Single-Trial Visual Evoked Potentials Estimation Using Support Vector Machine

Exploiting the novel human-machine interaction paradigm called brain-computer interface, we developed a mental speller by constructing a virtual keyboard in a computer screen for those with neuromuscular disorders and motor disabilities. The carriers of communication between brain and computer were induced by a so-called "imitating-natural- reading" paradigm. Single-trial feature estimation should be used in this online paradigm instead of grand average that usually used in the cognitive or clinical experiments. With carefully signal preprocess and feature selection procedure, we explored the single-trial estimation of EEG using v-support vector machine in three subjects, and gained a satisfied classification accuracy of 92.6%, 88.2% and 93.5%, respectively. These results demonstrated the advantages of the inducing paradigm in the construction of our mental speller.