System identification for the contribution of light variations to pupil diameter change

The pupil diameter (PD) is adjusted by muscles under the control of the sympathetic and parasympathetic divisions of the Autonomic Nervous System (ANS). It is well known that a major mechanism for PD adjustment is the pupillary light reflex (PLR), by which the PD decreases as a response to increased illumination on the retina. However, it has also been found that PD is modified by ANS changes due to affective variations in the subject. Ultimately, we pursue the reproduction and cancellation of PLR-driven PD changes from the PD signal as measured during human-computer interaction, using an Adaptive Interference Canceller (AIC), so that the PD changes not driven by PLR may be used to gauge the affective changes of the computer user. As a preliminary step towards that goal we studied the actual performance of an AIC in modeling changes in measured PD caused exclusively by changes in illumination, which were simultaneously recorded by a light sensor. Our results confirm that the AIC was able to converge, minimizing its error to acceptable levels for all of our 8 subjects. Furthermore, the impulse response model implicitly formed in the weights of the adapted AIC seemed to generally coincide with the impulse response that previous research by others would predict for the transfer function mediating between changes in illumination and changes of pupil diameter.