Spiking Neural Network based Brain-Computer-Interface
Spatio- and spectro-temporal data(SSTD), such as EEG, fMRI, and MEG readings of a human brain, is one of the most prominent types of data observed in daily basis. Their high dimensionality, inherently large and complex structure, and low signal-to-noise ratio pose great challenges in developing efficient spatio-temporal pattern recognition (STPR) mechanisms.
Clusters of spiking neurons combined with spike-timing-dependent plasticity(STDP) learning rules, which closely adheres to a brain-like information processing scheme, are proven to address the posed challenges in a fast, flexible, and adaptive fashion. Our research focus on further developing the evolving connectionist systems(ECOS) that can bridge the gap between human brain signals and assistive robots.
Brain-Computer Interface, Detection of various brain diseases, Emotion Recognition