Pupillometric System for Cognitive Load Estimation in Noisy-Speech Intelligibility Psychoacoustic Experiments: Preliminary Results

Abstract

Cognitive load (CL) is the mental effort required to perform a task. It may be estimated by different ways, such as by electroencephalography, or analysis of pupil dilation (pupillometry). Due to the recent applications of CL for hearing aid performance assessment, this work presents a pupillometric system for CL estimation in speech intelligibility experiments. It consists of a one eye image acquisition sensor used in clinical video-oculography, and a signal processing software specially developed for recording pupil’s response and estimating CL. To overcome the adversities found in very low intelligibility conditions, a new CL index was proposed. Psychoacoustic listening experiments with 7 normal-hearing volunteers and noisy-speech compared intelligibility scores with pupil features for a wide range of signal to noise ratios (SNR). Preliminary results show a consistent increase of the proposed CL index with the SNR decrease. This index may be a useful tool for CL estimation in speech intelligibility investigation under very difficult listening situations. However, the observed findings are limited by the small testing group.