Abstract: In order to interpret the world around us, the brain must first gather and process sensory information. In naturalistic settings, this sensory information is inherently multisensory. The brain seamlessly integrates information from multiple senses to ensure we detect stimuli and make decisions efficiently. Many inferences have been made about how the brain uses multisensory information during detection and decision-making processes.

However, this area of research is filled with nuance regarding not only the external conditions of the environmental cues but also the internal conditions of the participant's brain, making multisensory integration a rich area of research for mathematical modelling.

Both behavioural and electrophysiological experiments have revealed that strategies for combining multisensory cues evolve throughout early human development and become abnormal with ageing. Studies have demonstrated that maladaptive multisensory processing is a key indicator of fall-risk in older adults and in individuals with Parkinson's Disease. The models developed in this study investigate multisensory mechanisms in areas of cortical accumulation across healthy human development and ageing, and in individuals with various pathologies.

This study presents an adaptable framework of multisensory integration by extending a mean-field two-variable recurrent network model of decision-making dynamics from Wong and Wang (2006). All model configurations within the developed framework reproduce known behavioural measures from multisensory detection and decision-making tasks. Until now, research using cortical decision-making models has either been primarily unisensory or simulates dynamics at a neuronal level.

By assessing distinct model architectures and multisensory integration strategies, the findings of this study suggest the presence of co-activation in the human brain's decision-making areas during simple multimodal tasks. Furthermore, this study demonstrates how biologically motivated mathematical models can be used to probe the potential mechanisms responsible for the observed differences in unisensory and multisensory responses.

Joint work with John Butler