The Use of EEG-Based Authentication Models
Faculty Mentor Information
Dr. Sindhu Kalathur Gopal, Boise State University
Presentation Date
7-16-2026
Abstract
Biometric data such as fingerprints is increasingly being used for security and authentication purposes due to how unique this kind of information can be to each individual person. But what happens when this information is easily obtainable and is taken by a malicious actor? For this, we propose the use of EEG-based authentication. EEG data is highly unique and sensitive for each individual and leaves no visible traces unlike fingerprints, and they are also not visible unlike mannerisms. We compared and analyzed EEG data belonging to a control group and a group of participants with migraines to test the performance of an EEG-authentication model that could be used to grant access to specific users.The EEG data of the participants recorded them being exposed to various visual and auditory stimuli as well as resting states. We have recorded the experimental results and insights of the model that could be used for authentication and how it performs when trained to look for a specific user.
The Use of EEG-Based Authentication Models
Biometric data such as fingerprints is increasingly being used for security and authentication purposes due to how unique this kind of information can be to each individual person. But what happens when this information is easily obtainable and is taken by a malicious actor? For this, we propose the use of EEG-based authentication. EEG data is highly unique and sensitive for each individual and leaves no visible traces unlike fingerprints, and they are also not visible unlike mannerisms. We compared and analyzed EEG data belonging to a control group and a group of participants with migraines to test the performance of an EEG-authentication model that could be used to grant access to specific users.The EEG data of the participants recorded them being exposed to various visual and auditory stimuli as well as resting states. We have recorded the experimental results and insights of the model that could be used for authentication and how it performs when trained to look for a specific user.