Discrimination of bilateral finger photoplethysmogram responses to reactive hyperemia in diabetic and healthy subjects using a differential vascular model framework
| Autor: | Haleh Aghajani, Adib Keikhosravi, Edmond Zahedi |
|---|---|
| Rok vydání: | 2013 |
| Předmět: |
Male
medicine.medical_specialty Physiology Biomedical Engineering Biophysics Hyperemia Clinical settings Fingers Naive Bayes classifier Physiology (medical) Internal medicine Photoplethysmogram medicine.artery Diabetes Mellitus medicine Humans Endothelial dysfunction Brachial artery Photoplethysmography Reactive hyperemia Aged business.industry Healthy subjects Bayes Theorem Middle Aged medicine.disease Limb ischemia Surgery Vasodilation Cardiology Female Endothelium Vascular business |
| Zdroj: | Physiological Measurement. 34:513-525 |
| ISSN: | 1361-6579 0967-3334 |
| DOI: | 10.1088/0967-3334/34/5/513 |
| Popis: | Endothelial dysfunction assessment has received considerable attention due to its potential in early screening of cardiovascular diseases. Since the seminal work by Celermajer in flow-mediated dilation (FMD) based on B-mode ultrasound measurement of the brachial artery dilation following limb ischemia, many attempts have been made toward applying this method to clinical, non-invasive endothelial dysfunction assessment. One major obstacle toward achieving this objective has been the relative high cost of the required setup and skilled manpower. Such limitations have prompted the investigation of other non-invasively accessible signals such as the photoplethysmogram (PPG) in relation to FMD. It is in the above context that this paper proposes to use a modified version of an existing differential model of the human upper vasculature in order to discriminate between healthy and diabetic subjects. PPG from 46 subjects (23 healthy and 23 diabetic) were utilized to identify the model parameters. Once the model parameters were identified, singular value decomposition was applied to reduce the number of features and increase the separability. Finally, a naive Bayes classifier resulted in an overall accuracy of 93.5% (Spec. 87.0% and Sens. 100%). Taking into account subjects' gender further improved the overall accuracy. It is thought that the application of the proposed method to endothelial dysfunction assessment may positively impact the deployment of FMD in clinical settings. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|