کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

Analysis Therapeutic Effect of repetitive Transcranial Magnetic Stimulation on Recovery of Patients with Multiple Sclerosis based on Human EEG Signals via Machine Learning

نویسندگان :

Ali Abedi¹ Mehran Jahed²

1- دانشگاه صنعتی امیرکبیر(پلی تکنیک تهران) 2- دانشگاه صنعتی شریف

کلمات کلیدی :

repetitive Transcranial Magnetic Stimulation (rTMS)،Multiple Sclerosis (MS)،Support Vector Machine (SVM)،Machine Learning،Electroencephalography (EEG)

چکیده :

repetitive Transcranial Magnetic Stimulation (rTMS) is a noninvasive treatment for Multiple Sclerosis (MS), but results are inconsistent, and protocols are seldom personalized. The current paper proposes an EEG-based machine learning scheme to investigate the effects of rTMS by classifying Real Stimulation (RS) from Sham Stimulation (SS). Resting EEG was recorded pre/post 10-week rTMS in 37 MS patients (RS=19, SS=18). Features were extracted across time, frequency, time–frequency, and connectivity domains, and LASSO selected the informative features. A SVM with leave-one-out cross-validation achieved 92.57% accuracy, 90.0% specificity, 95.0% sensitivity, and AUC = 0.96, while PLV, spectral entropy, and band power featured as the most discriminative ones.