Pre-pandemic mental health and brain characteristics predict adolescent stress and emotions during the COVID-19 pandemic

Abstract

The COVID-19 pandemic had profound effects on developing adolescents that, to date, remain incompletely understood. Youth with preexisting mental health problems and associated brain alterations were at increased risk for higher stress and poor mental health. This study investigated impacts of adolescent pre-pandemic mental health problems and their neural correlates on stress, negative emotions and poor mental health during the first 15 months of the COVID-19 pandemic. N = 2,641 adolescents (median age = 12.0 years) from the Adolescent Brain Cognitive Development (ABCD) cohort were studied, who had pre-pandemic data on anxiety, depression, and behavioral (attention, aggression, social withdrawal, internalizing, externalizing) problems, longitudinal survey data on mental health, stress and emotions during the first 15 months following the outbreak, structural MRI, and resting-state fMRI. Data were analyzed using mixed effects mediation and moderation models. Preexisting mental health and behavioral problems predicted higher stress, negative affect and negative emotions (β = 0.09–0.21, CI=[0.03,0.32]), and lower positive affect (β = −0.21 to −0.09, CI=[−0.31,-0.01]) during the first ~6 months of the outbreak. Pre-pandemic structural characteristics of brain regions supporting social function and emotional processing (insula, superior temporal gyrus, orbitofrontal cortex, and the cerebellum) mediated some of these relationships (β = 0.10–0.15, CI=[0.01,0.24]). The organization of pre-pandemic brain circuits moderated (attenuated) associations between preexisting mental health and pandemic stress and negative emotions (β = −0.17 to −0.06, CI=[−0.27,-0.01]). Preexisting mental health problems and their structural brain correlates were risk factors for youth stress and negative emotions during the early months of the outbreak. In addition, the organization of some brain circuits was protective and attenuated the effects of preexisting mental health issues on youth responses to the pandemic’s stressors.

1. Introduction

Environmental and experiential factors during adolescence contribute to profound and complex biological changes [1–4] with important implications for physical and mental health [5–8]. Negative experiences and environmental stressors during this sensitive period significantly increase the risk of neural miswiring (aberrant connections between neurons, ensembles, and/or brain regions) [9–13], which, in turn, can lead to mental health issues and the emergence of neuropsychiatric disorders [7,14–16].

Recent studies have specifically sought to elucidate how environmental stressors associated with the COVID-19 pandemic impacted adolescent mental health and brain development. They have shown that prolonged stress, social isolation, school-related disruptions, decreased physical activity, and increased social media use, had extensive negative effects on brain development and mental health [17,18], especially in girls and youth from racio-ethnic minorities [19,20]. Preexisting mental health problems that are common in adolescents, such as anxiety, depression, internalizing and externalizing problems, increased vulnerability to these stressors [21–23]. A study on approximately 5,000 youth from the Adolescent Brain Cognitive Development (ABCD) cohort [24] reported exacerbated depressive symptoms, attention problems, and aggressive behaviors in youth with preexisting mental health problems [25]. Another study in about 8,000 youth from the same cohort showed that history of adverse childhood experiences was a significant risk factor for poor mental health, increased stress, and fear during the pandemic, especially in those with preexisting internalizing problems [19]. Recent work on brain characteristics as risk or protective factors has shown that stronger and more resilient neural circuits before the outbreak were protective, whereas weaker and more fragile ones increased risk of stress and negative emotions [26].

The relationship between youth pre-pandemic mental health/behavioral problems and their neural substrates, and their combined (mediating or moderating) impacts on mental health, emotional responses and stress during the pandemic are incompletely understood. A prior study in young adults showed that the strength of functional brain connections prior to the pandemic predicted pandemic-related anxiety [27]. Another study identified psychological resilience as a mediator of the association between functional connectivity of the Default Mode Network (DMN) and the inferior temporal gyrus and traumatic stress during the pandemic [28]. Our recent study in over 2,600 adolescents from the ABCD cohort reported predictive relationships between the pre-pandemic organization of the underdeveloped salience network, and the prefrontal cortex, and stress and negative affect during the pandemic [26]. However, this work did not examine the effects of pre-pandemic mental health and behavioral problems on developing brain circuits, which may have, in turn, have increased the risk for poor outcomes during the pandemic. Collectively, prior studies have examined pre-pandemic mental health problems and brain characteristics separately, but none has examined their relationships, interactions, and combined impacts on youth stress and affective responses to the pandemic’s stressors.

This study aimed to address this important gap in knowledge, and answer two questions: a) did preexisting mental health issues affect youth stress, emotions and overall mental health during the pandemic, and b) what role did the brain (which was likely impacted by these preexisting mental health issues) play in these relationships. Therefore, the study investigated direct but also indirect (mediating, moderating and interacting) relationships between preexisting mental health issues, brain characteristics, and pandemic outcomes, including overall mental health, stress, and emotional responses during the first 15 months following the outbreak. It specially examined measures of functional circuit organization (inter-regional connection patterns and their properties), and morphometric properties (including cortical thickness and cortical/subcortical volume) of individual brain regions. In 2,641 adolescents from the ABCD cohort with neuroimaging data and assessments of common mental health and behavioral problems, and longitudinal survey data collected during the pandemic, the study tested two hypotheses: a) pre-pandemic anxiety, depression and/or behavioral problems (attention problems, aggressive, internalizing and/or externalizing behaviors, preference for solitude, and/or social withdrawal) predicted worse mental health, higher stress and negative emotions during the pandemic, directly and through the mediating effects of brain circuits adversely modulated by these problems; b) the organization of brain circuits and structural characteristics of their constituent regions moderated the relationships between pre-pandemic mental health issues and pandemic responses. Finally, it also hypothesized that these relationships were positively affected by protective factors, specifically parental engagement and youth spirituality/religiosity. Prior studies have shown that these factors had positive effects on youth mental health during the pandemic [25,29].

2. Materials and methods

This study was approved by the Institutional Review Board at Boston Children’s Hospital. All analyzed data were from the ABCD release 4.0 and are publicly available through the National Institute of Mental Health Data Archive (NDA)

2.1. Participants

Youth from the 2-year follow-up ABCD cohort were studied (n = 2,641; median age at fMRI scan = 12.0 years, interquartile range (IQR) = 1.1 years). The sample included adolescents with common mental health problems, such as depressive symptoms and anxiety, but excluded those with psychiatric and neurodevelopmental disorders (including schizophrenia, psychotic disorders, Autism Spectrum Disorders, Attention-Deficit/Hyperactivity Disorder), all of which adversely modulate developing brain structures and circuits in disorder-specific ways and deserve dedicated studies. In addition, participants with any clinical MRI findings were also excluded. The analytic sample consisted of two overlapping cohorts to facilitate sensitivity and replication analyses: a) n = 1,414 (cohort A) with fMRI scans collected ≤9 months prior to any pandemic-related survey. This cutoff was used to minimize potential developmental brain changes from scan to survey (median scan-to-survey time was 3–7 months). In this cohort, 460 participants had fMRI scans collected during the pandemic; b) n = 2,174 (cohort B) with fMRI data collected prior to March 11, 2020, when the World Health Organization declared COVID-19 a pandemic [30] (scan-to-survey time was 10–22 months). Appropriate adjustments were included in statistical analyses to account for pandemic-related effects on the brain (in cohort A) and pubertal changes (in cohort B).

Youth responses from seven pandemic surveys were analyzed. Not all youth had data across surveys. Thus, seven partially overlapping sub-cohorts were studied with sample sizes from n = 802 (May 2020) to n = 218 (May 2021) in cohort A, and n = 1451 (May 2020) to n = 1135 (May 2021) in cohort B. Youth demographic characteristics in cohorts A and B are summarized in Table 1. Boys and girls were approximately equally distributed in the sample. Less than 50% were racio-ethnic minorities and about20% were Hispanic. Median (IQR) BMI was 19.3 (5.4) kg/m² and 19.1 5.2) kg/m² respectively. Over 30% were in mid-puberty, about 20% in early puberty and slightly over 20% in late puberty.

[Figure omitted. See PDF.]

2.2. Rapid Response Research (RRR) surveys

Surveys were sent electronically to all eligible youth and caregivers in May, June, August, October, December 2020, and March and May 2021. Youth responses to questions on overall mental health, stress and pandemic-related stress (separate questions) were analyzed following standardization. These questions asked: 1) “How do you think your mental health (emotional well-being) has changed in the past week compared to normal?” [1 = much worse to 5 = much better]; 2) “COVID-19 presents a lot of uncertainty about the future. In the past 7 days, including today, how stressful have you found this uncertainty to be?” [1 = not at all/very slightly to 5 = extremely]. An overall stress measure was estimated as a standardized sum of responses from the 4-item Perceived Stress Scale (PSS). Two of these responses–feeling confident in their ability to handle personal problems and feeling things were going their way–were reverse-coded before summing, for interpretability. Further, positive affect was calculated as the standardized sum of items from the NIH Toolbox Positive Affect Survey, and negative affect as the standardized sum of items from the NIH Toolbox Emotion Battery (Sadness Survey) [31]. Responses to additional questions on feeling sad, alone, lonely, angry, or scared were also analyzed on a scale from 1 = never to 5 = almost always. Not all questions were asked across surveys, but four of seven included the question on mental health, and all included the PSS measure of stress [26]. Positive and negative affect were analyzed as t-scores, and all others as z-scores. Response summary statistics are provided in supplemental S1 Table in S1 File.

