Abstract This chapter presents a functional near-infrared spectroscopy study to differentiate between drowsy and active states in drivers. Passive brain signals were acquired from the dorsolateral prefrontal cortex of five healthy subjects. After preprocessing the acquired signals, mean and slope values were extracted as features for classification. A support vector machine (SVM) and linear discriminant analysis (LDA) were used as classifiers to test and train the data. The average classification accuracies acquired using SVM and LDA were 74.3 ± 2.5% and 73.0 ± 2.7%, respectively, using signal mean and signal peak calculated within a 0–7 s time window. The results demonstrate the suitability of the proposed method for drowsiness detection with minimum delay, thereby adding another attribute for driver safety.