Analysis of Relationships between Combinations of Biological Signals and Subjective Interest

Abstract

—Interest greatly affects human behavior. Interest is often measured by a questionnaire or observational methods and is used for marketing or advertising purposes. However, there are some problems in the reliability of results of such conventional methods. We propose that using biological signals can help address these issues because such signals are associated with biological reactions to interest. This study describes an experiment and multiple linear regression analyses of the relationships between a viewer’s interest in an infomercial and the viewer’s biological signals such as electroencephalogram, hemoencephalography, blood volume pulse, skin conductance, and respiration. The subjective interests of viewers were measured by a questionnaire. A comprehensive index of interest based on the answer values was extracted by principal component analysis. We then constructed an individual regression equation to estimate the degree of interest from different combinations of biological signals. 