Brain-based Authentication: Towards A Scalable, Commercial Grade Solution Using Noninvasive Brain Signals

Abstract

In this study we report on a field test where we asked if it is feasible to deliver a scalable, commercial-grade solution for brain-based authentication currently given available head wearables. Sixty-two (62) participants living across the United States in autumn 2020 completed four (4) at-home sessions over a single (1) week. In each session there were six (6) authentication events consisting of rapid presentation of images (10Hz) that participants watched for 10 seconds while recording their brain signal with an off-the-shelf brain signal measuring headband. The non-stationary nature of the brain signal, and the fact that the signal results from a superposition of hundreds of simultaneous processes in the brain that respond to context makes the data unique in time, unrepeatable, and unpredictable. Even when a participant watched identical stimuli, we find no two periods of time to be alike (Fig. 4B) and furthermore, no two combinations of time periods are alike. Differences within people (intra-) and across people (inter- participant) from session to session were found to be significant, however stable processes do appear to be underlying the signal complexity and non-stationarity. We show a simplified brain-based authentication system that captures distinguishable information with reliable, commercial-grade performance from participants at their own homes. We conclude that noninvasively measured brain signals are an ideal candidate for biometric authentication, especially for head wearables such as headphones and AR/VR devices.

Passwordless Authentication Using a Combination of Cryptography, Steganography, and Biometrics

“…However, the human brain has also been observed to produce (unconscious) unique identification data in the form of electrical signals that can be captured using non-invasive means [5]. Based on this, brain-based authentication was proposed by Kopito et al (2021) [6] as a secure method of user authentication. In this scheme, the user's brain signals are captured and processed using machine learning models to extract features that are stored for use during authentication.…”

“…While iris-based systems have been recorded to have the highest degree of accuracy of all biometrics, implementing them as a password alternative requires extensive research. Take the enrolment process, for instance: "A specialised camera, typically within three feet of the subject, uses an infrared imager to illuminate the eye and capture an extremely high-resolution photograph" [6]. Unless current devices are upgraded with the capability for such functionality or future devices engineered for this purpose, enrolment will have to be conducted in conditions like that of brain-signal authentication, making it infeasible for mass adoption.…”

A Review of Password-less User Authentication Schemes.

“…Based on this, brain-based authentication was proposed by Kopito et. al, (2021) 22 as a secure method of user authentication. In this scheme, the user's brain signals are captured and processed using machine learning models to extract features that are stored for use during authentication.…”

“…While iris-based systems have been recorded to have the highest degree of accuracy of all biometrics, implementing it as a password alternative requires extensive research. Take the enrolment process for instance:" A specialised camera, typically within three feet of the subject, uses an infrared imager to illuminate the eye and capture an extremely highresolution photograph" 22 . Unless current devices are upgraded with the capability for such functionality or future devices engineered for this purpose, enrolment will have to be conducted in conditions like that of brain-signal authentication, making it infeasible for mass adoption.…”