Analysis of Electrodermal Activity Signal Collected During Visual Attention Oriented Tasks

Abstract

In this paper, we have measured Electrodermal Activity (also known as Galvanic Skin Response or EDA) signal of young adult human subjects, during tasks associated with Visual attention, in order to investigate whether these signals carry any information regarding human visual attentiveness. We find that the phasic component of the EDA signal carries more information about visual attentiveness compared to the tonic component. Specifically, zero crossing rate is a feature that is found to be useful in differentiating between periods of low and high mental activity during visual attention related tasks. This paper proposes two new features namely the number of Gaussian peaks in the histograms of zero crossing rate and the overlapping area between two Gaussians which have been used to predict the visual attentiveness of the subject during tasks associated with visual attentiveness. A total of thirty human subjects have volunteered for the current experiment where they have been asked to complete three tasks (each task twice). These tasks are standard tests for visual attentiveness measurement. The proposed model has been tested on the collected dataset. In 91.3% cases, the result aligned with manual visual attentiveness test results. We have tried to investigate the EDA signal during good, average, and poor stages of visual attentiveness. Furthermore, the method gives better performance compared to the methods carried out using Electroencephalogram signals by other research groups.

Keywords

• Electrodermal activity
• galvanic skin response
• visual attention
• Weitzman number
• Gaussian mixture model of human visual attentiveness (GMMHA)
• CART