A Bayesian approach to computational neuropsychology

Bayesian theories have profoundly impacted neuroscience by advancing our formal understanding of neurological disorders. In this talk, I will present a Bayesian framework to neuropsychology, utilising the rigour of generative modelling to explore syndromes such as speech disorders. In the first part of the talk, I will elucidate the construction of these models under the paradigm of active inference, detailing their ability to encapsulate complex environmental dynamics. I will show that these models can support context-sensitive planning and different inference and learning processes, offering computational insights into systemic resilience against perturbations. In the latter half, I will illustrate the application of these models in dissecting the mechanisms underlying neuropsychological dysfunction. This includes a computational perspective on the interaction between brain lesions and their symptomatic expressions. Here, I will introduce the concept of functional degeneracy, and show how it can be used to assess degenerate architectures and potential recovery mechanisms in both intact and impaired brains. I will conclude with a discussion of how these models could facilitate personalised interventions post-neurological damage, thereby paving the way for targeted therapeutic strategies in the future.