Quantifying State-Dependent Control Properties of Brain Dynamics from Perturbation Responses

Area of Science:

• Neuroscience
• Control Theory
• Systems Biology

Background:

• The brain functions as a dynamic control system, transitioning between various states like rest and motor activity.
• Network control theory offers tools to analyze brain dynamics, but prior studies often overlooked external perturbations' role in system identification.

Purpose of the Study:

• To develop a novel method for estimating the controllability Gramian matrix by integrating perturbation inputs with network control theory.
• To investigate brain state dynamics and their responses to external stimulation using this new framework.

Main Methods:

• A perturbation input paradigm was combined with network control theory.
• A novel method for estimating the controllability Gramian matrix was proposed and validated.
• The method was applied to transcranial magnetic stimulation-induced electroencephalographic (EEG) responses in motor-related and resting states.

Main Results:

• The proposed method provides insights into brain dynamics, quantifying overall controllability (eigenvalues) and specific controllable directions (eigenvectors).
• Controllable directions effectively differentiated between resting and motor-related brain states.
• Certain states, such as motor execution and motor imagery, were not distinguishable using these measures, suggesting shared intrinsic control properties.

Conclusions:

• Brain states exhibit distinct intrinsic control properties, influencing their dynamic responses to stimulation.
• The developed methodology offers a quantitative approach to assess brain state differences.
• This approach has potential applications in characterizing individual response variability and optimizing stimulation efficacy.