Jodi M.
Impairment was identified with 76.4 percent accuracy, 69.8 percent positive predictive value, and a 10 percent false-positive rate with machine learning models using fNIRS time course features and connectivity matrices and using convergent classification as ground truth.