A Pattern Recognition Neural Network Approach for an SSVEP-Based Brain-Computer Interface

This paper proposes an algorithm for feature extraction and classification for a Steady State Visual Evoked Potentials-based Brain-Computer Interface to control a device using four commands. An energy-based dataset created using Parseval's Energy Theorem is used for the feature extraction, while a Shallow Pattern Recognition Neural Network is used in feature classification. Further, the algorithm's performance is evaluated for different numbers of channels and combinations, and 96.7% accuracy is obtained for 16 electroencephalogram channels. This study's findings can help to develop brain controlled devices and assistive technology for individuals with motor disabilities for an improved quality of life.