EEG Biomarkers of Accelerated Prolonged Intermittent Theta-Burst Stimulation in Depression: Machine Learning with a Minimal-Electrode Montage

Resting-state electroencephalography (rsEEG) is a promising biomarker for treatment stratification in depression because it is noninvasive, relatively low cost, repeatable, and can capture network-level physiology relevant to brain stimulation [1–3]. However, the spatial sampling needed for clinically useful prediction remains uncertain [1,4]. Although high-density EEG offers broader coverage, its longer setup time, poorer tolerability, and greater technical burden may limit scalability [4]. This issue is especially relevant for accelerated protocols, where the appeal of rapid treatment may be offset by burdensome baseline phenotyping [3,5].