Detecting mild traumatic brain injury for athletes using SSVEP classification : a case study

Mild traumatic brain injury (mTBI) can have detrimental impacts on the well-being of individuals, especially athletes with millions of injury cases reported per year. Nevertheless, the current assessment and diagnostic tools for mTBI have limitations due to their subjectivity and the lack of accessibility. This study aimed to evaluate the potential of machine learning algorithms in combination with steady-state visual evoked potentials (SSVEP) to provide mTBI diagnoses. The participants of this study included 36 athletes diagnosed with mTBI, aged 17–54, and 400 matched healthy controls without mTBI. Altogether, we extracted 51 SSVEP-based features from the collected observations and transformed them via principal component analysis (PCA) for feature reduction. Several machine learning algorithms were trained and validated using the transformed features for further analysis and comparison. Linear Discriminant Analysis (LDA) was found to be the best-performing classifier with 62 % balanced accuracy and has the potential to improve further with additional data. Overall, the findings of this study indicate that machine learning models have the potentials to be utilized as a diagnostic tool for mTBI when used with SSVEP data.