Koosha Sadeghi

E-BIAS: A Pervasive EEG-Based Identification and Authentication System

Security systems using brain signals or Electroencephalography (EEG), is an emerging field of research. Brain signal characteristics such as chaotic nature and uniqueness, make it an appropriate information source to be used in security systems. In this paper, E-BIAS, a pervasive EEG-based security system with both identification and authentication functionalities is developed. The main challenges are: 1) accuracy, 2) timeliness, 3) energy efficiency, 4) usability, and 5) robustness. Therefore, we apply machine learning algorithms with low training times, multi-tier distributed computing architecture, and commercial single channel dry electrode wireless EEG headsets to respectively overcome the first four challenges. With only two minutes of training time and a simple rest task, the authentication and identification performance reaches 95% and 80%, respectively on 10 subjects. We finally test the robustness of our EEG-based seamless security system against three types of attacks: a) brain impersonation, b) database hacking, and c) communication snooping and discuss the system configurations which can avoid data leakage.

Optimization of Brain Mobile Interface Applications Using IoT

Pervasive Brain Mobile Interfaces (BMoI) can be made more accurate and time efficient when knowledge from other sensors and computation power from available devices in the Internet of Things (IoT) infrastructure are utilized. This paper takes the example of Neuro-Movie (nMovie), an interactive movie application that blurs movie scenes based on mental state, to illustrate and analyze optimization opportunities when BMoI is interfaced with IoT. The three way trade-off between accuracy, real-time operation, and energy efficiency can be optimized through usage of physiological responses from IoT sensors and prediction algorithms. Latency and power models of BMoI are developed for thorough analysis of the trade-offs. Experiments on 10 volunteers show that: a) utilizing electrocardiogram responses to psychological stimulus increases the accuracy of mental state recognition by almost 10%, b) predictive models cover computation and communication latencies in the system to satisfy real-time requirements, and c) use of predictive models allows duty cycling of smartphone WiFi that potentially saves upto 71.6% communication energy.

Toward Parametric Security Analysis of Machine Learning Based Cyber Forensic Biometric Systems

Machine learning algorithms are widely used in cyber forensic biometric systems to analyze a subjects truthfulness in an interrogation. An analytical method (rather than experimental) to evaluate the security strength of these systems under potential cyber attacks is essential. In this paper, we formalize a theoretical method for analyzing the immunity of a machine learning based cyber forensic system against evidence tampering attack. We apply our theory on brain signal based forensic systems that use neural networks to classify responses from a subject. Attack simulation is run to validate our theoretical analysis results.

Permanency analysis on human electroencephalogram signals for pervasive Brain-Computer Interface systems

Brain-Computer Interface (BCI) systems use some permanent features of brain signals to recognize their corresponding cognitive states with high accuracy. However, these features are not perfectly permanent, and BCI system should be continuously trained over time, which is tedious and time consuming. Thus, analyzing the permanency of signal features is essential in determining how often to repeat training. In this paper, we monitor electroencephalogram (EEG) signals, and analyze their behavior through continuous and relatively long period of time. In our experiment, we record EEG signals corresponding to rest state (eyes open and closed) from one subject everyday, for three and a half months. The results show that signal features such as auto-regression coefficients remain permanent through time, while others such as power spectral density specifically in 5–7 Hz frequency band are not permanent. In addition, eyes open EEG data shows more permanency than eyes closed data.