Childhood trauma, task-based functional connectivity and working memory performance in schizophrenia: the moderating role of diagnosis

Introduction: Schizophrenia is associated with aberrations in resting-state and task-based functional connectivity, particularly in the frontoparietal network, which is critical for higher-order cognitive functions such as working memory [1]. Evidence, primarily from resting-state fMRI studies, suggests that childhood trauma exposure may induce changes in brain function in clinical and non-clinical populations [2]. Behavioural findings further indicate that childhood trauma, particularly physical neglect, is linked to cognitive impairments [3]. It remains unclear how childhood trauma affects frontoparietal connectivity during active working memory processes, and whether such alterations contribute to working memory deficits in schizophrenia. We hypothesized that physical neglect would be associated with poorer working memory performance, and that this would be partially mediated by frontoparietal connectivity in a diagnosis-independent manner. Methods: Childhood trauma history, working memory performance, and task-based fMRI data were collected from 38 individuals with schizophrenia or schizoaffective disorder and 128 healthy controls. fMRI data were preprocessed and denoised using a modular preprocessing pipeline in the CONN toolbox [4]. Seed-to-voxel connectivity analyses were conducted using four predefined frontoparietal regions (right and left dorsolateral prefrontal cortices, right and left posterior parietal cortices) as seeds, and all other voxels brain-wide as targets. Connectivity patterns during the working memory task were contrasted across diagnostic groups to identify group differences in connectivity. Additionally, high versus low working memory load conditions were contrasted across the whole sample to isolate connectivity patterns associated with higher cognitive load. To examine the interactions between physical neglect, frontoparietal connectivity, and working memory performance, we conducted a moderated mediation analysis in RStudio using PROCESS Macro Model 59, with diagnosis as moderator. Results: Patients showed widespread patterns of both hyper- and hypo-connectivity within the frontoparietal network compared to controls. Across all participants, physical neglect was associated with reduced working memory performance (r = -0.2595, p = 0.0019). However, this association was not mediated by frontoparietal connectivity. Under high working memory load, there was suppression of medial frontal cortex (MedFC) activity by the left dorsolateral prefrontal cortex (dlPFC) (T(1, 165) = -5.4, p < 0.001). Diagnosis moderated both the association between physical neglect and left dlPFC–MedFC connectivity (R2 change = 2.98%, p = 0.0223), and the association between connectivity and working memory performance (R2 change = 2.75%, p = 0.0112). Although the relationship between physical neglect and left dlPFC-MedFC connectivity was not significant in either group, greater suppression of MedFC by left dlPFC was associated with better working memory performance in patients only (effect = -10.8334, p = 0.0054). Conclusion: Contrary to our hypothesis, frontoparietal connectivity did not mediate the relationship between physical neglect and impaired working memory performance. We provide evidence that suppression of the default mode network – namely, the MedFC – by the left dlPFC under high working memory load predicted better working memory performance in patients only. These findings support the view of schizophrenia as a disorder of functional dysconnectivity, and suggest that neural mechanisms supporting cognitive performance may differ by diagnosis. Understanding these mechanisms may inform future interventions targeting cognitive deficits in schizophrenia.