Pre-COVID brain functional connectome features prospectively predict emergence of distress symptoms after onset of the COVID-19 pandemic

Abstract

Background Persistent psychological distress associated with the coronavirus disease 2019 (COVID-19) pandemic has been well documented. This study aimed to identify pre-COVID brain functional connectome that predicts pandemic-related distress symptoms among young adults. Methods Baseline neuroimaging studies and assessment of general distress using the Depression, Anxiety and Stress Scale were performed with 100 healthy individuals prior to wide recognition of the health risks associated with the emergence of COVID-19. They were recontacted for the Impact of Event Scale-Revised and the Posttraumatic Stress Disorder Checklist in the period of community-level outbreaks, and for follow-up distress evaluation again 1 year later. We employed the network-based statistic approach to identify connectome that predicted the increase of distress based on 136-region-parcellation with assigned network membership. Predictive performance of connectome features and causal relations were examined by cross-validation and mediation analyses. Results The connectome features that predicted emergence of distress after COVID contained 70 neural connections. Most within-network connections were located in the default mode network (DMN), and affective network-DMN and dorsal attention network-DMN links largely constituted between-network pairs. The hippocampus emerged as the most critical hub region. Predictive models of the connectome remained robust in cross-validation. Mediation analyses demonstrated that COVID-related posttraumatic stress partially explained the correlation of connectome to the development of general distress. Conclusions Brain functional connectome may fingerprint individuals with vulnerability to psychological distress associated with the COVID pandemic. Individuals with brain neuromarkers may benefit from the corresponding interventions to reduce the risk or severity of distress related to fear of COVID-related challenges.

Aberrant brain network topology in youth with a familial risk for bipolar disorder: a task‐based fMRI connectome study

“…Further, research has linked DMN alterations to depressive rumination (Zhou et al., 2020) and suggests its protective role in mitigating negative emotions (Akiki, Averill, & Abdallah, 2017; Chin Fatt et al., 2020; Qin et al., 2022) as well as enhancing resilience among unaffected siblings of individuals with BD (Doucet, Bassett, Yao, Glahn, & Frangou, 2017). The disruption of intrinsic coherence within the DMN may contribute to the heterogeneous forms of emotional arousal associated with risk for BD (Lindquist, Wager, Kober, Bliss‐Moreau, & Barrett, 2012; Pan et al., 2022; Suo et al., 2022), and increased connectivity in the DMN may also predict depression treatment response (Singh et al., 2021). Intriguingly, genetic variations linked to heritable forms of BD may shed light on how DMN disorganization increases vulnerability to BD development or exacerbates the progression of depression and anxiety symptoms (Buckner & DiNicola, 2019).…”

“…We constructed the brain functional connectome using an atlas with a parcellation system in which seed regions were labeled according to a community of seven canonical brain networks (Chin Fatt et al., 2020; Pan et al., 2022; Yeo et al., 2011) that were derived from Yeo's template with 100 cortical parcels and 36 subcortical parcels of Brainnetome (for details of network distribution, see Figure S2) (Fan et al., 2016; Schaefer et al., 2018).…”

“…Briefly, NBS was implemented using the following steps: (a) establish a threshold for initial pairwise connections at p < .005 (striking a balance between conservatism and sensitivity to potentially meaningful effects) (Drysdale et al., 2017) to create a connectivity matrix with suprathreshold connections; (b) identify connected components within the suprathreshold connections using a breadth‐first search and quantify the size of the components; (c) conduct a test of nonparametric permutation (reallocate subjects into groups and record the maximal connected component) to assign a corrected p ‐value to each component based on its size (pNBS‐FDR < .05 as the network cluster threshold) (Cocchi et al., 2012; Zalesky et al., 2010). To enhance the interpretability of our network findings, pairs with identified connectivity were attributed to one of seven large‐scale brain networks to yield the network intensity, which determined the networks with the most connectivity alterations in the comparison of high‐risk BD and healthy controls (Pan et al., 2022; Yeo et al., 2011). We also performed supplementary analyses when additionally including measures of depression and anxiety symptoms as covariates to examine their effects on group differences.…”

Frequent Errors in Modeling by Machine Learning: A Prototype Case of Predicting the Timely Evolution of COVID-19 Pandemic

“…(iv) Brain features prior to the COVID-19 epidemic are useful for predicting the later emergence of distress [78]. Distress prediction was found significant by tenfold cross-validation with linear regression (brain features were employed as independent variables) [78].…”

Emotions in Times of Pandemic Crisis among Italian Children: A Systematic Review

“…Evidence of brain plasticity related to volume change patterns has been provided in healthy adults, who were not somatically affected by COVID-19 infection following the real-life external event of the global pandemic. The high rate of psychological distress (stress and anxiety) attributed to social effects of the COVID-19-related lockdown, such as social isolation and perceived uncertainty, seemed to be associated with the transient volumetric enlargement of some brain regions (amygdala, putamen, and ventral anterior temporal cortices) [ 9 ]. In addition, a recent fMRI study [ 10 ] focused on neural correlates of psychological general distress and showed that some brain features prior to the COVID-19 epidemic could predict the later emergence of distress, mainly including within and between the default mode network (DMN) connection patterns, with the left hippocampus emerging as the most critical hub region.…”