Data collected for A review on EEG-based music datasets

Description

This narrative review has gathered and contrasted the distinct attributes of the openly accessible EEG datasets they are commonly used to develop (affective) Brain-computer Music related algorithms. We explored the strengths and drawbacks inherent in the attributes of the EEG datasets. Drawing from our investigation, we pinpointed 19 characteristics that make the music-based EEG datasets unique. We also briefly looked into how specific characteristics of the datasets that are openly accessible impact the performance and results of a study and influence the selection of machine learning techniques and preprocessing steps required to develop methods to build Brain-computer music interface algorithms. In conclusion, this study provides a guideline for choosing publicly available music-based EEG datasets for musicians and scientists working to develop reproducible, generalizable, and effective brain-computer music algorithms.

Steps to reproduce

he identification of relevant scientific articles using electroencephalogram (EEG) datasets related to music was conducted in two phases, following the PRISMA criteria for systematic review and meta-analysis. A comprehensive search was performed across several academic databases, including Springer, IEEE, Scopus, ScienceDirect, ACM, DBLP, CiteSeerX, and PubMed, using the keyword "Music stimuli AND Electroencephalogram." The search was limited to papers published in the last eleven years with at least six citations to ensure relevance. To minimize bias and search errors, the literature search was conducted by experts in both music and EEG analysis. After filtering out duplicate studies and applying exclusion criteria, 161 relevant studies remained. Studies that did not use auditory stimuli, offer open datasets, provide stimuli files or algorithms, or investigate EEG data were excluded. Additionally, public EEG datasets related to music stimuli were searched using platforms like Google Dataset Search, Kaggle, and OpenNeuro. The final dataset only includes publicly available EEG datasets elicited by music stimuli.

Institutions

Categories

Funding

Licence