2.3. Mental health variables

Pre-pandemic anxiety, depression, aggressive behaviors, attention problems, internalizing and externalizing behaviors, preference for solitude, and social withdrawal were analyzed. Anxiety, depression, and internalizing and externalizing behaviors were extracted from the Child Behavior Checklist (CBCL) as t-scores. Scores in the range 65–69 are on the clinical borderline and scores ≥70 reflect clinically significant behavioral problems. Preference for solitude and social withdrawal (also extracted from the CBCL) were measured on a Likert scale from 0 = not true to 2 = very/often true. Prevalence of these problems in the analyzed cohorts and related summary statistics are provided in S2 Table in S1 File.

2.4. Other variables

Youth spirituality/religiosity was measured by the response to a question from the Parent Demographics Survey (PDEM) that asked ‘How important are your child’s religious and spiritual beliefs in his/her daily life?’, on a scale from 1 = not at all to 4 = very much. Supplemental S16 Table in S1 File provides the distribution of responses in the sample of n = 2,641. Parental engagement was measured as the standardized mean of responses to four items from the COVID-19 RRR surveys. These questions assessed parental knowledge of their child’s whereabouts, youth-parent communication, and interaction (parents knowing their child’s plans for the day and eating dinner with them). Responses were coded on a Likert scale from 1 = never to 5 = always or almost always.

2.5. Neuroimaging data analysis

2.5.1. fMRI preprocessing.

Participants in the ABCD study are measured at 21 sites across the US. Neuroimaging data are acquired using 3T Siemens, GE, or Philips scanners. The data are minimally preprocessed before being released [32]. Resting-state fMRI (rs-fMRI) data were analyzed (fMRI acquisition: 2.4 mm isotropic; TR = 0.8 s) following further processing by the Next Generation Neural Data Analysis platform (NGNDA) [33], to improve signal-to-noise ratio, suppress artifacts, and harmonize the data across scanners. The NGNDA platform integrates custom codes for fMRI analysis with well-established tools and packages for image preprocessing. In this study, SPM12 [34,35] was used for structural segmentation, initial fMRI frame removal, coregistration of each participant’s fMRI to their structural MRI (T1w), slice-time correction, and normalization to MNI152 space. Custom codes were then used to filter voxel-level fMRI time series in the frequency range 0.01 to 0.25 Hz, and spatially downsample them to parcel-level signals. Three atlases were used for this purpose, the Schaefer cortical atlas (1000 parcels) [36,37], Melbourne subcortical atlas (54 parcels) [38], and a probabilistic atlas of the cerebellum (34 parcels) [39]. Parcel-level time series were further processed to suppress residual artifacts [40]. Given differences between MRI scanners, all signal amplitudes were normalized, so that they can be comparable across brains. The highest-quality 5-min run (out of a maximum of 4) was analyzed. This run had the lowest median connectivity (assuming that the brain at rest is overall weakly correlated), and usually the lowest number of frames censored for motion (in this sample median (IQR) percent of censored frames was 1.0 (4.0)%).

2.5.2. Estimation of resting-state topological properties.

Topological properties of the entire brain connectome, individual networks (including large resting-state networks [36] and the reward [41], prefrontal cortical, social [42], and central executive networks [43]), smaller circuits (fronto-thalamic, fronto-amygdala, fronto-basal ganglia, and fronto-striatal circuits), and individual regions (network nodes) were estimated. For this purpose, connectivity matrices (estimated using the pairwise peak cross-correlation as a measure of connection strength between pairs of fMRI signals) were first thresholded to eliminate spurious connections. A conservative, cohort-wide threshold equal to the bootstrapped moderate outlier (median + 1.5*IQR) of peak cross-correlation values was used. Resulting weighted or binary adjacency matrices (depending on the requirements of each topological property algorithm) were then used to estimate network properties across spatial scales. These properties included median connectivity (within- and across-network), global (network) and regional clustering, modularity (i.e., community structure), segregation (measured as the ratio of connections within vs outside communities), topological stability (using the largest eigenvalue of the adjacency matrix as a proxy for it), fragility, and robustness, and were estimated using algorithms in NGNDA platform. The latter incorporates tools from the The Brain Connectivity Toolbox [44], and custom implementations of algorithms for estimating topological network robustness, fragility and stability [45,46] All analyses involving brain parameters were adjusted for time of day at which participants were scanned, since timing of fMRI data collection may impact network topologies [47], and percent of fMRI frames censored for motion (relative to scan length).

2.5.3. Structural MRI analysis.

Cortical thickness, cortical and subcortical volume, and white matter intensity were estimated from the T1w MRI. These measures are provided by the ABCD at the regional level following brain segmentation using cortical and subcortical atlases, resulting in 98 regions [48].

2.6. Statistical analysis

2.6.1. Mediation model.

The model shown in the diagram in supplemental S1 Fig in S1 File was used to test the hypothesis that pre-pandemic mental health problems directly and indirectly (through their effects on the pre-pandemic organization of brain circuits–the mediator) impacted youth responses during the pandemic. Across paths, linear mixed effect models were developed to account for potential ABCD site differences (through the inclusion of a random intercept and slope for site at which participants were scanned), given variable state responses and measures to contain the spread of the SARS-CoV-2 virus. All parameter p-values were corrected for the False Discovery Rate (FDR) [49]. This correction was applied across topological properties or the connectome, individual networks or anatomical region, and for each regional topological property, across nodes within a particular network. Mediation analyses were conducted using well-established approaches, and the Sobel test was used to assess the significance of the mediation [50].

Depending on the path assessed by the model, the dependent or independent variables were either the outcomes from the COVID-19 surveys, individual topological or morphometric brain properties, or pre-pandemic mental health measures. To account for sampling differences between sites, all analyses were adjusted using propensity scores provided by the ABCD. Statistical models also included sex, age, pubertal stage, race-ethnicity, family income, BMI, pandemic onset-to-survey time (months), and scan-to-survey time. Models that included brain parameters were also adjusted for time of scan [47] and percent of frames censored for motion. The primary reported results were consistent across cohorts. Additional cohort-specific findings are reported in Supplemental Materials in S1 File.

2.6.2. Moderated multiple regression (MMR) models.

MMR models [51] that included interaction terms between individual pre-pandemic mental health issues and individual brain characteristics were also developed to investigate moderation of the relationship between pre-pandemic mental health and pandemic responses by the organization of brain circuits and the morphometric characteristics of their constituent structures. The sign and significance of the interaction term were then examined to assess moderation.

2.6.3. Secondary analyses.

Additional mixed-effects models were used to determine whether protective factors moderated the relationship between pre-pandemic mental health and responses during the pandemic. Parental engagement and youth spirituality were examined as moderators. Finally, models were developed with adjustments for youth responses to COVID survey questions in previous surveys to evaluate the pandemic’s impact over time. Specifically, each model was adjusted in two ways: 1) for the most recent prior survey for an outcome (for example when evaluating associations for stress in survey 3, the response on stress in survey 2 was included in the model), and 2) for the average across all prior surveys for an outcome to capture cumulative effects.

3. Results

3.1. Associations between pre-pandemic mental health and pandemic responses

About 24.0% of youth had preexisting anxiety, and 7.1% had depression. Pre-pandemic mental health and behavior t-scores were, on average, in the range of 41–51. Only a small fraction of youth had scores ≥65 (~2–5.5%). Less than 20% preferred solitude, at least some times, and ~6% were socially withdrawn. Summary statistics for pre-pandemic mental health and behavioral issues are summarized in supplemental S2 Table in S1 File.

