Mind your privacy: Privacy leakage through BCI applications using machine learning methods

摘要: Abstract With the digitization of almost every aspect our lives, privacy leakage in cyber space has become a pressing concern. Brain–Computer Interface (BCI) systems have more popular recent years and are now being used for variety applications. BCI data represents an individual’s brain activity at given time. Like many other kinds data, can be utilized malicious purposes. Electroencephalography (EEG) is one most acquisition methods More specifically, games, represent main EEG However, application (e.g. game) could allow attacker to take advantage unsuspecting user happily enjoying game record user’s activity; by analyzing this infer private information characteristics regarding user, without his/her consent or awareness. This study first demonstrate ability predict meaningful personality traits cognitive abilities resting-state (rsEEG) recordings using machine learning methods. A comprehensive set raw rsEEG scans, along with dissociation level executive function (EF) performance measures, 162 subjects were evaluation. The best results we achieved accuracy 73% classification less than 16% mean absolute error predicting all examined EFs. These encouraging better those presented prior research, both terms data-validity dataset size.