Novel EEG Security Inspection for Identifying Insider Intrusions by ERP features and P300 signal using Artificial Intelligence.

  • Alanoud S. Almemari

Student thesis: Master's Thesis


Despite the growing interest in the field of security, there is yet to be developed a system that can inspect insider intrusion intent before it happens. However, many systems have been developed to protect critical places from outsider intrusion attacks by applying various authentication mechanisms or restrictive technologies to detect any attack attempts. All this great focus led to overlook the knowledge or access privileges that an employee has which in turn can maliciously misuse their role to harm the place or damage the system. Therefore, this research project aims to implement a reliable and accurate system to recognize any intention to cause damage or even investigate any previous incidents that might has a relation with insider attack. This can be achieved through detecting concealed information stored in the brain by measuring Event relatedpotential signal features and P300 signal. Brain wave signals will be recorded using EMOTIV headgear after asking specific questions to individuals to check their will in damaging or attacking the place. Next, the system will pre-process and analyze the signal to check the presence of specific ERP features and ERP-P300 signal that occurs when the subject recognizes the stimuli or question and has prior knowledge about it and lied about it using machine learning or deep learning algorithms. In order to train artificial intelligent model, an experiment is designed to collect data for the training purpose. Artificial intelligent algorithms will be used to investigate the most reliable and accurate algorithm. At this stage, data will be trained using convolutional neural network model in python, as the data will be represented by continues wavelet transform to deal with the low signal-to-noise ratio (SNR) before the training process.
Date of AwardJul 2021
Original languageAmerican English


  • Insider Intrusion
  • Signal-to-noise Ratio
  • Event Related potential
  • Artificial Intelligence
  • Deep Learning
  • Machine learning.

Cite this