My research activity has always been focused on one of the most innovative and fascinating areas of bioengineering applied to neuroscience, the brain-computer interface (BCI), defined as “a system that measures Central nervous System (CNS) activity and converts it into artificial output that replaces, restores, enhances, supplements, or improves natural CNS output and thereby changes the ongoing interactions between the CNS and its external or internal environment, Wolpaw et al., 2012”. In this regard, I had the possibility to work with different types of BCI systems, by the involvement in many national and international projects (see sections VIII and XI), in particular (i) as assistive technology (i.e. communication and control), (ii) for rehabilitation purposes (i.e. motor imagery) and (iii) for “passive” monitoring of internal states of the user (i.e. workload, attention, stress, etc) while dealing with a task (i.e. driving a car or piloting an aircraft). My specific background as bioengineer, is focused on the (i) processing and features extraction of different kind of biosignals (i.e. electroencephalography-EEG, electrocardiography-ECG, photoplethysmography-PPG, Electro Dermal Activity-EDA, Electromyography-EMG, Electrooculography-EOG), and (ii) machine learning techniques able to employ such mentioned features to maximize BCI performances.
BCI for communication & control: At the beginning of my activity I worked with BCI systems for communication and control, for locked-in patients. In particular, it can be possible to decode some specific features extracted from the EEG signal of the subjects, and employ them as a communication and/or control channel. In this regard, I got great knowledge in processing EEG signals in time domain, extract and analyse Event Related Potentials (i.e. ERPs, P300 and N200 potentials). In this regard, I have developed an algorithm able to maximize the signal to noise ratio for an improved extraction of ERPs from the background EEG noise. At the same time, I had the possibility to deal with machine learning techniques (both linear and non linear) applied to such mentioned features, to be used to enhance BCI performances.
BCI for rehabilitation: The principle at the basis of this kind of BCI is that the system can be used to “reinforce” specific brain patterns of post-stroke patients, while performing simple tasks (e.g. grasping an object), and so fasten the rehabilitation phase. I have generated in this regard a hybrid BCI system that employ at the same time EEG and EMG signals, to maximize the reinforcement of physiological brain patterns, inhibiting the activation of pathological patterns. During this activity I got expertise in analysing frequency domain features of EEG signals.