Analysis of EEG based functional connectivity during mental arithmetic task

Recent studies in computational neuroscience have shown that functional connectivity graphs contain significant information about mental states. In this research, we analyzed functional connectivity graphs constructed based on EEG signals recorded during mental arithmetic tasks. We implemented various functions using mathematical tools in order to construct the functional connectivity to check whether it is possible to classify the quality of mental calculations with them or not. We used both unsupervised and supervised methods to ensure that the data can be separated even if we ignore labels. Based on classification and clustering, we showed that the subjects can be separated based on quality-quantity characteristics. Using methods like spectral analysis, mutual information and granger causality, we also found some channels and connectivities statistically more significant than others, mostly close to cortical areas previously detected as related to mental arithmetics.