Quantifying brain state transition cost via Schrödinger Bridge.

Area of Science:

• Systems Neuroscience
• Computational Neuroscience
• Cognitive Neuroscience

Background:

• Quantifying brain state transition cost is crucial in systems neuroscience.
• Previous methods used deterministic models, neglecting neural system stochasticity.
• Accurate quantification requires accounting for inherent randomness in neural dynamics.

Purpose of the Study:

• To propose a novel framework for quantifying brain state transition cost in stochastic systems.
• To introduce the Schrödinger Bridge as a measure of transition cost.
• To apply this framework to neuroimaging data and assess its utility in cognitive tasks.

Main Methods:

• Developed a novel framework based on optimal control in stochastic systems.
• Quantified transition cost using Kullback-Leibler divergence (Schrödinger Bridge).
• Applied the framework to Human Connectome Project functional magnetic resonance imaging (fMRI) data.

Main Results:

• Successfully computed brain state transition cost during cognitive tasks.
• Demonstrated a correspondence between computed transition cost and task difficulty.
• Validated the framework's utility in analyzing cognitive functions.

Conclusions:

• The proposed stochastic optimal control framework offers a more accurate measure of brain state transition cost.
• Brain state transition cost is linked to cognitive effort and task demands.
• This framework provides a general theoretical tool for investigating cognitive functions through the lens of transition cost.