THE EMOTIONAL RECOGNITION RESEARCH ON THE F0 EFFECTS OF ERP COMPONENTS OF SPEECH SIGNAL

As an effective method to study brain recognition process of emotional speech, the event-related potential (ERP) provided numerous empirical evidence demonstrating that the early recognition of acoustic signal is influenced by acoustic parameters. This paper reported a study investigating whether different fundamental frequency (F0 mean) with the same emotion would influence the ERP components of N100, P200 and N300. In the experiment, in order to help the superposition of ERP waveforms, all stimuli signals were preprocessed by the endpoint detection method to synchronize the ERP starting position with its corresponding signal content. To ensure the naturalisticness and effectiveness of the emotional speech, the emotional sentences was excerpted from Chinese radio plays as stimuli, in which three types of emotional speech (sad, happy, and neutral) were divided into two groups by F0 mean. The findings indicated that the amplitudes of the N100 components were associated with F0 mean, while the amplitudes of the P200 and N300 components were not completely influenced by F0 mean. In addition, there were no differences of ERP amplitudes between left and right hemisphere.