Motor imagery EEG classification via wavelet-packet synthetic augmentation and entropy-based channel selection

Introduction Motor-imagery (MI) brain–computer interfaces often suffer from limited EEG datasets and redundant channels, hampering both accuracy and clinical usability. We address these bottlenecks by presenting a unified framework that simultaneously boosts classification performance, reduces th…