A first realization of reinforcement learning-based closed-loop EEG-TMS

Transcranial magnetic stimulation (TMS) is a powerful tool to investigate neurophysiology of the human brain and treat brain disorders. Traditionally, therapeutic TMS has been applied in a one-size-fits-all approach, disregarding inter- and intraindividual differences. Brain state-dependent EEG-TMS, such as coupling TMS with a pre-specified phase of the sensorimotor μ-rhythm, enables the induction of differential neuroplastic effects depending on the targeted phase. But this approach is still user-dependent as it requires defining an a-priori target phase. Here, we present a first realization of a machine-learning-based, closed-loop real-time EEG-TMS setup to identify user-independently the individual μ-rhythm phase associated with high- vs. low-corticospinal excitability states. We applied EEG-TMS to 25 participants targeting the supplementary motor area- primary motor cortex network and used a reinforcement learning algorithm to identify the μ- rhythm phase associated with high- vs. low corticospinal excitability. We employed linear mixed effects models and Bayesian analysis to assess effects of reinforced learning on corticospinal excitability indexed by motor evoked potential amplitude, and functional connectivity indexed by the imaginary part of resting-state EEG coherence. Our results show that reinforcement learning identified the μ-rhythm phase associated with high- vs. low-excitability states, and their repetitive stimulation resulted in long-term increases vs. decreases in functional connectivity in the stimulated sensorimotor network. Our work demonstrated the feasibility of a reinforcement learning-based adaptive closed-loop EEG-TMS approach in a full experimental cohort, enabling user-independent identification of μ-rhythm phases associated with corticospinal excitability states.