Flashing checkerboard and attentional effort: SSVEPs signals analysis and adjustment of stimulation frequencies

Patients in the complete locked-in state (CLIS) are unable to interact with the reality around them and to manifest their condition, although they are aware of what is happening around them. Due to a complete interruption between the cortex neurons of the Central Nervous System and the motor neurons responsible for movement, the impulses from the brain cannot reach all the extremities involved in movement. The only technological support able to compensate for the lack of functionality of these neurons is the BCI. Specifically, in SSVEP-based BCIs, the signal components that are derived in response to a visual stimulus are the SSVEPs, which can be recorded on the primary visual cortex with a non-invasive electroencephalogram (EEG). Most of the BCIs available on the market use the eye movement of LIS (Locked-In Syndrome) patients, but CLIS (Complete Locked-in Syndrome) patients are also unable to perform this type of movement: in order to investigate this aspect, the study carried out in this thesis is aimed precisely at patients who are in the latter stage of the disease. In SSVEP-based BCIs, the type of visual stimulus administered influences both the goodness of the results and the stress caused to the patient. Using an effective but low-stress stimulus can improve the already problematic communication in such subjects: in this work, a type of visual stimulation that tries to satisfy this condition was tested. A visual stimulus consisting of a checkerboard was designed to allow the subject to provide a binary response. It is made up of very small squares, close to each other, coloured with two different colours and in which each colour flashes at a certain frequency, whose values are within the alpha band of the EEG signal. In this way the patient, unable to move his eyes, has less difficulty in concentrating on one or the other colour, depending on the response he wants to give: he does not have to concentrate on the single square object, but on the whole colour associated with the response to be given. In addition, stimulation frequencies adapted to the subject were sought. These frequencies, which may vary from subject to subject, elicit the best responses for the individual subject. In order to explore all these aspects, a non-invasive EEG signal sampling system was used, allowing us to collect, on healthy subjects, the electroencephalographic signal during periods of visual stimulation at different frequencies and during periods of rest. During stimulation, particular strategies were used to reproduce the fixed gaze condition that characterises CLIS patients. Thus, offline signals were analysed and processing techniques were implemented in order to build an algorithm able to derive the best stimulation frequencies for each subject and to guarantee a binary response.