A Review on Automatic Detection Methods of Sleep Spindles

Journal article

Authors/Editors

Strategic Research Themes

Publication Details

Author list: Shinanang Promkaew;Pornwanut Kitudom;Isaree Suwansombat;Thamonwan Thongbun;Kaewklao Thavornwattana;Santitham Prom-on;Woranich Hinthong

Publication year: 2023

Volume number: 5

Issue number: 1

Start page: 13

End page: 26

Number of pages: 14

ISSN: 3027-7418

eISSN:  2697-5203

URL: https://he02.tci-thaijo.org/index.php/jcra/article/view/255126

Abstract

Sleep spindle is a pattern of brainwave during non-rapid eye movement (NREM) in the second stage of sleep. There are several indications that Sleep spindles might play an important role in memory consolidation, and neurodegenerative disorders such as Alzheimer’s disease andinsomnia. Sleep spindle is commonly annotated by visual inspection of experts which is time consuming, and the task has risk of high error due to over-reliance on the experts’ skill or variations in characteristics of Sleep spindle. Modern research aims to study machine learning to develop the efficient automatic Sleep spindle detectors, emulate human annotations, and solve the stated problems. However, there are few review articles on the subject. This article summarizes and compares the automatic detection methods of Sleep spindle. The workflow can be summarized into five steps: data collection, data preprocessing, feature extraction, modeling, and model evaluation. This article reveals the variation of Machine Learning application and study on supervised model can detect Sleep spindle equivalent to experts. Nevertheless, the model need customization to suit data diversity. In addition, the discussion includes recommendation and possibility for further studies.

Keywords

การตรวจวัดคลื่นไฟฟ้าสมอง, การประมวลผลสัญญาณ, การเรียนรู้ของเครื่อง, คลื่นสมอง Sleep spindle, วิธีการตรวจจับแบบอัตโนมัติ