Use of EEG-Based Machine Learning to Predict Music-Related Brain Activity
Location
Room 232, Schewel Hall
Access Type
Open Access
Entry Number
86
Start Date
4-5-2023 11:15 AM
End Date
4-5-2023 11:30 AM
College
Lynchburg College of Arts and Sciences
Department
Biomedical Science
Keywords
EEG, Brainwaves, Machine Learning, Prediction, Music, Melody, Therapy, Thought
Abstract
Music has many awe-inspiring characteristics. Some may refer to it as a “universal language” with the ability to transcend the barriers of speech, while others may describe its ability to evoke intense emotional experiences for the listener. Regardless of the description, it is a commonly held view that music can have many profound effects. Studies of music’s effects have found these beliefs to be more than pure conjecture, finding that music interacts with and changes our brains, ranging from physical changes to emotional healing.
Music can even have clinical applications, such as music therapy. This type of therapy has been shown to be beneficial in many areas, ranging from stroke rehabilitation to mental health treatment. The mechanisms behind music’s therapeutic benefit has to do with neuroplastic effects; Being able to harness this benefit in a therapeutic setting could make treatments for mental disorders and brain injuries even more effective.
This thesis aimed to discover whether musical thoughts could be interpreted by a program, potentially opening the door to the use of thought-based musical training for therapeutic benefit. For this study, EEG data was collected while people were thinking of 5 melodies, then machine learning models were trained on labeled datasets of subjects thinking of the 5 melodies. The models were then tasked with categorizing unlabeled sets of EEG data - in other words, predicting which melody a subject was thinking of while the data was being recorded. The accuracy of the predictions ranged from 45% to 80%, which means that the programs were 2-4 times better than random guessing. This shows that these programs could potentially be used to help promote neuroplasticity via musical thinking.
While this topic is still exploratory and requires more research, the results could lead to a promising future of development of music-based brain-computer interfaces.
Faculty Mentor(s)
Dr. David Freier Dr. Price Blair Dr. Cynthia Ramsey
Rights Statement
The right to download or print any portion of this material is granted by the copyright owner only for personal or educational use. The author/creator retains all proprietary rights, including copyright ownership. Any editing, other reproduction or other use of this material by any means requires the express written permission of the copyright owner. Except as provided above, or for any other use that is allowed by fair use (Title 17, §107 U.S.C.), you may not reproduce, republish, post, transmit or distribute any material from this web site in any physical or digital form without the permission of the copyright owner of the material.
Use of EEG-Based Machine Learning to Predict Music-Related Brain Activity
Room 232, Schewel Hall
Music has many awe-inspiring characteristics. Some may refer to it as a “universal language” with the ability to transcend the barriers of speech, while others may describe its ability to evoke intense emotional experiences for the listener. Regardless of the description, it is a commonly held view that music can have many profound effects. Studies of music’s effects have found these beliefs to be more than pure conjecture, finding that music interacts with and changes our brains, ranging from physical changes to emotional healing.
Music can even have clinical applications, such as music therapy. This type of therapy has been shown to be beneficial in many areas, ranging from stroke rehabilitation to mental health treatment. The mechanisms behind music’s therapeutic benefit has to do with neuroplastic effects; Being able to harness this benefit in a therapeutic setting could make treatments for mental disorders and brain injuries even more effective.
This thesis aimed to discover whether musical thoughts could be interpreted by a program, potentially opening the door to the use of thought-based musical training for therapeutic benefit. For this study, EEG data was collected while people were thinking of 5 melodies, then machine learning models were trained on labeled datasets of subjects thinking of the 5 melodies. The models were then tasked with categorizing unlabeled sets of EEG data - in other words, predicting which melody a subject was thinking of while the data was being recorded. The accuracy of the predictions ranged from 45% to 80%, which means that the programs were 2-4 times better than random guessing. This shows that these programs could potentially be used to help promote neuroplasticity via musical thinking.
While this topic is still exploratory and requires more research, the results could lead to a promising future of development of music-based brain-computer interfaces.