Personalized whole-brain activity patterns predict human corticospinal tract activation in real-time

Area of Science:

• Neuroscience
• Rehabilitation Medicine
• Biomedical Engineering

Background:

• Transcranial magnetic stimulation (TMS) shows promise for stroke-related motor impairments, but treatment effects are inconsistent.
• Brain state-dependent TMS is a potential solution, yet individual differences in stroke lesions and brain activity complicate its application.
• Personalized approaches are crucial to overcome these challenges and optimize TMS efficacy in post-stroke patients.

Purpose of the Study:

• To develop and validate a novel machine learning-based system for real-time identification of personalized brain activity patterns.
• To investigate the ability of this system to predict corticospinal tract (CST) activation states (strong and weak) in real-time.
• To assess the impact of real-time, personalized brain state-dependent TMS on motor evoked potential (MEP) amplitudes and variability.

Main Methods:

• A machine learning-based EEG-TMS system was developed to identify personalized strong and weak CST states in real-time.
• Participants underwent a single session involving TMS-EEG-EMG data acquisition, personalized classifier training, and real-time TMS guided by the classifier.
• Single-pulse TMS was delivered during classifier-predicted personalized CST states (strong, weak, or random) in healthy adults.

Main Results:

• Motor evoked potential (MEP) amplitudes were significantly larger during classifier-predicted strong CST states compared to weak and random states.
• MEP amplitude variability was significantly reduced during strong CST states relative to weak CST states.
• Personalized strong and weak CST states were identified in real-time, lasting approximately 1-2 seconds, with unique spectro-spatial EEG patterns differentiating them between individuals.

Conclusions:

• This study demonstrates for the first time that personalized whole-brain EEG patterns can predict CST activation in real-time in humans.
• These findings represent a significant advancement towards personalized, brain state-dependent TMS interventions for enhancing post-stroke motor function.
• The developed system offers a pathway to more effective and targeted neuromodulation strategies for stroke rehabilitation.