Amir Fallah Ramezannezhad - Department of Computer Engineering, Lahijan Branch, Islamic Azad University, Lahijan, Iran,
Kamrad Khoshhal Roudposhti - Department of Computer Engineering, Lahijan Branch, Islamic Azad University, Lahijan, Iran,
Mohsen Falah Rad - Department of Computer Engineering, Lahijan Branch, Islamic Azad University, Lahijan, Iran,

The human brain's activity is represented by the brain signal, a trustworthy biometric marker. To gain permanent access to systems, a continual process called human authentication is used. The effectiveness of the human authentication systems ought to be steady over time. By examining the variables that affect the system's long-term accuracy using machine learning approaches, this study intends to address the dearth of research on brain signal-based biometric human identification. The study will concentrate on various time periods and brain signals related to motor imagery. The results of the study showed that the length of sample affected how well systems performed. The proposed technique attempts to be unaffected by temporal variables and different kinds of motor imagery. Extracting autoregression model and the Wavelet Transform features from occipital and central electrodes as support vector machines inputs, showed successful performance in human authentication during motor imagery throughout a range of time intervals.

Authentication،Biometrics،Brain Signals،Motor Imagery