A Cybersecure P300-Based Brain-to-Computer Interface against Noise-Based and Fake P300 Cyberattacks

Area of Science:

• Neuroscience
• Computer Science
• Cybersecurity

Background:

• The increasing integration of Internet of Things (IoT), ubiquitous computing, and artificial intelligence necessitates robust cybersecurity measures.
• Current cybersecurity in brain-to-computer interfaces (BCIs) is underdeveloped, posing significant risks to user safety.
• Standard algorithms in BCI systems are insufficient to detect and mitigate cyberattacks.

Purpose of the Study:

• To address the cybersecurity vulnerabilities in brain-to-computer interfaces (BCIs).
• To propose and evaluate a novel solution for enhancing the security of P300-based BCI systems using EEG data.
• To demonstrate the effectiveness of the proposed method against simulated cyberattacks.

Main Methods:

• Simulated cyberattacks, including fake P300 signals and noise-based attacks, were used to test the inadequacy of Support Vector Machine (SVM) algorithms in detecting intrusions.
• Compared the performance of SVM models using both real and simulated hacked P300 datasets.
• Implemented an EEG channel mixing approach to identify anomalies indicative of hacking in the BCI transmission channel.

Main Results:

• Validated that standard SVM algorithms fail to identify simulated cyberattacks in P300-based BCI systems.
• The proposed EEG channel mixing architecture successfully identified 99.996% of simulated cyberattacks.
• The implemented counteraction preserved the majority of BCI functionalities while mitigating threats.

Conclusions:

• Existing SVM algorithms are insufficient for securing P300-based BCIs against sophisticated cyber threats.
• The developed EEG channel mixing approach offers a highly effective solution for real-time cyberattack detection in BCIs.
• This research significantly advances BCI security, ensuring user safety and system integrity.