You are here

EEG-based Investigation of Music Familiarity and Emotion

EEG-based Investigation of Music Familiarity and Emotion

Nattapong Thammasan, Koichi Moriyama, Ken-ichi Fukui, Masayuki Numao
The 29th Annual Conference of the Japanese Society for Artificial Intelligence, 2015

Familiarity is a crucial factor in music engagement, but very few study focuses on its neural correlates. We investigated effects of familiarity based on electroencephalogram in music-emotion recognition. Our research focused on self-reporting and continuous annotation based on the hypothesis that emotion in music experiencing is subjective and varies over time. Our methodology allowed subject to select up to 16 MIDI songs, comprising of eight familiar songs and the remaining were unfamiliar songs. We found that prior knowledge of the music indicated by familiarity level affected brain activities reflected by our indicators in frequency domain analysis. Furthermore, using entirelyunfamiliar songs was superior to mixed data in emotion classification. Therefore, unfamiliar songs are recommended to be used in emotion recognition system construction.

For more information, please click here >>


© Copyright 1999 - 2024 ANT Neuro | | Terms of Use | Privacy Statement | Contact | USA Customers