Exploratory Analysis and Modeling of EEG Spectral Characteristics and Treatment Factors in ECT Patients

Electroconvulsive therapy (ECT) is an effective treatment for individuals with treatment-resistant depression. It involves applying external electrical stimulation to the brain in patients under general anesthesia to induce ictal activity followed by postictal activity which are observable on an electroencephalogram (EEG). Many factors influence the overall therapeutic outcome of ECT. In this paper, we use patient and treatment information along with EEG data from six patients undergoing ECT to explore the impact that these factors have on patient outcomes. We observed distinct spectral differences in the EEG across four distinctly defined phases of the treatment: baseline, pre-ECT, ECT, and post-ECT. Based on these insights, we developed two regression models to predict ictal and postictal durations as outputs that are proxies for patient outcomes. These models incorporate patient and treatment features alongside EEG spectral data to estimate target outputs. Our exploratory analysis highlights ECT-related features that may influence both immediate and long-term treatment outcomes.