Comprehensive Summary
Brain computer interface (BCI) is able to provide individuals with brain injuries a method of communication. A BCI model surrounding electroencephalogram data was proposed. This model would contain Common Spatial Patterns as well as Temporal Patterns to differentiate between EEG signals. In this study, EEG data was collected from 15 paralyzed individuals with ten words in sets and the Common Spatial Patterns and Temporal Patterns were analyzed using the model. The results showed the average classification accuracies for the model, as well as maximum accuracies. The pairwise classification was 97.78%, which shows solving a multiclass problem is the most efficient by breaking it down into tasks, and the multiclass classification was 79.22%. The maximum accuracies are high for these models, which shows the potential positive outcome of implementing this BCI model.
Outcomes and Implications
The research in this study is important because degenerative diseases and traumatic injuries to the nervous system can render individuals unable to speak or communicate. The inability to speak can directly impact an individual's quality of life and BCI models can be created to bridge this gap. This work applies to medicine as scientists and clinicians are constantly looking for ways to help patients, as well as to use models that are more efficient in differentiation diagnoses and allowing individuals to perform necessary functions. Communication is an important part of a person's daily life, so being able for a person to communicate if they are unable to due to extenuating circumstances is necessary. The BCI model containing Common Spatial Patterns and Temporal Patterns was proposed to address the problem that patients with brain disorders, injuries, and degenerative diseases possess.