Neuroacoustic Patterns: Constant Q Cepstral Coefficients for the Classification of Neurodegenerative Disorders
Keywords: Neurodegenerative Disorder, Constant Q Cepstral Coefficient, Form Invariance, Random Forest, SVM.
Abstract: Early identification of neurodegenerative diseases is crucial for effective diagnosis in neurological disorders. However, the quasi-periodic nature of vocal tract sampling often results in inadequate spectral resolution in traditional spectral features, such as Mel Frequency Cepstral Coefficients (MFCC), thereby limiting their classification effectiveness. In this study, we propose the use of Constant Q Cepstral Coefficients (CQCC), which leverage geometrically spaced frequency bins to provide superior spectrotemporal resolution, particularly for capturing the fundamental frequency and its harmonics in speech signals associated with neurodegenerative disorders. Our results demonstrate that CQCC, when integrated with Random Forest and Support Vector Machine classifiers, significantly outperform MFCC, achieving absolute improvements of 5.6 % and 7.7 %, respectively. Furthermore, CQCC show enhanced performance over traditional acoustic measures, such as Jitter, Shimmer, and Teager Energy. The effectiveness of CQCC is underpinned by the form-invariance property of the Constant Q Transform (CQT), which ensures consistent feature representation across varying pitch and tonal conditions, thereby enhancing classification robustness. Furthermore, the robustness of CQCC features against MFCC features are validated using LDA plots. These findings are validated using the Italian Parkinson’s database and the Minsk2019 database of Amyotrophic Lateral Sclerosis, underscoring the potential of CQCC to advance the classification of neurodegenerative disorders.
Primary Area: applications to neuroscience & cognitive science
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 14296
Loading