Abstract
Previous studies have suggested that patients with schizophrenia and healthy controls exhibit differential activation of and connectivity involving the dorsolateral prefrontal cortex (DLPFC) during working memory tasks, though findings remain inconsistent. This study explores how working memory performance modulates these functional interactions. Using fMRI and psychophysiological interaction (PPI) analyses on 45 healthy controls and 45 patients with schizophrenia performing a 2-back task, researchers found that performance acts as a critical modulator. An interaction effect was observed between performance and diagnosis on functional connectivity between the DLPFC and posterior regions. This effect was driven by a negative correlation in healthy controls (efficiency) and a positive correlation in patients (compensation), shedding new light on the neural basis of working memory deficits in schizophrenia.
Introduction
Working memory (WM) deficits are a core feature of schizophrenia, often linked to dysfunctions in the dorsolateral prefrontal cortex (DLPFC). However, neuroimaging literature has been inconsistent, reporting both hyper- and hypo-activation of the DLPFC in patients. The "functional integration" perspective suggests these discrepancies may arise from ignoring inter-individual differences in task performance. This study investigates whether task performance modulates the functional connectivity between the DLPFC and other brain regions differently in patients versus healthy controls.
Methods
The study involved 90 participants: 45 patients with schizophrenia (SZ) and 45 healthy controls (HC). All participants performed a 2-back task during fMRI scanning.
- Performance Metrics: Participants were categorized into "high" and "low" performers based on their $d'$ scores, a measure of sensitivity in distinguishing targets from distractors.
- Connectivity Analysis: Psychophysiological interaction (PPI) analysis was employed to measure task-specific changes in functional connectivity, using the left and right DLPFC as seed regions.
- Clinical Assessment: Patient symptoms were evaluated using the Positive and Negative Syndrome Scale (PANSS).
Results
The behavioral data confirmed a significant performance gap, with patients showing lower $d'$ scores than controls ($p < 0.001$). Neuroimaging revealed a complex relationship between performance and brain connectivity.
The Performance-Connectivity Interaction
The most significant finding was an interaction effect between diagnosis and performance on the connectivity between the DLPFC and posterior regions, including the angular cortex, fusiform gyrus, and middle occipital gyrus.
- Healthy Controls (Efficiency): In the HC group, better performance was associated with decreased functional connectivity. This suggests a "neural efficiency" model where high-performing healthy brains require less communication between regions to process information.
- Schizophrenia Patients (Compensation): Conversely, in the SZ group, better performance was associated with increased functional connectivity. This indicates a compensatory mechanism where patients must recruit additional neural resources and increase network "chatter" to maintain task performance.
Discussion
These findings support the "cortical inefficiency" model of schizophrenia. While healthy individuals optimize their neural pathways for efficiency, patients with schizophrenia appear to rely on increased connectivity to overcome underlying processing deficits. When this compensatory capacity is exceeded, performance declines.
This "cross-over" effect explains why previous studies—which often grouped patients regardless of performance—produced conflicting results. By accounting for individual performance levels, researchers can better map the functional architecture of the schizophrenic brain.
Limitations and Future Directions
The study's reliance on a single task load (2-back) and the potential influence of antipsychotic medication are noted limitations. Future research should employ varying levels of task difficulty to identify the specific "breaking point" where compensatory connectivity fails in patients and where efficiency transitions to compensation in healthy controls.
Summary updated