Abstract
Discrete visual perceptual disturbances (VisDis) are often neglected in clinical examination of schizophrenia patients though being reported by 50% of patients. Moreover, the relevance of VisDis in patients presenting with recent-onset psychosis (ROP) or a clinical high-risk state for psychosis (CHR) is unclear. Deeper insights into visual dysfunction in early psychosis states could not only hold beneficial information for clinical practice but foster understanding of basic perceptual disease mechanisms mediating susceptibility to psychosis. Here, we systematically investigated the relationship between VisDis and (a) core clinical measures across early phase psychiatric conditions, and (b) intrinsic brain systems using a novel machine learning approach. VisDis, clinical measures, and resting-state fMRI were assessed in ROP (N=147), CHR (N=143) and, for comparison reasons, in recent onset depression (ROD, N=151) and healthy controls (HC, N=280). Our machine-learning approach used pairwise functional connectivity within occipital (ON) and frontoparietal (FPN) networks implicated in visual information processing to predict VisDis. VisDis were reported more frequently in ROP (50.3%) and CHR (55.9%) than in ROD (16.6%) and HC (4.3%), (p<0.0001). Higher VisDis burden was associated with worse psycho-functional remission in both CHR and ROP, and lower quality of life (Qol) and higher depressiveness in CHR specifically (all p<0.05). Functional connectivity within ON predicted VisDis in ROP (balanced accuracy (BAC) 60.2%, p=0.0001) and CHR (BAC 67.4 %, p=0.029). Additionally, VisDis were predicted by functional connectivity within FPN in the combined ROP-CHR sample (BAC 61.1%, p=0.006). These large-sample study findings imply that VisDis are highly relevant in ROP and, especially, CHR being related to measures of psycho-functional outcome and emotional states. Results from the multivariate machine learning approach suggest that functional intrinsic activity within ON and FPN is associated with the VisDis phenomenon. This strongly supports a model of the visual system being implicated in core disease mechanisms of psychosis.