TRAITEMENT DU SIGNAL, vol.38, no.6, pp.1767-1773, 2021 (SCI-Expanded)
The COVID-19, which has rapidly spread and infected millions of people from all over the world, causes various problems including psychiatric, economic, educational as well as health. Many studies have been reported that COVID-19 can be characterized by vascular damage predominantly involving micro vessels. In this study, we proposed a method to examine whether COVID-19 effects on brain computer interface (BCI) performance or not. We collected P300 based electroencephalogram (EEG) signals from six subjects before and after the COVID-19 infection. For classifying the P300 and non-P300 EEG signals, single output and two-layer artificial neural network was utilized. Based on the t-test analysis, it was observed that there was a significant difference between the before and after COVID-19 infection test groups especially on Oz channel in occipital region for alpha=0.05 percent while that of for alpha=0.01 percent shows no statistical difference for P300 classification results. The latency values, on the other hand, before and after COVID-19 infection did not represent any difference for both significance levels. It is clearly understood from the literature that COVID-19 negatively effects to the microvascular bed. Therefore, it might be expected that it could cause to reduce the P300 based BCI performance. This was the first study to investigate the impact of COVID-19 on P300-based BCI performance, taking into account the EEG signals of the COVID-19 infection. The obtained results showed that although the COVID-19 infection did not generally effected P300 based BCI application performance and latency values, the performance of the occipital region electrodes slightly effected.