Regional superficial amygdala resting-state functional connectivity in adults infers childhood maltreatment severity

Qi Liu,Xinwei Song,Xinqi Zhou,Linghong Huang,Xiaodong Zhang,Lan Wang,Siyu Zhu,Chunmei Lan,Wenxu Yang,Weihua Zhao
DOI: https://doi.org/10.1093/psyrad/kkad004
2023-04-18
Psychoradiology
Abstract:Background: Childhood maltreatment (CM) is a potential risk factor for some neuropsychiatric disorders in adulthood (e.g. depression and anxiety) and alters trajectories of brain development. Accumulating evidence suggests that functional connectivity of the limbic system, especially the amygdala, is highly associated with childhood maltreatment, although not all studies have found this. These inconsistent results may be due to differential alterations of amygdala resting-state functional connectivity (rsFC) following childhood maltreatment. Objective: Our aim was to investigate the relationship between the rsFC of amygdala subregions and CM severity, as well as to develop a stable rsFC-based model for inferring the severity of CM. Methods: In this study, we employed the Childhood Trauma Questionnaire (CTQ) to assess CM severity in each individual. We explored the relationship between the rsFC of amygdala subregions (i.e. centromedial -CMA, basolateral -BLA, superficial-SFA amygdala) and CM experience in a discovery dataset of n = 110 healthy Chinese participants by linear multiple regression analysis. Subsequent dimensional and categorical approach were performed to elucidate the relationship between rsFCs and CM severity and CM subtypes, respectively. A support vector regression model was then conducted to validate the associations between rsFCs and total CTQ scores. Moreover, we also verified the model into another independent replication dataset (n = 38). Results: Our findings suggested that childhood maltreatment was negatively associated with rsFC between the right superficial amygdala and perigenual anterior cingulate cortex (pgACC)/postcentral gyrus (PCG) but not the other two amygdala subregions. Moreover, SFA-pgACC coupling was more associated with physical neglect whereas the SFA-PCG was more related to emotional neglect. In addition, supervised machine learning confirmed that using these two rsFCs as predictors could stably estimate continuous maltreatment severity in both discovery and replication datasets. Conclusion: The current study supports that the rsFCs of superficial amygdala are related to childhood maltreatment and which may be a potential biomarker for the effects of childhood maltreatment-related psychiatric disorders (i.e. depression and anxiety).
What problem does this paper attempt to address?