Predicting Individual Cognitive Status Based on EEG Data Fit to Power Law Distribution

As the immersion of the ageing population in digital technologies and cyberspace becomes more and more ubiquitous, techniques for assessment of cognitive status gain in importance. For instance, it is known that electroencephalography (EEG) data can characterize various functional conditions of the brain, including pathological ones. In our chapter, we investigate whether EEG can be used to predict the Montreal Mental State Scale (MMSE) scores that are widely used as the metrics of cognitive status in elder people. Inspired by existing studies that found the relationship between changes in EEG spectral power of the theta rhythm and cognitive status, we relied on power law to describe the distribution of the spectral power over the theta rhythm in the right temporal-parietal site of the cortex. We found statistically significant association of the cognitive status assessment scores (MMSE) and the goodness-of-fit to power law statistics (GOF), obtained with Kolmogorov-Smirnov test in combination with maximum-likelihood, for the distribution of spectral power over the TP8 electrode. The regression model we constructed with the GOF and the subjects’ age factors was highly significant and had R2 = 0.513, which implies that the application of power law for EEG analysis in order to classify the functional state of the brain can be considered promising.