Mild Cognitive Impairment (MCI) and Subjective Cognitive Impairment (SCI) are brain disorders with a high risk to progress to Alzheimer’s disease. This work investigates the possible EEG differences in MCI and SMC subjects, with respect to control subjects (CS), both during rest and cognitive stimulation. EEG data were analyzed by decomposing the power spectral density into frequency sub-bands and by calculating the power-law beta exponent parameter. Moreover, from a nonlinear point of view, also Higuchi’s fractal dimension and Poincaré plot indexes were estimated. Feature extraction was performed by using the Principal Component Analysis method, in an effort to distinguish CS, MCI and SCI patients. Results show that some parameters present statistically significant differences between SCI and control subjects, whereas MCI patients present intermediate values between the other two groups. The use of the principal component analysis allows a preliminary visual separation of CS and SCI, even if the difficulty in distinguishing MCI subjects persists.
