Detection of Sympathetic Activation through Measurement and Adaptive Processing of the Pupil Diameter for Affective Assessment of Computer Users

Future human-computer interactions could be enhanced by enabling the computer to detect the user‟s emotional states (e.g. stress), and adjust its interaction behavior appropriately. However, providing the computer with the ability to detect important changes in the user‟s affective state is a challenging problem. We propose that increases in sympathetic activation which typically accompany the onset of stress could be detected noninvasively by monitoring and processing of the pupil diameter (PD) in the computer user. Our previous work indicated that, under constant illumination conditions, there are measurable increases of the pupil diameter in response to stress stimuli, which we have termed the Pupillary Affective Response (PAR). On the other hand, it is known that the Pupillary Light Reflex (PLR) plays a significant role in modifying the pupil diameter and, therefore, the PAR must be separated from the PLR to be useful towards affective assessment of a computer user in environments with variable illumination. To that end, this study adds the measurement of illumination (IL) in the neighborhood of the subject‟s eyes and applies it to the processing of the PD signal through an Adaptive Interference Canceller (AIC) using the H ∞ time-varying (HITV) adaptive algorithm, to remove the influence of the PLR from the pupil diameter variations originally measured. The output of the AIC, the Modified Pupil Diameter (MPD) signal, is subjected to additional processing, to result in a detection feature that is compared against a threshold to obtain an indication of subject “stress”. The performance of this stress detector, along with 9 others derived from Galvanic Skin Response (GSR) and Blood Volume Pulse (BVP) signals is studied through their Receiver Operating Characteristic (ROC) curves.