Evaluation of MEG power spectral density as a biomarker of dementia
-
- Okumura Naohiro
- Medical Imaging Business Centre, Healthcare Business Group, RICOH Company, L td., Tokyo, Japan
-
- Hoshi Hideyuki
- Medical Imaging Business Centre, Healthcare Business Group, RICOH Company, L td., Tokyo, Japan Precision Medicine Centre, Hokuto Hospital, Japan
-
- Hirata Yoko
- Department of Neurosurgery, Kumagaya General Hospital, Japan Department of Neurosurgery, Ohashi Medical Center, Toho University, Japan
-
- Kobayashi Momoko
- Precision Medicine Centre, Kumagaya General Hospital, Japan
-
- Sakamoto Yuki
- Precision Medicine Centre, Kumagaya General Hospital, Japan
-
- Fukasawa Keisuke
- Clinical Laboratory Centre, Kumagaya General Hospital, Japan
-
- Ichikawa Sayuri
- Clinical Laboratory Centre, Kumagaya General Hospital, Japan
-
- Kanzawa Takao
- Isesaki Clinic, Japan Department of Stroke, Mihara Memorial Hospital, Japan
-
- Shigihara Yoshihito
- Precision Medicine Centre, Hokuto Hospital, Japan Precision Medicine Centre, Kumagaya General Hospital, Japan
Bibliographic Information
- Other Title
-
- 脳磁計パワースペクトル密度の認知症バイオマーカーとしての性能評価
Abstract
<p>Background: Power spectral density (PSD) is often used for frequency analysis of brain signal measured by magnetoencephalography (MEG). Parameters calculated from the PSD have been studied as potential biomarkers of dementia and mild cognitive impairment (MCI). In this study, to evaluate classification performance of the PSD-derived parameters quantitatively, we compared them between healthy controls (CTR), patients with MCI, and dementia (DEM), using machine learning approach.Method: PSD was calculated from spontaneous MEG signals (139 CTR, 38 MCI, and 57 DEM) measured at the Kumagaya General Hospital. As the parameters, median frequency, individual alpha frequency, and Shannon entropy were calculated. The parameters were compared between the groups and the classification performance was evaluated using SVM classifier.Results: All parameters were significantly different between the groups. The classification accuracy was improved using SVM.Discussion: The classification performance of PSD-derived parameters were replicated. Each parameter captures distinctive features of PSD.</p>
Journal
-
- Transactions of Japanese Society for Medical and Biological Engineering
-
Transactions of Japanese Society for Medical and Biological Engineering Annual59 (Abstract), 426-426, 2021
Japanese Society for Medical and Biological Engineering
- Tweet
Details 詳細情報について
-
- CRID
- 1390852714993985920
-
- NII Article ID
- 130008105375
-
- ISSN
- 18814379
- 1347443X
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- CiNii Articles
-
- Abstract License Flag
- Disallowed