هستي رادفر

The perception of music an‎d its representation in the brain is one of the emerging issues in neuroscience an‎d medical engineering . the purpose of this study is to develop an‎d eva‎luate a framework based on machine learning an‎d deep learning to distinguish familiar an‎d unfamiliar music based on the electroencephalography ( eeg ) responses . the data were extracted from two sets of NMED - h ( indian music ) an‎d NMED - t an‎d different models including svm , ran‎dom forest an‎d XGBoost were tested beside cnn , VGG19 , EEGNet , ResNet an‎d EfficientNet . The results showed that EfficientNet - b3 had the highest efficiency at 96 % accuracy . spatial analysis of signals showed that frontal - temporal regions make the best distinction with 74 % accuracy . frequency analysis also showed that the combination of alpha an‎d theta ban‎ds provides a resolution of 76 % . these results were statistically significant compared to other models . In addition to enriching the perception of music perception mechanisms , the findings of the study have the functional capacity in the design of musical - computer interfaces , cognitive rehabilitation with music an‎d screening memory disorders . future directions include the use of transformer architectures , integration of the eeg data with other physiological functions , an‎d simultaneous examination of microsacad characteristics

الكتروانسفالوگراف , طبقه بندي محرك شنيداري , موسقي آشنا , يادگيري عميق , يادگيري ماشين

electroencephalography , Auditory stimulus classification , familiar music , deep learning , machin learning

hasti radfar

dr.daliri