Date of Award
Brain Computer Interface, BCI, Assistive Technology, Neurotechnology, Event Related Potentials, ERP
Brain Computer Interfaces (BCIs) allow individuals to operate technology using (typically consciously controllable) aspects of their brain activity. Auditory BCIs utilize principles of Auditory Event Related Potentials or Auditory Evoked Potentials as a reproducible controllable features that individuals can use to operate a BCI. These Auditory BCIs in their most basic format can allow users to answer yes or no questions by listening to either one auditory stimuli or the other. Current accuracy in intended response detection for these kinds of BCIs can be as good as mean accuracy of 77 \%. BCI research tends to optimize the computer side of the system however the ease of use for the human operating the system is an important point of consideration as well. This research project aimed to determine what factors make a human operator able to achieve the highest accuracy using a given previously successfully demonstrated classifier. This research project primarily sought to answer the questions; to what degree people can improve their accuracy in operating an Auditory BCI and what factors of the stimulus used can be altered to achieve this. The results of this project, obtained through the data collected from six individuals, found that slower stimuli speeds for eliciting Auditory Event Related Potentials were significantly better at achieving higher prediction accuracies compared to faster stimulus speeds. The amount of time spent using the system appeared to result in diminishing returns in accuracy regardless of condition however not before an initial spike in greater classifier prediction accuracy for the second condition run on each subject. Although further research is needed to gain more conclusive evidence for or against the hypothesis, the results of this research may be able suggest that individuals can improve their performance using Auditory BCIs with practice at optimal parameters albeit within a given time frame before experiencing diminishing returns. These findings would stand to provide benefit both to continued research in making optimal non-invasive alternative communication technologies as well as making progress in finding the potential ceiling in accuracy that an Auditory BCI can have in interpreting brain activity for the intended action of the user.
Ferrisi, Leonardo M., "Optimizing an assistive Brain Computer Interface that uses Auditory Attention as Input" (2023). Honors Theses. 2705.