Classifying cognitive load from galvanic skin response time domain features

Abstract

When performing a task, a certain amount of working memory is used. The amount of information a human is trying to process in working memory at any time is known as cognitive load. Biometrics can be used to determine this load. In this thesis galvanic skin response measurements were used to classify the cognitive load of a person. To study this, a pilot experiment was conducted in which every participant performed four tasks with varying di culty levels, each associated with a certain cognitive load level. An average 4-class classi cation accuracy of roughly 35%, and an average 2- class classi cation accuracy of around 62% was achieved.