Across both cohorts A and B, several direct associations between existing mental health and behavioral problems and responses to the pandemic stressors were identified. Pre-pandemic anxiety, depression, internalizing, externalizing, and aggressive behaviors were associated with more frequent feelings of anger and loneliness at the May 2020 assessment, i.e., two months following the outbreak (β = 0.09 to 0.18, CI = [0.02, 0.26], p < 0.03). In addition, anxiety, depression, internalizing and/or externalizing behaviors were associated with more frequent sadness, higher stress, and negative affect (β = 0.09 to 0.20, CI = [0.019, 0.30], p < 0.03). Higher pre-pandemic anxiety, depression and internalizing behavior scores were associated with feeling scared in June 2020 (β = 0.13 to 0.21, CI = [0.06, 0.31], p < 0.01), while lower scores and less frequent social withdrawal and preference for solitude were associated with higher positive affect (β = −0.21 to −0.09, CI = [−0.30, −0.02], p < 0.02). Pre-pandemic mental health issues were also positively associated with feeling alone/lonely, angry, and sad in August 2020 (β = 0.09 to 0.21, CI = [0.01, 0.28], p < 0.05), and aggressive behaviors were also associated with higher stress (β = 0.13 to 0.14 in the two cohorts, CI = [0.02, 0.23], p = 0.04). Anxiety and depression were negatively associated with positive affect (β = −0.21 to −0.11, CI = [−0.31, −0.05], p < 0.01), and positively associated with stress in October 2020 (β = 0.12 to 0.14, CI = [0.03, 0.25], p < 0.04). Finally, in May 2021, i.e., ~ 15 months following the outbreak, pre-pandemic aggressive behaviors were positively associated with sadness (β = 0.12 to 0.20 in the two cohorts, CI = [0.06, 0.32], p < 0.01). Detailed model statistics for results that were consistent across cohorts are provided in Table 2. Positive associations between stress in the first ~7 months of the pandemic (at multiple assessments: May, August, October 2020) and pre-pandemic anxiety and aggressive behaviors are shown in Fig 1a. Negative associations between positive affect during this period (two assessments: June and October 2020) and pre-pandemic anxiety are shown in Fig 1b. Cohort-specific model statistics are summarized in supplemental S3 and S4 Tables in S1 File.

[Figure omitted. See PDF.]

[Figure omitted. See PDF.]

Stress and positive affect were adjusted for demographic and other participant characteristics, as well as variability across study sites.

Additional associations were estimated between responses at different time points during the first 15 months of the pandemic. Positive affect ~3 months following the outbreak (June 2020 assessment) was associated with positive affect ~4 months later (October 2020; β = 0.49 to 0.55, CI = [0.43, 0.65], p < 0.01) in models that were adjusted for pre-pandemic mental health and behavioral issues. Higher negative affect ~2 months following the outbreak (May 2020 assessment) was associated with higher negative affect ~3 months later (August 2020, β = 0.68, CI = [0.63, 0.73], p < 0.01), and similarly for feeling alone (β = 0.58, CI = [0.50, 0.61], p < 0.01]). Feeling lonely in the first few months of the pandemic (May and August 2020) was also associated with feeling lonely in December 2020 (β = 0.51, CI = [0.46, 0.57], p < 0.01).

3.2. Associations between pre-pandemic mental health and brain characteristics

Brain circuit characteristics: In the context of the mediation analysis, associations between pre-pandemic mental health and pre-pandemic brain circuitry were examined in the same samples as those with survey data at each of the seven assessments. In samples from both cohorts A and B, higher anxiety was associated with higher connectivity between the right limbic network and the rest of the brain (β = 0.08 to 0.11, CI = [0.03, 0.19], p < 0.03). Higher attention problems and/or aggressive behavior scores were associated with higher fragility of the left frontoparietal control network (β = 0.08 to 0.13, CI = [0.03, 0.21], p < 0.04), and the right central executive network (β = 0.09 to 0.10, CI = 0.03, 0.17, p < 0.04). Preference for solitude was associated with higher modularity and fragility of the right somatomotor network (β = 0.08 to 0.10, CI = [0.01, 0.18], p < 0.05) and higher segregation of the right temporoparietal network (β = 0.11 to 0.12, CI = [0.04, 0.20], p < 0.05). Model statistics for survey-specific samples are provided in Table 3. Cohort-specific associations between pre-pandemic mental health and topological properties at the connectome and individual network scales are provided in supplemental S5-S8 Tables in S1 File. At the regional (node) levels, there were no significant associations in either cohort.

[Figure omitted. See PDF.]

Brain morphometric characteristics: Higher anxiety was associated with lower thickness of the right temporal gyrus and lower cerebral white matter volume (β = −0.11 to −0.05, CI = [−0.19, −0.001], p < 0.05). Higher depression was associated with higher thickness of the bilateral parahippocampal gyrus and greater white matter intensity in the left temporal pole and right superior frontal gyrus (β = 0.04 to 0.10, CI = [0.01, 0.17], p < 0.05) and lower volume in the right precentral gyrus (β = −0.12 to −0.08, CI = [−0.20, −0.03], p < 0.04). Attention problems were associated with lower volume of the bilateral rostral anterior cingulate cortex, bilateral cerebellum white matter, right caudal middle frontal gyrus, left pallidum, and left rostral middle frontal and precentral gyri (β = −0.17 to −0.06, CI = [−0.27, −0.002], p < 0.05). Preference for solitude was associated with lower volume of the right nucleus accumbens (β = −0.11 to −0.06, CI = [−0.18, −0.006], p ≤ 0.03) and higher volume of the right hippocampus (β = 0.06 to 0.10, CI = [0.01, 0.18], p < 0.04), while social withdrawal was associated with lower volume of the right paracentral lobule (β = −0.11 to −0.09, CI = [−0.18, −0.03], p < 0.02). Internalizing behaviors were associated with lower volume of the right lateral orbitofrontal cortex and bilateral cerebral white matter (β = −0.14 to −0.05, CI = [−0.23, −0.002], p ≤ 0.04) and higher white matter intensity in the right superior temporal gyrus (β = 0.04 to 0.06, CI = [0.01, 0.10], p ≤ 0.03). Externalizing behaviors were associated with higher thickness of the right pars triangularis and higher white matter intensity in the right temporal pole and right transverse temporal gyrus (β = 0.04 to 0.11, CI = [0.01, 0.17], p < 0.05) and lower volume of the bilateral cerebellum and cerebral white matter (β = −0.14 to −0.05, CI = [−0.27, −0.003], p < 0.04). Finally, aggressive behaviors were associated with lower white matter volume of the left cerebellum, bilateral cerebral white matter, left hippocampus, and bilateral superior temporal gyrus (β = −0.11 to −0.06, CI = [−0.17, −0.003], p < 0.04). Model statistics for survey-specific samples are provided in Table 4.

[Figure omitted. See PDF.]

3.3. Mediation by the brain

No statistically significant mediation of relationships between pre-pandemic mental health and pandemic responses by the brain’s topological organization were identified. Significant mediation of these relationships by brain morphology was estimated only in individual cohorts and primarily in the first assessment, ~ 2 months following the outbreak. In cohort A, the volume of the left insula partially mediated the relationship between pre-pandemic depression and feeling alone in May 2020, the volume of the left superior temporal gyrus partially mediated the relationships between aggressive behaviors and feeling angry, and similarly for externalizing behaviors and anger at the same assessment. White matter volume of the right cerebellum also mediated the relationship between externalizing behaviors and anger in May 2020 (β = 0.10 to 0.14, CI = [0.01, 0.24], p ≤ 0.03). In cohort B, the volume of the right lateral orbitofrontal cortex and left hippocampus partially mediated the relationship between pre-pandemic aggressive behaviors and feeling lonely and angry, respectively, in May 2020 (β = 0.08, CI = [0.01, 0.15], p < 0.04). Model statistics for all mediation results are provided in supplemental S9–S10 Tables in S1 File.

3.4. Moderation by the brain

The topological organization of multiple brain networks moderated the relationship between pre-pandemic mental health issues and pandemic responses, primarily in the first ~3 months following the outbreak. In either cohort A or B, negative associations between pre-pandemic anxiety and depression and overall mental health in May 2020 were attenuated by stronger connections within the left somatomotor network and between this network and the rest of the brain (β = −0.14 to −0.11, CI = [−0.21, 00.04],

p < 0.02). In addition, positive associations between pre-pandemic depression and social withdrawal and stress in May 2020 were attenuated by the community organization (modularity) of the left somatomotor network (β = −0.08, CI = [−0.14, −0.03], p < 0.02). In contrast, strength of connections within the hippocampal network amplified the positive association between pre-pandemic depression and feeling alone in May 2020 (β = 0.11, CI = [0.04, 0.17], p = 0.02). Similar interactions were estimated in June and August 2020. Specifically, modularity of distributed networks, including the prefrontal cortex, frontoparietal control network, and circuits connecting frontal regions to the thalamus, basal ganglia, and/or amygdala attenuated positive relationships between pre-pandemic mental health and behavioral issues and pandemic-related stress in June 2020. In addition, modularity of the right default and central executive networks, and left prefrontal cortex and its connections with the amygdala and the basal ganglia, attenuated positive relationships between pre-pandemic mental health and behavioral problems and feeling scared at the same assessment (β = −0.11 to −0.06, CI = [−0.17, −0.01], p < 0.05). In contrast, stronger connections within the left temporoparietal network and left hippocampus, amplified positive associations between pre-pandemic attention problems and pandemic-related stress, and depression and feeling scared, respectively (β = 0.08 to 0.10, CI = [0.02, 0.17], p < 0.03). Finally, modularity of the left somatomotor and left temporoparietal networks attenuated positive associations between pre-pandemic aggressive behaviors, attention problems, and/or social withdrawal and feeling sad and alone in August 2020 (β = −0.12 to −0.07, CI = [−0.17, −0.01], p < 0.04). Detailed model statistics are provided in supplemental S11 Table in S1 File.

