Towards real-time EEG decoding for closed-loop, brain state-dependent TMS

Combining TMS with EEG offers a promising avenue for closed-loop, brain state-dependent stimulation. By aligning TMS with periods of increased or decreased cortical excitability, our goal is to enhance neuromodulatory effects of TMS that outlast the effects of the stimulation, potentially increasing the efficacy of therapeutic interventions. To this end, we are developing TMS protocols that can algorithmically modulate and sustain individual target brain states in real-time, ensuring stimulation is both personalized and optimally timed. Moving beyond pre-defined biomarkers, we are employing machine learning techniques to identify individual predictive brain activity patterns for immediate TMS effects on motor cortical networks. To facilitate the translation of offline decoding insights into real-time applications, we incorporate EEG data from previous sessions of the same individual as well as a broader subject pool. This approach improves the generalizability and robustness of our machine learning models and reduces the time required for classifier personalization.

Project contributors: Lisa Haxel with Oskari Ahola, Miriam Kirchhoff, David Vetter, Dania Humaidan, Andreas Jooß and in collaboration with our ERC synergy project partners at Aalto University and University of Chieti–Pescara

Funding: European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (ConnectToBrain, ERC synergy grant agreement No. 810377)