Event-Related Potentials' jitter compensation for Brain-Computer Interfaces applications

Electroencephalogram (EEG) is a measure of brain activity that contains abundant information about human brain function. For this reason, recent clinical brain researches and Brain-Computer Interface (BCI) studies use EEG signals in many applications. We studied a particular Event-related Potential, named Readiness Potential, connected to volitional motor intention, to explore cognitive functions in patients with Disorders of Consciousness as Coma, Vegetative State, and Minimally Consciousness State. We studied the neural correlates of intention to move in order to investigate the self-awareness and of the environment. In the early 1980s, B. Libet found that RP starts about 550 ms prior to freely voluntary acts, but the consciousness of the act occurs about 200 ms before the EMG onset. The Readiness Potential or Bereitschaftpotential (BP) can be distinguished into two segments: the “early BP”, related to the intention to move, and the “late BP” related to the movement itself. The BP is maximum in correspondence of the contralateral motor area to the movement, thus, it is possible to understand whether a subject is going to move the left or the right limb. We carried out an experimental session in healthy volunteers under different experimental conditions. We detected RPs to distinguish between voluntary, semi-voluntary and involuntary movements. We used EEGLAB, its plug-in MRCPLAB particularly, to analyze the EEG and EMG recordings. The RP is obtained by averaging the EEG over many trials, aligning the epochs by using the onset of the movement as a trigger. Averaging is very sensitive to latency jitter, which is a time misalignment of the epochs used to compute the RP. The misalignment is due to small changes in the brain times across all the trials and to movements' onsets timing detection. The goal of this thesis work is to implement a Matlab algorithm, based on Woody’s method to compensate the jitter between trials, realigning the epochs. This algorithm improves the averaging and enhances the RP signal quality. To verify the reliability of the algorithm, the SNR (signal-noise ratio) is calculated. The algorithm about jitter compensation is added as a function of MRCPLAB.