Sensitive and Diagnostic Mental Workload Classifier

We propose to develop a sensitive and selective workload classifier, called PHYSIOPRINT (Physiology and Performance Research Integration Tool), that will ultimately operate in real time on multiple physiological signals (EEG, EKG, EOG, EMG) acquired and processed by our wearable and wireless X24 system. The raw signals will be converted into input variables for the classifier using a suite of proprietary real-time algorithms that include noise reduction, spectral decomposition and topographic mapping of the EEG signals, extraction of event-related potentials, detection of eye blinks and fixations, calculation of heart rate and heart rate variability, detection of EMG bursts and tonic activity, calculation of respiratory rate and detection of the head/body position and movements. PHYSIOPRINT will be designed around the IMPRINT model of mental workload and will discriminate between seven workload types (visual, auditory, cognitive, speech, tactile, fine and gross motor). PHYSIOPRINT will also provide a measure of overall workload construed to account for potential conflicts between different types of workload. Phase I research, which will include analysis of a large database of physiological data acquired during military-relevant tasks and a pilot study in a driving simulator, will define the PHYSIOPRINT design concept and development approach for Phase II.