Relationships between pre-pandemic mental health and behavioral issues and pandemic responses, primarily during the first ~5−6 months of the pandemic, were also moderated by structural brain characteristics. Higher volume of the right nucleus accumbens, right frontal pole, superior parietal lobule and the hippocampus attenuated relationships between pandemic stress, feeling angry, or feeling alone and pre-pandemic depression, anxiety, and/or internalizing behaviors (β = −0.13 to −0.05, CI = [−0.24, −0.01], p < 0.05). Also, lower thickness of the right anterior cingulate cortex attenuated the relationship between pre-pandemic internalizing behaviors and pandemic-related stress (β = −0.17, CI = [−0.27, −0.07], p = 0.04). Model statistics for cohorts A and B are summarized in supplemental S12 and S13 Tables in S1 File, respectively.

3.5. Additional associations

Higher parental engagement was associated with higher positive affect (β = 0.08 to 0.40, CI = [0.03, 0.50], p < 0.01) and lower negative affect, perceived stress, and feelings of sadness, loneliness, anger, and fear (β = −0.32 to −0.10, CI = [−0.40, −0.05], p < 0.01). Parental engagement also moderated relationships between pre-pandemic mental health issues and negative emotions and stress in the first few months following the outbreak, although this moderation was not consistent across cohorts. Specifically, it attenuated associations between pre-pandemic anxiety and feeling alone, and pre-pandemic aggressive behaviors and feeling angry, as well as negative affect in the first ~2 months of the pandemic (β = −0.11 to −0.10, CI = [−0.18, −0.03], p < 0.05). Model statistics are provided in supplemental S14-S15 Tables in S1 File. In addition, for over 50% of youth, religious beliefs were at least somewhat important in their daily lives (supplemental S16 Table in S1 File). Across cohorts, religiosity/spirituality was associated with lower negative affect, less frequent feelings of anger, loneliness, and sadness, and with lower stress, primarily in the first ~two months of the pandemic, and to a lesser extent ~5−7 months following the outbreak (β = −0.20 to −0.06, CI = [−0.28, −0.01], p < 0.03). It was also associated with higher positive affect at the October 2020 assessment (β = 0.11 to 0.12 CI = [0.01, 0.22], p < 0.03). Model statistics are summarized in supplemental S17 Table in S1 File. Youth from higher-income families reported lower positive affect (β = −0.157 to −0.090, CI = [−0.28, −0.01], p < 0.03) and higher negative affect, including higher sadness, loneliness, anger, and fear (β = 0.08 to 0.22, CI = [0.01, 0.37], p < 0.049). Girls were more likely to report lower positive affect (β = −2.735 to −0.190, CI = [−4.415, −0.035], p < 0.016) and higher negative affect, as well as higher perceived and pandemic-related stress and higher sadness, loneliness, anger, and fear (β = 0.075 to 4.724, CI = [0.006, 6.117], p < 0.036).

4. Discussion

In over 2,600 adolescents, we have investigated direct and indirect impacts of preexisting (pre-pandemic) mental health and behavioral problems on their stress, emotions, and overall mental health during the first 15 months of the COVID-19 outbreak. At multiple assessments during this period, especially in the first ~7 months of the outbreak, pre-pandemic mental health and behavior problems predicted negative emotions and stress, and (to a lesser extent) lower positive affect. Across most assessments, girls reported higher stress, negative affect and emotions, and lower positive affect, in agreement with prior reports of worse mental health and higher stress in girls during the COVID-19 outbreak [20]. Our findings are also in agreement with prior reports of preexisting mental health and behavioral problems as risk factors that amplified the adverse effects of the pandemic on adolescent mental and emotional health [52].

Youth with preexisting mental health and behavioral problems also had differences in the organization and structural characteristics of brain circuits and regions compared to those who did not have such problems. Specifically, they had stronger connections between the limbic network and the rest of the brain, and more topologically fragile frontoparietal control, somatomotor, and central executive networks. Prior studies have linked aberrant connection strength within and between these networks (and their topological patterns) with mental health problems, including depression and anxiety [53,54]. Topological fragility of these networks may increase vulnerability to environmental stressors, and risk of neural miswiring. Our prior work in the ABCD cohort has shown that lower pre-pandemic topological robustness (higher fragility) of some of these networks were predictive of negative emotions and higher stress during the pandemic [26].

Preexisting mental health problems were also associated with morphometric alterations, especially lower volume of cortical and subcortical regions, including the anterior cingulate cortex, frontal gyrus, orbitofrontal cortex, nucleus accumbens, cerebellum, and overall cerebral white matter. Prior studies have reported associations between lower volume of the anterior cingulate (a brain hub that is widely involved in cognitive function) and mental health problems in children and adolescents [55], and lower volume of the frontal gyrus (also widely involved in cognitive function, including cognitive control and emotion regulation) and other frontal regions with a similar range of problems, including depression, anxiety, and aggressive behaviors [56–59]. Furthermore, alterations in white matter volume of the cerebellum (which has extensive connections with cortical regions that undergo profound rewiring in adolescence) have been associated with widespread cognitive and mental health issues [60,61]. Of note is that a number of structural differences in youth with preexisting mental health issues were identified in developing regions that support social function. Social isolation during the pandemic had a profound impact on youth social function [62], which was likely amplified in youth with an already vulnerable social brain. Together, these findings suggest extensive links between pre-pandemic mental health issues and their topological and morphometric neural substrates, especially in regions supporting cognitive control, emotional regulation and social function, which together increased youth vulnerability to pandemic stressors, leading to higher stress and negative emotions and poor mental health.

To gain a better understanding of complex interactions between the brain’s pre-pandemic structural and circuit characteristics, preexisting mental health/behavioral issues, and youth responses during the pandemic, we also examined the brain as a potential mediator of relationships between preexisting mental health and youth pandemic responses. We hypothesized that these preexisting issues impacted youth responses during the pandemic also indirectly, through their neuromodulatory effects. Only morphometric brain characteristics mediated relationships, and only in the early months of the pandemic. Specifically, the volume of the insula, superior temporal gyrus, cerebellum, and orbitofrontal cortex partially mediated associations between pre-pandemic aggressive and/or externalizing behaviors and feelings of anger, and similarly for pre-pandemic depression and feeling lonely in first two months following the outbreak. Together, these structures play important roles in emotional processing and regulation but also social function [42,63–65]. Our findings provide mechanistic insights into the brain’s role in how youth responded to the pandemic, especially regions that are central to social function and emotional health, and were already impacted by preexisting mental health issues.

We also examined factors that may positively or negatively moderate the relationships between preexisting mental health issues and stress, emotional and mental health during the pandemic. First, we examined brain characteristics as potential moderators. The organization of multiple networks (including the prefrontal cortex and its subcortical connections, the frontoparietal control, and the somatomotor network), especially their community structure and connection strength, attenuated negative associations between preexisting mental health and stress, being scared, sad, and/or lonely during the first 2–5 months of the pandemic. Community structure (modularity) is a fundamental topological property of the adult connectome, is associated with its resilience, and increases as a function of neural maturation [66]. Thus, more modular and/or strongly connected brain (and thus more resilient) networks prior to the pandemic may have mitigated the negative impacts of preexisting mental health issues that increased vulnerability to the pandemic stressors. These findings are in agreement with prior work showing that more resilient circuits measured prior to the pandemic were protective for mental and emotional health and stress during the first 15 months of the outbreak [26]. In contrast, stronger connections of the left temporoparietal network and the hippocampus were risk factors for emotional health and stress during the pandemic, as they amplified relationships between preexisting behavioral issues and stress and negative emotions (especially feeling scared) in the first few months of the pandemic. Prior research has identified positive associations between connectivity in the hippocampus and stress [67], and positive associations between connectivity of the temporoparietal network and emotional responses including fear [68]. In addition, a recent study in adolescents and young adults found that stronger connections between the hippocampus and frontal regions, measured prior to the pandemic, were associated with higher stress during the pandemic [69]. Our findings are thus aligned with prior work, and further highlight the critical role of the brain’s organization prior to the pandemic as a protective or risk factor for youth responses to the pandemic’s stressors. Finally, parental engagement and the importance of religious beliefs in the youth’s life were also examined as potential protective factors during the pandemic. Although both were associated with lower stress and negative emotions in the first ~6–7 months of the pandemic, only parental engagement positively moderated negative associations between pre-pandemic mental mental health and behavioral issues and negative emotions during the pandemic, and only in the first few months following the outbreak. These findings are aligned with prior studies that have shown that parent involvement and/or religious affiliation were protective factors for adolescent outcomes during the pandemic [25,29], especially in the first few months of the pandemic, during which social isolation, uncertainty and profound changes in everyday life led to heightened stress and negative emotions.

