On the Viability of SSVEP Brain Computer Interfaces Based on Purely Imaged Stimuli

A Brain-Computer Interface (BCI) is a communication system that is able to translate the electrical activity of the brain into commands capable of driving an application. Usually the electrical potential is recorded by an electrophysiological monitoring method called Electroencephalograph (EEG), which is a device able to recognize the neural activity by means of a set of electrodes placed on the scalp of the user. In this case an EEG-based BCI system is mainly composed by 3 blocks: (1) a brain acquisition system made up of several sensors responsible for the data acquisition, (2) a computational system which extracts from the obtained data the task performed by the user, (3) a control system that elaborates these data and provides a feedback to the user. Previous works focused on a particular BCI category that relies on Steady-State-Visual-Evoked Potentials (SSVEPs), which means that the brain signals, evaluated by the system, were triggered by external stimuli. In our case the stimulus is a flickering light at constant frequency that generates a similar signal at the neuronal level. However when the subject has to imagine a visual stimulus without receiving any stimulus from the outside, we talk about visual imagery. The purpose of this work is to evaluate whether it is possible to imagine a visual stimulus, such as a flashing light with a certain frequency, and to be able to pick up a signal that is similar to the real one and therefore distinguishable from the rest. Actually, starting from the SSVEP paradigm, the aim of the work is to investigate brain activation during visual imagery and in particular whether visual perception and visual imagery involve the same neuronal activation. Finally, staring at a flashing light for a long time can also cause several problems that would not occur with visual imagery. For this reason this work was designed to allow paralyzed patients to carry out various activities related to several types of interactive applications involved with neuroscience.