Distinguishing acute encephalopathy by acute phase using machine lerning and self-organizing map

2021 Fiscal Year Final Research Report

• Project/Area Number: 19K17362
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 52050:Embryonic medicine and pediatrics-related
• Research Institution: Kagawa Prefectural College of Health Sciences
• Principal Investigator: Masayoshi Oguri 香川県立保健医療大学, 保健医療学部, 講師 (70791078)
• Project Period (FY): 2019-04-01 – 2022-03-31
• Keywords: 急性脳症 / 小児 / 脳波解析 / power spectrum / phase lag index / 自己組織化マップ

Outline of Final Research Achievements

This study aimed at developing a test method that can quantitatively analyze brain waves in bipolar induction of AESD and FS and distinguish individual patients in the early stage of onset. The results conveyed that AESD showed a significantly lower value than FS due to alpha and fast wave components in EEG analysis. Further, the functional connectivity of the brain was increased. In addition, 16 out of 20 AESD cases could be diagnosed from the AESD diagnosis map using the self-organizing map.
By performing EEG analysis for AESD within 48 hours of onset, it is possible that U-fiber disorders may be detected at an early stage before consulting imaging results. Furthermore, it is a distinguishing point from FS, which shows slow waves after status epilepticus.

Free Research Field

小児神経

Academic Significance and Societal Importance of the Research Achievements

本研究により、発症48時間以内の脳波解析を用いることでAESDとFSの客観的な鑑別点を明らかにすることができた。また、脳波解析結果を用いることで、発症後急性期におけるAESDとFSの鑑別を行える可能性が考えられた。また、自己組織化マップを用いることで、AESDとFSの自動鑑別診断の可能性を示すことができた。
今後、症例数を増やして自動診断マップの有用性を確認することで、臨床現場で用いられている脳波計に組み込めるソフトを開発できる。ソフトが完成し、臨床の脳波計に組み込むことが可能となれば、発症急性期にAESDとFSの鑑別診断補助システムとして稼働でき、急性脳症の早期鑑別が可能となる。