4.1. Strengths and limitations

The study has a number of strengths, including a large sample size of over 2,600 adolescents, longitudinal surveys assessing youth mental health, stress and emotions during the pandemic, and comprehensive investigation of topological and morphometric brain characteristics, and their complex associations with preexisting mental health and behavior issues, and outcomes during the pandemic. In contrast with previous studies that have focused on direct effects of preexisting mental health issues and youth outcomes during the pandemic, this investigation has allowed us to examine both direct predictive relationships between such preexisting issues and youth responses during the pandemic, but also indirect impacts of these issues through their neuromodulatory effects on underlying neural circuits and structures. It has also allowed us to examine the role of the brain in these predictive relationships, and identify circuits, structures and their characteristics that mediated these relationships and others that moderated the negative effects of preexisting mental health issues on youth pandemic responses. Thus, this investigation has provided mechanistic insights into underlying factors driving (at least partly) the vulnerability of youth with preexisting mental health issues during the pandemic.

Despite these strengths, the study also had some limitations. As a retrospective investigation, this study relied on the data provided by the ABCD, and was limited to measures selected by its investigators. In addition, there are potentially many other risk and protective factors that were not measured or investigated, and may have interacted with preexisting mental health problems. The analyses were adjusted for sociodemographic and other individual factors, but there were likely unmeasured, pandemic-related environmental factors (for example factors in the youth’s immediate social environment, such as family dynamics) that may have affected youth outcomes during the outbreak, and could not be accounted for. Despite these limitations, this study makes a significant contribution towards our incomplete understanding of neural, mental health, and environmental/experiential factors that were either protective or increased risk for mental health issues during the COVID-19 pandemic

5. Conclusions

Preexisting anxiety, depression, and/or behavioral problems increased adolescent vulnerability to the COVID-19 pandemic’s stressors, leading to worse mental health, higher stress and negative emotions, directly and through their adverse effects on emotion- and stress-regulatory brain circuits and their constituent structures, especially in girls. Brain structural characteristics, adversely modulated by preexisting mental health problems, were additional risk factors for negative youth outcomes during the panemic. In contrast, the organization of some brain circuits was protective and attenuated the effects of preexisting mental health issues on youth responses to the pandemic’s stressors. The study identified adolescents who were especially vulnerable during the pandemic (both in terms of their mental health and brain characteristics), and should be followed longitudinally to assess their longer-term (post-pandemic) mental health outcomes. It also identified specific neural circuits and structures that could be targeted by interventions to improve these outcomes.

Supporting information

S1 File. Supplemental Materials.

https://doi.org/10.1371/journal.pone.0334028.s001

(DOCX)

References

1. 1. Spear LP. The adolescent brain and age-related behavioral manifestations. Neurosci Biobehav Rev. 2000;24(4):417–63. pmid:10817843

* View Article

* PubMed/NCBI

* Google Scholar

2. 2. Richter LM. Studying adolescence. Science. 2006;312(5782):1902–5. pmid:16809526

* View Article

* PubMed/NCBI

* Google Scholar

3. 3. Hazen E, Schlozman S, Beresin E. Adolescent psychological development: a review. Pediatr Rev. 2008;29(5):161–7; quiz 168. pmid:18450837

* View Article

* PubMed/NCBI

* Google Scholar

4. 4. Hollenstein T, Lougheed JP. Beyond storm and stress: typicality, transactions, timing, and temperament to account for adolescent change. Am Psychol. 2013;68(6):444–54. pmid:23915399

* View Article

* PubMed/NCBI

* Google Scholar

5. 5. Susman EJ, Rogol A. Puberty and psychological development. Handbook of Adolescent Psychology. 2013. pp. 15–44. doi: https://doi.org/10.1002/9780471726746.ch2

6. 6. Steinberg L. Cognitive and affective development in adolescence. Trends Cogn Sci. 2005;9(2):69–74. doi: https://doi.org/10.1016/j.tics.2004.12.005 pmid:15668099

* View Article

* PubMed/NCBI

* Google Scholar

7. 7. Paus T, Keshavan M, Giedd JN. Why do many psychiatric disorders emerge during adolescence? Nat Rev Neurosci. 2008;9(12):947–57. pmid:19002191

* View Article

* PubMed/NCBI

* Google Scholar

8. 8. Sawyer SM, Afifi RA, Bearinger LH, Blakemore S-J, Dick B, Ezeh AC, et al. Adolescence: a foundation for future health. Lancet. 2012;379(9826):1630–40.

* View Article

* Google Scholar

9. 9. Dahl RE. Adolescent brain development: a period of vulnerabilities and opportunities. Keynote address. Ann N Y Acad Sci. 2004;1021:1–22. pmid:15251869

* View Article

* PubMed/NCBI

* Google Scholar

10. 10. Casey BJ, Getz S, Galvan A. The adolescent brain. Dev Rev. 2008;28(1):62–77. pmid:18688292

* View Article

* PubMed/NCBI

* Google Scholar

11. 11. Arain M, Haque M, Johal L, Mathur P, Nel W, Rais A, et al. Maturation of the adolescent brain. Neuropsychiatr Dis Treat. 2013;9:449–61. pmid:23579318

* View Article

* PubMed/NCBI

* Google Scholar

12. 12. Fuhrmann D, Knoll LJ, Blakemore S-J. Adolescence as a sensitive period of brain development. Trends Cogn Sci. 2015;19(10):558–66. pmid:26419496

* View Article

* PubMed/NCBI

* Google Scholar

13. 13. Lamblin M, Murawski C, Whittle S, Fornito A. Social connectedness, mental health and the adolescent brain. Neurosci Biobehav Rev. 2017;80:57–68. pmid:28506925

* View Article

* PubMed/NCBI

* Google Scholar

14. 14. Merikangas KR, He J-P, Burstein M, Swanson SA, Avenevoli S, Cui L, et al. Lifetime prevalence of mental disorders in U.S. adolescents: results from the National Comorbidity Survey Replication--Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychiatry. 2010;49(10):980–9. pmid:20855043

* View Article

* PubMed/NCBI

* Google Scholar

15. 15. de Girolamo G, Dagani J, Purcell R, Cocchi A, McGorry PD. Age of onset of mental disorders and use of mental health services: needs, opportunities and obstacles. Epidemiol Psychiatr Sci. 2011;21(1):47–57.

* View Article

* Google Scholar

16. 16. Blakemore S-J. Adolescence and mental health. Lancet. 2019;393(10185):2030–1. pmid:31106741

* View Article

* PubMed/NCBI

* Google Scholar

17. 17. de Figueiredo CS, Sandre PC, Portugal LCL, Mázala-de-Oliveira T, da Silva Chagas L, Raony Í, et al. COVID-19 pandemic impact on children and adolescents’ mental health: biological, environmental, and social factors. Progr Neuro-Psychopharmacol Biol Psychiatry. 2021;106:110171.

* View Article

* Google Scholar

18. 18. Panchal U, Salazar de Pablo G, Franco M, Moreno C, Parellada M, Arango C, et al. The impact of COVID-19 lockdown on child and adolescent mental health: systematic review. Eur Child Adolesc Psychiatry. 2021;32(7):1151–77.

* View Article

* Google Scholar

19. 19. Stinson EA, Sullivan RM, Peteet BJ, Tapert SF, Baker FC, Breslin FJ, et al. Longitudinal impact of childhood adversity on early adolescent mental health during the COVID-19 pandemic in the ABCD study cohort: does race or ethnicity moderate findings? Biol Psychiatry Glob Open Sci. 2021;1(4):324–35. pmid:34608463

* View Article

* PubMed/NCBI

* Google Scholar

20. 20. Magson NR, Freeman JYA, Rapee RM, Richardson CE, Oar EL, Fardouly J. Risk and protective factors for prospective changes in adolescent mental health during the COVID-19 pandemic. J Youth Adolesc. 2021;50(1):44–57. pmid:33108542

* View Article

* PubMed/NCBI

* Google Scholar

21. 21. Liu CH, Stevens C, Conrad RC, Hahm HC. Evidence for elevated psychiatric distress, poor sleep, and quality of life concerns during the COVID-19 pandemic among U.S. young adults with suspected and reported psychiatric diagnoses. Psychiatry Res. 2020;292:113345. pmid:32745794

* View Article

* PubMed/NCBI

* Google Scholar

22. 22. Hards E, Loades ME, Higson-Sweeney N, Shafran R, Serafimova T, Brigden A, et al. Loneliness and mental health in children and adolescents with pre-existing mental health problems: a rapid systematic review. Br J Clin Psychol. 2022;61(2):313–34. pmid:34529837

* View Article

* PubMed/NCBI

* Google Scholar

23. 23. Ng CSM, Ng SSL. Impact of the COVID-19 pandemic on children’s mental health: a systematic review. Front Psychiatry. 2022;13:975936. pmid:36329921

* View Article

* PubMed/NCBI

* Google Scholar

24. 24. Casey BJ, Cannonier T, Conley MI, Cohen AO, Barch DM, Heitzeg MM, et al. The Adolescent Brain Cognitive Development (ABCD) study: imaging acquisition across 21 sites. Dev Cogn Neurosci. 2018;32:43–54. pmid:29567376

* View Article

* PubMed/NCBI

* Google Scholar

25. 25. Hamatani S, Hiraoka D, Makita K, Tomoda A, Mizuno Y. Longitudinal impact of COVID-19 pandemic on mental health of children in the ABCD study cohort. Sci Rep. 2022;12(1):19601. pmid:36379997

* View Article

* PubMed/NCBI

* Google Scholar

26. 26. Hu L, Stamoulis C. Strength and resilience of developing brain circuits predict adolescent emotional and stress responses during the COVID-19 pandemic. Cereb Cortex. 2024;34(4):bhae164. pmid:38669008

* View Article

* PubMed/NCBI

* Google Scholar

27. 27. He L, Wei D, Yang F, Zhang J, Cheng W, Feng J, et al. Functional connectome prediction of anxiety related to the COVID-19 pandemic. Am J Psychiatry. 2021;178(6):530–40.

