Precision phase targeting of event-related oscillations using real-time closed-loop TMS-EEG

Abstract

ObjectiveCurrent closed-loop TMS-EEG systems rely on phase prediction algorithms that require highly periodic signals, limiting their ability to target brief, event-related activity. We developed a real-time closed-loop (RT-CL) TMS-EEG system that directly detects oscillatory phase without prediction, enabling phase-locked stimulation within microseconds. MethodsWe validated the system against a prediction-based approach using simulated sine waves and human EEG data (N=18), without active TMS delivery. ResultsAcross frequency-modulated sweeps and spontaneous occipital alpha oscillations (eyes-open vs. closed), the RT-CL system achieved higher triggering probability (11-24%) and reduced the phase error variability (2-10{degrees}). Importantly, when targeting event-related theta oscillations during two spatial navigation tasks, RT-CL produced [~]20% higher triggering probabilities and [~]17{degrees} lower phase error variability than phase prediction. ConclusionThese findings validate the RT-CL system for probing phase-dependent mechanisms during active cognition and pathological brain states. SignificanceBy precisely targeting brief, variable neural signals, RT-CL could be used for the development of personalized TMS interventions for neuropsychiatric disorders during symptom provocation.