Discrimination of Emotional Type by Heartbeat Signal Information

Information obtained from heartbeat signal that is associated with emotional responses was extracted and discriminant model for estimating the type of emotion from the information was developed. In healthy subjects, ECG was recorded continuously during watching movies, during which they pressed a button according to the type of emotion. Power concentration of high-frequency (HF, 0.15–0.45 Hz) component (Hsi) that reflects regularity of respiration, the standard deviation of the amplitude of HF component, and the amplitude of low-frequency (LF, 0.04–0.15 Hz) were extracted as the useful discriminator for the type of emotion. Canonical discriminant model with these variables revealed that periods with joy, worry, sorrow, surprise, and aversion are expanded separately on a plane consisting of two canonical variables reflecting the levels of valence and arousal, respectively.