Accurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learning
<p>This study focuses on the development of an objective, automated method to extract clinically useful information from sustained vowel phonations in the context of Parkinson’s disease (PD). The aim is twofold: (a) differentiate PD subjects from healthy controls, and (b) replicate the Unified...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2012
|
| Subjects: |
| _version_ | 1797059864925569024 |
|---|---|
| author | Tsanas, A |
| author2 | Little, M |
| author_facet | Little, M Tsanas, A |
| author_sort | Tsanas, A |
| collection | OXFORD |
| description | <p>This study focuses on the development of an objective, automated method to extract clinically useful information from sustained vowel phonations in the context of Parkinson’s disease (PD). The aim is twofold: (a) differentiate PD subjects from healthy controls, and (b) replicate the Unified Parkinson’s Disease Rating Scale (UPDRS) metric which provides a <em>clinical impression</em> of PD symptom severity. This metric spans the range 0 to 176, where 0 denotes a healthy person and 176 total disability. Currently, UPDRS assessment requires the physical presence of the subject in the clinic, is <em>subjective</em> relying on the clinical rater’s expertise, and logistically costly for national health systems. Hence, the practical frequency of symptom tracking is typically confined to once every several months, hindering recruitment for large-scale clinical trials and under-representing the true time scale of PD fluctuations.</p><p>We develop a comprehensive framework to analyze speech signals by: (1) extracting novel, distinctive signal <em>features</em>, (2) using robust <em>feature selection</em> techniques to obtain a parsimonious subset of those features, and (3a) differentiating PD subjects from healthy controls, or (3b) determining UPDRS using powerful <em>statistical machine learning</em> tools. Towards this aim, we also investigate 10 existing fundamental frequency (F_0) estimation algorithms to determine the most useful algorithm for this application, and propose a novel ensemble F_0 estimation algorithm which leads to a 10% improvement in accuracy over the best individual approach. Moreover, we propose novel feature selection schemes which are shown to be very competitive against widely-used schemes which are more complex. We demonstrate that we can successfully differentiate PD subjects from healthy controls with 98.5% overall accuracy, and also provide <em>rapid</em>, <em>objective</em>, and <em>remote</em> replication of UPDRS assessment with clinically useful accuracy (approximately 2 UPDRS points from the clinicians’ estimates), using only simple, self-administered, and non-invasive speech tests.</p><p>The findings of this study strongly support the use of speech signal analysis as an objective basis for practical clinical decision support tools in the context of PD assessment.</p> |
| first_indexed | 2024-03-06T20:10:06Z |
| format | Thesis |
| id | oxford-uuid:2a43b92a-9cd5-4646-8f0f-81dbe2ba9d74 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-06T20:10:06Z |
| publishDate | 2012 |
| record_format | dspace |
| spelling | oxford-uuid:2a43b92a-9cd5-4646-8f0f-81dbe2ba9d742022-03-26T12:24:01ZAccurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learningThesishttp://purl.org/coar/resource_type/c_db06uuid:2a43b92a-9cd5-4646-8f0f-81dbe2ba9d74Biomedical engineeringArtificial IntelligencePattern recognition (statistics)Mathematical biologyBioinformatics (technology)Mathematical modeling (engineering)Machine learningSignal processingEnglishOxford University Research Archive - Valet2012Tsanas, ALittle, MMcSharry, PHowell, P<p>This study focuses on the development of an objective, automated method to extract clinically useful information from sustained vowel phonations in the context of Parkinson’s disease (PD). The aim is twofold: (a) differentiate PD subjects from healthy controls, and (b) replicate the Unified Parkinson’s Disease Rating Scale (UPDRS) metric which provides a <em>clinical impression</em> of PD symptom severity. This metric spans the range 0 to 176, where 0 denotes a healthy person and 176 total disability. Currently, UPDRS assessment requires the physical presence of the subject in the clinic, is <em>subjective</em> relying on the clinical rater’s expertise, and logistically costly for national health systems. Hence, the practical frequency of symptom tracking is typically confined to once every several months, hindering recruitment for large-scale clinical trials and under-representing the true time scale of PD fluctuations.</p><p>We develop a comprehensive framework to analyze speech signals by: (1) extracting novel, distinctive signal <em>features</em>, (2) using robust <em>feature selection</em> techniques to obtain a parsimonious subset of those features, and (3a) differentiating PD subjects from healthy controls, or (3b) determining UPDRS using powerful <em>statistical machine learning</em> tools. Towards this aim, we also investigate 10 existing fundamental frequency (F_0) estimation algorithms to determine the most useful algorithm for this application, and propose a novel ensemble F_0 estimation algorithm which leads to a 10% improvement in accuracy over the best individual approach. Moreover, we propose novel feature selection schemes which are shown to be very competitive against widely-used schemes which are more complex. We demonstrate that we can successfully differentiate PD subjects from healthy controls with 98.5% overall accuracy, and also provide <em>rapid</em>, <em>objective</em>, and <em>remote</em> replication of UPDRS assessment with clinically useful accuracy (approximately 2 UPDRS points from the clinicians’ estimates), using only simple, self-administered, and non-invasive speech tests.</p><p>The findings of this study strongly support the use of speech signal analysis as an objective basis for practical clinical decision support tools in the context of PD assessment.</p> |
| spellingShingle | Biomedical engineering Artificial Intelligence Pattern recognition (statistics) Mathematical biology Bioinformatics (technology) Mathematical modeling (engineering) Machine learning Signal processing Tsanas, A Accurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learning |
| title | Accurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learning |
| title_full | Accurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learning |
| title_fullStr | Accurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learning |
| title_full_unstemmed | Accurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learning |
| title_short | Accurate telemonitoring of Parkinson’s disease symptom severity using nonlinear speech signal processing and statistical machine learning |
| title_sort | accurate telemonitoring of parkinson s disease symptom severity using nonlinear speech signal processing and statistical machine learning |
| topic | Biomedical engineering Artificial Intelligence Pattern recognition (statistics) Mathematical biology Bioinformatics (technology) Mathematical modeling (engineering) Machine learning Signal processing |
| work_keys_str_mv | AT tsanasa accuratetelemonitoringofparkinsonsdiseasesymptomseverityusingnonlinearspeechsignalprocessingandstatisticalmachinelearning |