* View Article

* Google Scholar

28. 28. Liu X, Zhao Y, Suo X, Zhang X, Pan N, Kemp GJ, et al. Psychological resilience mediates the protective role of default-mode network functional connectivity against COVID-19 vicarious traumatization. Transl Psychiatry. 2023;13(1):231. pmid:37380702

* View Article

* PubMed/NCBI

* Google Scholar

29. 29. Dyer WJ, Crandall A, Hanson CL. COVID-19 stress, religious affiliation, and mental health outcomes among adolescents. J Adolesc Health. 2023;72(6):892–8. pmid:36809866

* View Article

* PubMed/NCBI

* Google Scholar

30. 30. WHO Director-General’s Opening Remarks at the Media Briefing on COVID-19 - 11 March 2020. World Health Organization. 2020. https://www.who.int/director-general/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-covid-19---11-march-2020

31. 31. Salsman JM, Butt Z, Pilkonis PA, Cyranowski JM, Zill N, Hendrie HC, et al. Emotion assessment using the NIH Toolbox. Neurology. 2013;80(11 Suppl 3):S76–86. pmid:23479549

* View Article

* PubMed/NCBI

* Google Scholar

32. 32. Hagler DJ Jr, Hatton S, Cornejo MD, Makowski C, Fair DA, Dick AS, et al. Image processing and analysis methods for the Adolescent Brain Cognitive Development Study. Neuroimage. 2019;202:116091. pmid:31415884

* View Article

* PubMed/NCBI

* Google Scholar

33. 33. Next-Generation Neural Data Analysis (NGNDA). https://github.com/cstamoulis1/Next-Generation-Neural-Data-Analysis-NGNDA-

34. 34. Friston KJ, Holmes AP, Worsley KJ, Poline JP, Frith CD, Frackowiak RSJ. Statistical parametric maps in functional imaging: a general linear approach. Hum Brain Mapping. 1994;2(4):189–210.

* View Article

* Google Scholar

35. 35. Friston KJ, Ashburner JT, Kiebel SJ. Statistical parametric mapping: The analysis of functional brain images. Academic Press; 2007.

36. 36. Yeo BTT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106(3):1125–65. pmid:21653723

* View Article

* PubMed/NCBI

* Google Scholar

37. 37. Schaefer A, Kong R, Gordon EM, Laumann TO, Zuo X-N, Holmes AJ, et al. Local-global parcellation of the human cerebral cortex from intrinsic functional connectivity MRI. Cereb Cortex. 2018;28(9):3095–114. pmid:28981612

* View Article

* PubMed/NCBI

* Google Scholar

38. 38. Tian Y, Margulies DS, Breakspear M, Zalesky A. Topographic organization of the human subcortex unveiled with functional connectivity gradients. Nat Neurosci. 2020;23(11):1421–32. pmid:32989295

* View Article

* PubMed/NCBI

* Google Scholar

39. 39. Diedrichsen J, Balsters JH, Flavell J, Cussans E, Ramnani N. A probabilistic MR atlas of the human cerebellum. Neuroimage. 2009;46(1):39–46. pmid:19457380

* View Article

* PubMed/NCBI

* Google Scholar

40. 40. Brooks SJ, Parks SM, Stamoulis C. Widespread positive direct and indirect effects of regular physical activity on the developing functional connectome in early adolescence. Cereb Cortex. 2021;31(10):4840–52. pmid:33987673

* View Article

* PubMed/NCBI

* Google Scholar

41. 41. Haber SN, Knutson B. The reward circuit: linking primate anatomy and human imaging. Neuropsychopharmacology. 2010;35(1):4–26. pmid:19812543

* View Article

* PubMed/NCBI

* Google Scholar

42. 42. Blakemore S-J. The social brain in adolescence. Nat Rev Neurosci. 2008;9(4):267–77. pmid:18354399

* View Article

* PubMed/NCBI

* Google Scholar

43. 43. Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci. 2002;3(3):201–15. pmid:11994752

* View Article

* PubMed/NCBI

* Google Scholar

44. 44. Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and interpretations. Neuroimage. 2010;52(3):1059–69. pmid:19819337

* View Article

* PubMed/NCBI

* Google Scholar

45. 45. Pasqualetti F, Zhao S, Favaretto C, Zampieri S. Fragility limits performance in complex networks. Sci Rep. 2020;10(1):1774. pmid:32019963

* View Article

* PubMed/NCBI

* Google Scholar

46. 46. Wu J, Barahona M, Tan Y-J, Deng H-Z. Spectral measure of structural robustness in complex networks. IEEE Trans Syst, Man, Cybern A. 2011;41(6):1244–52.

* View Article

* Google Scholar

47. 47. Hu L, Katz ES, Stamoulis C. Modulatory effects of fMRI acquisition time of day, week and year on adolescent functional connectomes across spatial scales: implications for inference. NeuroImage. 2023;284:120459.

* View Article

* Google Scholar

48. 48. Hagler DJ Jr, Hatton SeanN, Cornejo MD, Makowski C, Fair DA, Dick AS, et al. Image processing and analysis methods for the adolescent brain cognitive development study. NeuroImage. 2019;202:116091.

* View Article

* Google Scholar

49. 49. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B: Stat Methodol. 1995;57(1):289–300.

* View Article

* Google Scholar

50. 50. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173–82. pmid:3806354

* View Article

* PubMed/NCBI

* Google Scholar

51. 51. McClelland GH, Irwin JR, Disatnik D, Sivan L. Multicollinearity is a red herring in the search for moderator variables: a guide to interpreting moderated multiple regression models and a critique of Iacobucci, Schneider, Popovich, and Bakamitsos (2016). Behav Res Methods. 2017;49(1):394–402. pmid:27531361

* View Article

* PubMed/NCBI

* Google Scholar

52. 52. Bera L, Souchon M, Ladsous A, Colin V, Lopez-Castroman J. Emotional and Behavioral Impact of the COVID-19 Epidemic in Adolescents. Curr Psychiatry Rep. 2022;24(1):37–46. pmid:35102536

* View Article

* PubMed/NCBI

* Google Scholar

53. 53. Li B-J, Friston K, Mody M, Wang H-N, Lu H-B, Hu D-W. A brain network model for depression: from symptom understanding to disease intervention. CNS Neurosci Ther. 2018;24(11):1004–19. pmid:29931740

* View Article

* PubMed/NCBI

* Google Scholar

54. 54. Martin EI, Ressler KJ, Binder E, Nemeroff CB. The neurobiology of anxiety disorders: brain imaging, genetics, and psychoneuroendocrinology. Psychiatr Clin North Am. 2009;32(3):549–75. pmid:19716990

* View Article

* PubMed/NCBI

* Google Scholar

55. 55. Boes AD, McCormick LM, Coryell WH, Nopoulos P. Rostral anterior cingulate cortex volume correlates with depressed mood in normal healthy children. Biol Psychiatry. 2008;63(4):391–7. pmid:17916329

* View Article

* PubMed/NCBI

* Google Scholar

56. 56. Shad MU, Muddasani S, Rao U. Gray matter differences between healthy and depressed adolescents: a voxel-based morphometry study. J Child Adolesc Psychopharmacol. 2012;22(3):190–7. pmid:22537357

* View Article

* PubMed/NCBI

* Google Scholar

57. 57. Bashford‐Largo J, Zhang R, Mathur A, Elowsky J, Schwartz A, Dobbertin M, et al. Reduced cortical volume of the default mode network in adolescents with generalized anxiety disorder. Depression Anxiety. 2022;39(6):485–95.

* View Article

* Google Scholar

58. 58. Straub J, Brown R, Malejko K, Bonenberger M, Grön G, Plener PL, et al. Adolescent depression and brain development: evidence from voxel-based morphometry. J Psychiatry Neurosci. 2019;44(4):237–45.

