AbstractModels of hallucinations across disorders emphasise an imbalance between sensory input and top-down influences over perception. However, the psychological and mechanistic correlates of this imbalance remain underspecified. Visual hallucinations in Parkinson’s disease (PD) are associated with impairments in lower level visual processes and attention, accompanied by over activity and connectivity in higher-order association brain networks. PD therefore provides an attractive framework to explore the relative contributions of bottom-up versus top-down disturbances in hallucinations. Here, we characterised sensory processing in PD patients with and without visual hallucinations, and in healthy controls, by fitting a hierarchical drift diffusion model (hDDM) to an attentional task. The hDDM uses Bayesian estimates to decompose reaction time and response output into parameters reflecting drift rates of evidence accumulation, decision thresholds and non-decision time. We observed slower drift rates in PD patients with hallucinations, which were insensitive to changes in task demand. In contrast, wider decision boundaries and shorter non-decision times relative to controls were found in PD regardless of hallucinator status. Inefficient and less flexible sensory evidence accumulation emerge as unique features of PD hallucinators. We integrate these results with current models of hallucinations, suggesting that slow and inefficient sensory input in PD is less informative, and may therefore be down-weighted leading to an over reliance on top-down influences. Our findings provide a novel computational framework to better specify the impairments in dynamic sensory processing that are a risk factor for visual hallucinations.
Visual Hallucinations Are Characterized by Impaired Sensory Evidence Accumulation: Insights From Hierarchical Drift Diffusion Modeling in Parkinson’s Disease