Brain Computer Interface and Eye Tracking Systems as Assistive Technologies

Electrical signals from the brain can be employed to create a direct information channel between the human brain and the computer. Recognizing patterns of these signals can let people’s intentions be expressed with no need of movements. Tracking the user’s gaze on a computer screen is another hands-free method for human-machine interaction. In this thesis, both technologies are combined to develop a hybrid spelling system for people affected by severe disabilities. The Brain Computer Interface system exploits auditory evoked potentials to select letters in order to write words. The Eye Tracking is combined with this system to let the user delete an unwanted letter or to save a word. A user of this system would only listen to auditory stimuli and make simple eyes movements. No other movements are needed. Experiments have been conducted on 10 healthy subjects, resulting in a 81% sensitivity for the Brain Computer Interface system and a 99% accuracy for the Eye Tracking system. The project combines electroencephalographic signal acquisition, processing and classification with gaze detection, and makes the interaction with the user possible through a graphic user interface.