Breaching Subjects' Thoughts Privacy: A Study with Visual Stimuli and Brain-Computer Interfaces.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computer Science

Background:

• Brain-computer interfaces (BCIs) are expanding beyond clinical use into entertainment and learning.
• Monitoring neuronal activity, like evoked potentials, is crucial for understanding central nervous system responses.
• Data sensitivity in BCIs necessitates secure transmission methods, with blockchain offering integrity solutions.

Purpose of the Study:

• To investigate the potential for extracting sensitive information from evoked potentials using visual stimuli in BCIs.
• To address the gap in literature regarding the impact of subliminal visual stimuli on BCI data.
• To analyze the relationship between visual stimuli, P300 evoked potentials, and subject privacy.

Main Methods:

• Conducted five experiments on ten subjects to assess the impact of visual stimuli on brain responses.
• Presented familiar visual stimuli, gradually decreasing image sampling time from supraliminal to subliminal.
• Monitored for P300 evoked potentials in response to visual stimuli across varying exposure levels.

Main Results:

• Supraliminal visual stimuli elicited P300 potentials in approximately 50% of trials across subjects.
• Decreasing image sampling time reduced the effectiveness of potential data extraction.
• The impact of subliminal visual stimuli on P300 potentials was not definitively confirmed.
• Younger subjects generally exhibited shorter response latencies to visual stimuli.

Conclusions:

• This study confirms that sensitive data can be extracted from subjects using visual stimuli and P300 potentials.
• The findings highlight significant privacy concerns associated with BCI technology, particularly concerning visual information processing.
• Further research is needed to fully understand the effects of subliminal stimuli and to develop robust privacy-preserving techniques for BCIs.