* View Article

* Google Scholar

59. 59. Hostetler N, Tavares TP, Ritchie MB, Oliver LD, Chen VV, Greening S, et al. Prefrontal cortex structural and developmental associations with callous-unemotional traits and aggression. Sci Rep. 2024;14(1):4087. pmid:38374428

* View Article

* PubMed/NCBI

* Google Scholar

60. 60. Carta I, Chen CH, Schott AL, Dorizan S, Khodakhah K. Cerebellar modulation of the reward circuitry and social behavior. Science. 2019;363(6424):eaav0581. pmid:30655412

* View Article

* PubMed/NCBI

* Google Scholar

61. 61. Van Overwalle F, Manto M, Cattaneo Z, Clausi S, Ferrari C, Gabrieli JDE, et al. Consensus paper: cerebellum and social cognition. Cerebellum. 2020;19(6):833–68. pmid:32632709

* View Article

* PubMed/NCBI

* Google Scholar

62. 62. Breaux R, Cash AR, Lewis J, Garcia KM, Dvorsky MR, Becker SP. Impacts of COVID-19 quarantine and isolation on adolescent social functioning. Curr Opin Psychol. 2023;52:101613. pmid:37364468

* View Article

* PubMed/NCBI

* Google Scholar

63. 63. Uddin LQ, Nomi JS, Hébert-Seropian B, Ghaziri J, Boucher O. Structure and function of the human insula. J Clin Neurophysiol. 2017;34(4):300–6.

* View Article

* Google Scholar

64. 64. Rolls ET. The functions of the orbitofrontal cortex. Brain Cogn. 2004;55(1):11–29. pmid:15134840

* View Article

* PubMed/NCBI

* Google Scholar

65. 65. Gu X, Hof PR, Friston KJ, Fan J. Anterior insular cortex and emotional awareness. J Comp Neurol. 2013;521(15):3371–88. pmid:23749500

* View Article

* PubMed/NCBI

* Google Scholar

66. 66. Chen M, Deem MW. Development of modularity in the neural activity of children’s brains. Phys Biol. 2015;12(1):016009. pmid:25619207

* View Article

* PubMed/NCBI

* Google Scholar

67. 67. Goldfarb EV, Rosenberg MD, Seo D, Constable RT, Sinha R. Hippocampal seed connectome-based modeling predicts the feeling of stress. Nat Commun. 2020;11(1):2650. pmid:32461583

* View Article

* PubMed/NCBI

* Google Scholar

68. 68. Shin LM, Liberzon I. The neurocircuitry of fear, stress, and anxiety disorders. Neuropsychopharmacology. 2010;35(1):169–91. pmid:19625997

* View Article

* PubMed/NCBI

* Google Scholar

69. 69. Perica MI, Ravindranath O, Calabro FJ, Foran W, Luna B. Hippocampal-prefrontal connectivity prior to the COVID-19 pandemic predicts stress reactivity. Biol Psychiatry Glob Open Sci. 2021;1(4):283–90. pmid:34849503

* View Article

* PubMed/NCBI

* Google Scholar

Citation: Risner M, Hu L, Stamoulis C (2025) Pre-pandemic mental health and brain characteristics predict adolescent stress and emotions during the COVID-19 pandemic. PLoS One 20(10): e0334028. https://doi.org/10.1371/journal.pone.0334028

About the Authors:

Matthew Risner

Roles: Formal analysis, Investigation, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

Affiliation: Department of Pediatrics, Division of Adolescent and Young Adult Medicine, Boston Children’s Hospital, Boston Massachusetts, United States of America

Linfeng Hu

Roles: Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

Affiliations: Department of Pediatrics, Division of Adolescent and Young Adult Medicine, Boston Children’s Hospital, Boston Massachusetts, United States of America, Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, Massachusetts, United States of America

Catherine Stamoulis

Roles: Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Writing – original draft, Writing – review & editing

E-mail: [email protected]

Affiliations: Department of Pediatrics, Division of Adolescent and Young Adult Medicine, Boston Children’s Hospital, Boston Massachusetts, United States of America, Department of Pediatrics, Harvard Medical School, Boston, Massachusetts, United States of America

ORICD: https://orcid.org/0000-0001-9814-1802

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

References

1. Spear LP. The adolescent brain and age-related behavioral manifestations. Neurosci Biobehav Rev. 2000;24(4):417–63. pmid:10817843

2. Richter LM. Studying adolescence. Science. 2006;312(5782):1902–5. pmid:16809526

3. Hazen E, Schlozman S, Beresin E. Adolescent psychological development: a review. Pediatr Rev. 2008;29(5):161–7; quiz 168. pmid:18450837

4. Hollenstein T, Lougheed JP. Beyond storm and stress: typicality, transactions, timing, and temperament to account for adolescent change. Am Psychol. 2013;68(6):444–54. pmid:23915399

5. Susman EJ, Rogol A. Puberty and psychological development. Handbook of Adolescent Psychology. 2013. pp. 15–44. doi: https://doi.org/10.1002/9780471726746.ch2

6. Steinberg L. Cognitive and affective development in adolescence. Trends Cogn Sci. 2005;9(2):69–74. doi: https://doi.org/10.1016/j.tics.2004.12.005 pmid:15668099

7. Paus T, Keshavan M, Giedd JN. Why do many psychiatric disorders emerge during adolescence? Nat Rev Neurosci. 2008;9(12):947–57. pmid:19002191

8. Sawyer SM, Afifi RA, Bearinger LH, Blakemore S-J, Dick B, Ezeh AC, et al. Adolescence: a foundation for future health. Lancet. 2012;379(9826):1630–40.

9. Dahl RE. Adolescent brain development: a period of vulnerabilities and opportunities. Keynote address. Ann N Y Acad Sci. 2004;1021:1–22. pmid:15251869

10. Casey BJ, Getz S, Galvan A. The adolescent brain. Dev Rev. 2008;28(1):62–77. pmid:18688292

11. Arain M, Haque M, Johal L, Mathur P, Nel W, Rais A, et al. Maturation of the adolescent brain. Neuropsychiatr Dis Treat. 2013;9:449–61. pmid:23579318

12. Fuhrmann D, Knoll LJ, Blakemore S-J. Adolescence as a sensitive period of brain development. Trends Cogn Sci. 2015;19(10):558–66. pmid:26419496

13. Lamblin M, Murawski C, Whittle S, Fornito A. Social connectedness, mental health and the adolescent brain. Neurosci Biobehav Rev. 2017;80:57–68. pmid:28506925

14. Merikangas KR, He J-P, Burstein M, Swanson SA, Avenevoli S, Cui L, et al. Lifetime prevalence of mental disorders in U.S. adolescents: results from the National Comorbidity Survey Replication--Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychiatry. 2010;49(10):980–9. pmid:20855043

15. de Girolamo G, Dagani J, Purcell R, Cocchi A, McGorry PD. Age of onset of mental disorders and use of mental health services: needs, opportunities and obstacles. Epidemiol Psychiatr Sci. 2011;21(1):47–57.

16. Blakemore S-J. Adolescence and mental health. Lancet. 2019;393(10185):2030–1. pmid:31106741

17. de Figueiredo CS, Sandre PC, Portugal LCL, Mázala-de-Oliveira T, da Silva Chagas L, Raony Í, et al. COVID-19 pandemic impact on children and adolescents’ mental health: biological, environmental, and social factors. Progr Neuro-Psychopharmacol Biol Psychiatry. 2021;106:110171.

18. Panchal U, Salazar de Pablo G, Franco M, Moreno C, Parellada M, Arango C, et al. The impact of COVID-19 lockdown on child and adolescent mental health: systematic review. Eur Child Adolesc Psychiatry. 2021;32(7):1151–77.

19. Stinson EA, Sullivan RM, Peteet BJ, Tapert SF, Baker FC, Breslin FJ, et al. Longitudinal impact of childhood adversity on early adolescent mental health during the COVID-19 pandemic in the ABCD study cohort: does race or ethnicity moderate findings? Biol Psychiatry Glob Open Sci. 2021;1(4):324–35. pmid:34608463

20. Magson NR, Freeman JYA, Rapee RM, Richardson CE, Oar EL, Fardouly J. Risk and protective factors for prospective changes in adolescent mental health during the COVID-19 pandemic. J Youth Adolesc. 2021;50(1):44–57. pmid:33108542

21. Liu CH, Stevens C, Conrad RC, Hahm HC. Evidence for elevated psychiatric distress, poor sleep, and quality of life concerns during the COVID-19 pandemic among U.S. young adults with suspected and reported psychiatric diagnoses. Psychiatry Res. 2020;292:113345. pmid:32745794

22. Hards E, Loades ME, Higson-Sweeney N, Shafran R, Serafimova T, Brigden A, et al. Loneliness and mental health in children and adolescents with pre-existing mental health problems: a rapid systematic review. Br J Clin Psychol. 2022;61(2):313–34. pmid:34529837

23. Ng CSM, Ng SSL. Impact of the COVID-19 pandemic on children’s mental health: a systematic review. Front Psychiatry. 2022;13:975936. pmid:36329921

24. Casey BJ, Cannonier T, Conley MI, Cohen AO, Barch DM, Heitzeg MM, et al. The Adolescent Brain Cognitive Development (ABCD) study: imaging acquisition across 21 sites. Dev Cogn Neurosci. 2018;32:43–54. pmid:29567376

25. Hamatani S, Hiraoka D, Makita K, Tomoda A, Mizuno Y. Longitudinal impact of COVID-19 pandemic on mental health of children in the ABCD study cohort. Sci Rep. 2022;12(1):19601. pmid:36379997

26. Hu L, Stamoulis C. Strength and resilience of developing brain circuits predict adolescent emotional and stress responses during the COVID-19 pandemic. Cereb Cortex. 2024;34(4):bhae164. pmid:38669008

27. He L, Wei D, Yang F, Zhang J, Cheng W, Feng J, et al. Functional connectome prediction of anxiety related to the COVID-19 pandemic. Am J Psychiatry. 2021;178(6):530–40.

