Brain plasticity in dyslexia after computer training: Spectral analysis based on statistical t-test

Dyslexia is a brain-based reading disability characterized by deficit in phonological processing. In this paper the effect of computer working memory (WM) training on EEG signals of 15 dyslexic children has been investigated. For this purpose, three sets of EEG were recorded from the subjects before and after the computer treatment. Each set of EEG was recorded while the subjects were doing a visual task that involves the working memory or the attention. Three spectrum feature types have been extracted from the recorded signals, t-test as a statistical approach for finding the most affected features due to the treatment, has been applied on the spectrum features. By analyzing the results, it could be found that slow wave index (SWI) and relative spectral power (RSP) features are more affected by the treatments, comparing to the harmonic parameters (HPs). In addition, it has been observed that the extracted features from theta and delta bands and RSP in theta sub-bands are more affected by the computer intervention. Finally, by analyzing the most repeated electrodes, among the responsive features to the treatments, the anatomic regions of brain affected by the treatment have been found.