A novel approach for ensuring the privacy of EEG signals using application-specific feature extraction and AES algorithm

Abstract

Human brain known commonly as an individual's computer is one of the complex organs in the body. The brain serves to control and coordinate the mental and physical actions. It has the capacity to store innumerable data. This data includes not only childhood memories but also those related to the present age. Tracking the electrical potentials or impulses, commonly known as brain waves, can be utilized for various purposes ranging from diagnosing diseases to decoding personality traits. This can be done using the technique of Electroencephalography (EEG). The amount of data available is exploding so also the opportunities of data mining thereby mounting privacy risks. The current practice either prevents the reuse of raw EEG signals or disrespects participants' right to privacy by using this data for purposes other than those consented to, by the individual. The aim of this work is to put a stop to such privacy risks by extracting and transferring only those features related to an application. Feature is a specific characteristic measurement of a pattern segment. EEG signal processing pertains mainly to feature extraction. There are different methods for extracting features from EEG signals. In this work, the features related to neurological diseases, namely epilepsy and Alzheimer's disease, are extracted using Discrete Wavelet Transform (DWT) and those related to Schizophrenia are extracted using Autoregressive (AR) method. For securing the raw data, it is doubly encrypted using Advanced Encryption Standard (AES) algorithm and is kept under the control of each user. The proposed system thus ensures the privacy of an individual's EEG signals, and mitigates any risk that may arise if raw data is made available as such to the public.