Reshaping computational neuropsychiatry beyond synaptopathy.

Area of Science:

• Computational Neuropsychiatry
• Neuroscience
• Psychiatry

Background:

• Computational neuropsychiatry explains psychopathology via neuronal networks and belief propagation.
• Active Inference (AI) models dysfunctional signal processing by minimizing variational free energy.
• Current AI approaches link aberrant belief updating to neuropsychiatric symptoms.

Purpose of the Study:

• To critically evaluate the applicability of Active Inference to the full spectrum of neuropsychiatric and neurological disorders.
• To address the challenge posed by the distinction between inference disorders (synaptopathies) and brain function disorders (disconnection).
• To propose an improved modeling framework for neuropsychiatric symptoms.

Main Methods:

• Review and conceptual analysis of Active Inference principles.
• Examination of the distinction between synaptopathies and disconnection disorders.
• Theoretical integration of these concepts for future modeling.

Main Results:

• Active Inference successfully models disorders of inference by focusing on neuronal message passing and belief updating.
• A clear distinction exists between disorders of inference (e.g., psychiatric) and disorders of brain function (e.g., neurological).
• This distinction highlights limitations in applying the free energy principle solely based on neuronal message passing.

Conclusions:

• Future computational models of neuropsychiatric symptoms must incorporate the distinction between inference and brain function disorders.
• Considering factors beyond neuronal message passing, such as neuromodulatory precision and structural disconnection, is crucial.
• This nuanced approach will enhance the explanatory power of computational models for a wider range of brain disorders.