EEG waveform identification based on deep learning techniques

"The use of Brain-Computer Interfaces can provide substantial improvements to the quality of life of patients with diseases such as severe Amyotrophic lateral sclerosis that cause Locked-in syndrome, by creating new avenues in which these people can communicate and interact with the outside world. The P300 speller is an interface which provide the patients the ability to spell letters and eventually words, so that they can speak while unable to use their mouth. The P300 speller works by reading signals from the brain using an Electroencephalogram. Traditionally, these signals were plotted and interpreted by specialized technicians or neurologists, but the development of Machine learning algorithms for classification allow the computers to perform this analysis and detect the P300 signals, which is an Event Related Potential triggered when certain stimuli such as a bright light is triggered on a place that the patient is focused on. In this thesis we used a Convolutional Neural Network to train multi-channel EEG readings, and attempted to detect P300 signals from a P300 speller. The results are corroborated against a public ALS dataset." Proyecto final Ingeniería Informática (grado) - Instituto Tecnológico de Buenos Aires, Buenos Aires, 2022