Characterizing early auditory deficits in schizophrenia using drift-diffusion modeling

Abstract

Introduction: While tone-matching disturbances have been extensively documented in schizophrenia, it remains unclear if impaired tone-matching performance is solely related to early sensory dysfunction or if the ability to appropriately translate early levels of processing into behavioral responses contributes to the deficit. Using an evidence accumulation model to analyze tone-matching performance, we aim to deconstruct perceptual decision-making into its core computational components, providing a more precise characterization of tone-matching deficits in schizophrenia.

Methods: Thirty (30) individuals diagnosed with schizophrenia and 30 matched healthy controls performed a task requiring discrimination between pairs of pure tones. Behavioral performance was evaluated using Bayesian mixed-effects models of accuracy and reaction times, and further decomposed via a Drift Diffusion Model (DDM) to quantify underlying decision-making parameters.

Results: Individuals with schizophrenia exhibit significantly lower accuracy and prolonged reaction times relative to controls, with a diminished impact of pitch difference levels on performance. DDM analysis revealed that individuals with schizophrenia accumulate sensory evidence at a slower rate (lower drift rate) and demonstrate increased non-decision time, reflecting delays in sensory encoding and/or motor execution. No significant differences were observed in starting point bias or boundary separation.

Conclusions: These findings suggest that tone-matching impairments in schizophrenia are primarily attributable to deficits in evidence accumulation and the translation of sensory information into behavioral responses. This study provides novel insights into the cognitive pathophysiology of schizophrenia and underscores the utility of computational modeling to elucidate the mechanisms underlying perceptual and decisional deficits in this population.

Keywords: Auditory processing; Computational modeling; Drift-diffusion model; Response time; Schizophrenia; Tone-matching.