Hybrid Deep Learning in Binary Mental Intention Recognition and EEG-Based Brain-Computer Interfaces
Year: 2026
Type: conference-paper
Abstract: Brain-Computer Interface (BCI) systems provide critical communication channels for individuals with severe motor or speech impairments. This paper presents an affordable, real-time EEG-based BCI system for binary mental intention recognition, classifying affirmative (yes) and negative (no) responses. A hybrid deep learning framework combining a Convolutional Neural Network (CNN) for spatial feature extraction, a Recurrent Neural Network (RNN) for temporal sequence modeling, and a Support Vector Machine (SVM) for binary classification is proposed. The system was validated on EEG data collected from 10 healthy participants aged 20-26 years as a controlled proof-of-concept study. Experimental results demonstrate a mean binary classification accuracy of 95.7% ± 2.3% with cross-validation, statistically outperforming all single-model baselines$(p<0.01)$. The system operates in realtime with a mean response latency of 1.2 seconds, demonstrating feasibility as a practical assistive communication tool.
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: hybrid
OpenAlex ID: W7203753170