A hierarchical stochastic model for bistable perception.

Area of Science:

• Cognitive Neuroscience
• Computational Psychiatry
• Perception Research

Background:

• Bistable perception involves alternating between interpretations of ambiguous stimuli.
• Response patterns in bistable perception show variability and differences between schizophrenia patients and healthy controls.
• Existing models for bistable perception are complex and difficult to parameterize.

Purpose of the Study:

• To develop a parsimonious stochastic model for analyzing bistable perception response patterns.
• To link empirical data analysis with detailed models of underlying neuronal processes.
• To investigate group differences in bistable perception between schizophrenia patients and healthy controls.

Main Methods:

• Utilized a Hidden Markov Model (HMM) to analyze the timing of percept changes under continuous and intermittent stimulus presentation.
• Developed a hierarchical Brownian model (HBM) to simulate response patterns and relate them to neuronal mechanisms.
• The HBM models neuronal activity as Brownian motion with drift, reflecting competing neuronal populations.

Main Results:

• The HMM captured observed differences in response patterns between schizophrenia patients and healthy controls.
• The HBM successfully reproduced similar response patterns and group differences with a minimal parameter set.
• The HBM provides a mechanistic explanation for bistable perception and group differences observed in schizophrenia.

Conclusions:

• The proposed hierarchical Brownian model (HBM) offers a parsimonious and mechanistically relevant approach to studying bistable perception.
• The HBM effectively models response patterns and highlights potential underlying neuronal mechanisms contributing to group differences.
• This model facilitates a bridge between empirical observations of bistable perception and theoretical neuronal models.