Relative Measures to Characterize EEG Signals for Early Detection of Alzheimer

Number of Alzheimer's patients are continuously increasing worldwide. Early detection of Alzheimer will improve patients life and world health care cost. A lot of modalities are there with the help of which Alzheimer can be detected at an early stage. But choosing a potential feature is a challenging task. In the present study we investigate frequency relative energy, frequency relative power and relative entropy from the dataset of Mild Cognitive Impairment (MCI), Normal and Dementia subjects. Bhattacharya distance is used to rank the aforementioned features that can classify between MCI, Control and Dementia subjects. Relative Entropy is identified as the best feature for classification. All the findings are statistically validated using Kruskal-Walis test. The selection of the relevant feature will be beneficial for early Alzheimer detection and may increase the quality of life of the patients suffering from the disease.