM Using a Macroscopic Anatomical Parcellation of the MNI MRI Single-Subject Brain. NeuroImage, 15(1), 273–289. 10.1006/nimg.2001.0978 [DOI] [PubMed] [Google Scholar]
  46. Ulrich-Lai YM, Christiansen AM, Ostrander MM, Jones AA, Jones KR, Choi DC, … Herman JP (2010). Pleasurable behaviors reduce stress via brain reward pathways. Proceedings of the National Academy of Sciences, 107(47), 20529–20534. 10.1073/pnas.1007740107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Vidal-Ribas P, Benson B, Aria D. Vitale, Keren H, Harrewijn A, Fox NA, … Stringaris A. (2019). Bidirectional Associations Between Stress and Reward Processing in Children and Adolescents: A Longitudinal Neuroimaging Study. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 4(10), 893–901. 10.1016/j.bpsc.2019.05.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wang KS, Smith DV, & Delgado MR (2016). Using fMRI to study reward processing in humans: Past, present, and future. Journal of Neurophysiology, 115(3), 1664–1678. 10.1152/jn.00333.2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Wassum KM, & Izquierdo A. (2015). The basolateral amygdala in reward learning and addiction. Neuroscience & Biobehavioral Reviews, 57, 271–283. 10.1016/j.neubiorev.2015.08.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Watson D, AnnaClark L, Weber K, & Assenheimer JS (1995). Testing a Tripartite Model: II. Exploring the Symptom Structure of Anxiety and Depression in Student, Adult, and Patient Samples. Journal of Abnormal Psychology, 104(1), 15–25. [DOI] [PubMed] [Google Scholar]
  51. Watson D, Weber K, SmithAssenheimer J, Strauss ME, & McCormick RA (1995). Testing a Tripartite Model: I. Evaluating the Convergent and Discriminant Validity of Anxiety and Depression Symptom Scales. Journal of Abnormal Psychology, 104(1), 3–14. [DOI] [PubMed] [Google Scholar]
  52. Yarkoni T, Poldrack RA, Nichols TE, Van Essen DC, & Wager TD (2011). Large-scale automated synthesis of human functional neuroimaging data. Nature Methods, 8(8), 665–670. 10.1038/nmeth.163 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1884494-supplement-1.docx (5MB, docx)

RESOURCES