28. Liu X, Zhao Y, Suo X, Zhang X, Pan N, Kemp GJ, et al. Psychological resilience mediates the protective role of default-mode network functional connectivity against COVID-19 vicarious traumatization. Transl Psychiatry. 2023;13(1):231. pmid:37380702

29. Dyer WJ, Crandall A, Hanson CL. COVID-19 stress, religious affiliation, and mental health outcomes among adolescents. J Adolesc Health. 2023;72(6):892–8. pmid:36809866

30. WHO Director-General’s Opening Remarks at the Media Briefing on COVID-19 - 11 March 2020. World Health Organization. 2020. https://www.who.int/director-general/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-covid-19---11-march-2020

31. Salsman JM, Butt Z, Pilkonis PA, Cyranowski JM, Zill N, Hendrie HC, et al. Emotion assessment using the NIH Toolbox. Neurology. 2013;80(11 Suppl 3):S76–86. pmid:23479549

32. Hagler DJ Jr, Hatton S, Cornejo MD, Makowski C, Fair DA, Dick AS, et al. Image processing and analysis methods for the Adolescent Brain Cognitive Development Study. Neuroimage. 2019;202:116091. pmid:31415884

33. Next-Generation Neural Data Analysis (NGNDA). https://github.com/cstamoulis1/Next-Generation-Neural-Data-Analysis-NGNDA-

34. Friston KJ, Holmes AP, Worsley KJ, Poline JP, Frith CD, Frackowiak RSJ. Statistical parametric maps in functional imaging: a general linear approach. Hum Brain Mapping. 1994;2(4):189–210.

35. Friston KJ, Ashburner JT, Kiebel SJ. Statistical parametric mapping: The analysis of functional brain images. Academic Press; 2007.

36. Yeo BTT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106(3):1125–65. pmid:21653723

37. Schaefer A, Kong R, Gordon EM, Laumann TO, Zuo X-N, Holmes AJ, et al. Local-global parcellation of the human cerebral cortex from intrinsic functional connectivity MRI. Cereb Cortex. 2018;28(9):3095–114. pmid:28981612

38. Tian Y, Margulies DS, Breakspear M, Zalesky A. Topographic organization of the human subcortex unveiled with functional connectivity gradients. Nat Neurosci. 2020;23(11):1421–32. pmid:32989295

39. Diedrichsen J, Balsters JH, Flavell J, Cussans E, Ramnani N. A probabilistic MR atlas of the human cerebellum. Neuroimage. 2009;46(1):39–46. pmid:19457380

40. Brooks SJ, Parks SM, Stamoulis C. Widespread positive direct and indirect effects of regular physical activity on the developing functional connectome in early adolescence. Cereb Cortex. 2021;31(10):4840–52. pmid:33987673

41. Haber SN, Knutson B. The reward circuit: linking primate anatomy and human imaging. Neuropsychopharmacology. 2010;35(1):4–26. pmid:19812543

42. Blakemore S-J. The social brain in adolescence. Nat Rev Neurosci. 2008;9(4):267–77. pmid:18354399

43. Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci. 2002;3(3):201–15. pmid:11994752

44. Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and interpretations. Neuroimage. 2010;52(3):1059–69. pmid:19819337

45. Pasqualetti F, Zhao S, Favaretto C, Zampieri S. Fragility limits performance in complex networks. Sci Rep. 2020;10(1):1774. pmid:32019963

46. Wu J, Barahona M, Tan Y-J, Deng H-Z. Spectral measure of structural robustness in complex networks. IEEE Trans Syst, Man, Cybern A. 2011;41(6):1244–52.

47. Hu L, Katz ES, Stamoulis C. Modulatory effects of fMRI acquisition time of day, week and year on adolescent functional connectomes across spatial scales: implications for inference. NeuroImage. 2023;284:120459.

48. Hagler DJ Jr, Hatton SeanN, Cornejo MD, Makowski C, Fair DA, Dick AS, et al. Image processing and analysis methods for the adolescent brain cognitive development study. NeuroImage. 2019;202:116091.

49. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B: Stat Methodol. 1995;57(1):289–300.

50. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173–82. pmid:3806354

51. McClelland GH, Irwin JR, Disatnik D, Sivan L. Multicollinearity is a red herring in the search for moderator variables: a guide to interpreting moderated multiple regression models and a critique of Iacobucci, Schneider, Popovich, and Bakamitsos (2016). Behav Res Methods. 2017;49(1):394–402. pmid:27531361

52. Bera L, Souchon M, Ladsous A, Colin V, Lopez-Castroman J. Emotional and Behavioral Impact of the COVID-19 Epidemic in Adolescents. Curr Psychiatry Rep. 2022;24(1):37–46. pmid:35102536

53. Li B-J, Friston K, Mody M, Wang H-N, Lu H-B, Hu D-W. A brain network model for depression: from symptom understanding to disease intervention. CNS Neurosci Ther. 2018;24(11):1004–19. pmid:29931740

54. Martin EI, Ressler KJ, Binder E, Nemeroff CB. The neurobiology of anxiety disorders: brain imaging, genetics, and psychoneuroendocrinology. Psychiatr Clin North Am. 2009;32(3):549–75. pmid:19716990

55. Boes AD, McCormick LM, Coryell WH, Nopoulos P. Rostral anterior cingulate cortex volume correlates with depressed mood in normal healthy children. Biol Psychiatry. 2008;63(4):391–7. pmid:17916329

56. Shad MU, Muddasani S, Rao U. Gray matter differences between healthy and depressed adolescents: a voxel-based morphometry study. J Child Adolesc Psychopharmacol. 2012;22(3):190–7. pmid:22537357

57. Bashford‐Largo J, Zhang R, Mathur A, Elowsky J, Schwartz A, Dobbertin M, et al. Reduced cortical volume of the default mode network in adolescents with generalized anxiety disorder. Depression Anxiety. 2022;39(6):485–95.

58. Straub J, Brown R, Malejko K, Bonenberger M, Grön G, Plener PL, et al. Adolescent depression and brain development: evidence from voxel-based morphometry. J Psychiatry Neurosci. 2019;44(4):237–45.

59. Hostetler N, Tavares TP, Ritchie MB, Oliver LD, Chen VV, Greening S, et al. Prefrontal cortex structural and developmental associations with callous-unemotional traits and aggression. Sci Rep. 2024;14(1):4087. pmid:38374428

60. Carta I, Chen CH, Schott AL, Dorizan S, Khodakhah K. Cerebellar modulation of the reward circuitry and social behavior. Science. 2019;363(6424):eaav0581. pmid:30655412

61. Van Overwalle F, Manto M, Cattaneo Z, Clausi S, Ferrari C, Gabrieli JDE, et al. Consensus paper: cerebellum and social cognition. Cerebellum. 2020;19(6):833–68. pmid:32632709

62. Breaux R, Cash AR, Lewis J, Garcia KM, Dvorsky MR, Becker SP. Impacts of COVID-19 quarantine and isolation on adolescent social functioning. Curr Opin Psychol. 2023;52:101613. pmid:37364468

63. Uddin LQ, Nomi JS, Hébert-Seropian B, Ghaziri J, Boucher O. Structure and function of the human insula. J Clin Neurophysiol. 2017;34(4):300–6.

64. Rolls ET. The functions of the orbitofrontal cortex. Brain Cogn. 2004;55(1):11–29. pmid:15134840

65. Gu X, Hof PR, Friston KJ, Fan J. Anterior insular cortex and emotional awareness. J Comp Neurol. 2013;521(15):3371–88. pmid:23749500

66. Chen M, Deem MW. Development of modularity in the neural activity of children’s brains. Phys Biol. 2015;12(1):016009. pmid:25619207

67. Goldfarb EV, Rosenberg MD, Seo D, Constable RT, Sinha R. Hippocampal seed connectome-based modeling predicts the feeling of stress. Nat Commun. 2020;11(1):2650. pmid:32461583

68. Shin LM, Liberzon I. The neurocircuitry of fear, stress, and anxiety disorders. Neuropsychopharmacology. 2010;35(1):169–91. pmid:19625997

69. Perica MI, Ravindranath O, Calabro FJ, Foran W, Luna B. Hippocampal-prefrontal connectivity prior to the COVID-19 pandemic predicts stress reactivity. Biol Psychiatry Glob Open Sci. 2021;1(4):283–90. pmid:34849503