Project/Area Number |
08458288
|
Research Category |
Grant-in-Aid for Scientific Research (B)
|
Allocation Type | Single-year Grants |
Section | 一般 |
Research Field |
Biomedical engineering/Biological material science
|
Research Institution | The University of Tokushima |
Principal Investigator |
KINOUCHI Yohsuke Faculty of Eng., The University of Tokushima Professor, 工学部, 教授 (80035807)
|
Co-Investigator(Kenkyū-buntansha) |
SHICHIJO Fumio School of Medicine, The University of Tokushima Assistant Professor, 医学部, 講師 (20145022)
NAGASHINO Hirofumi Faculty of Eng., The University of Tokushima Assistant Professor, 工学部, 講師 (40035655)
|
Project Period (FY) |
1996 – 1997
|
Project Status |
Completed (Fiscal Year 1997)
|
Budget Amount *help |
¥700,000 (Direct Cost: ¥700,000)
Fiscal Year 1997: ¥700,000 (Direct Cost: ¥700,000)
|
Keywords | EEG / Brain source / Neural network / Inverse problem / Epilepsy / Source localization / 脳波 / EEG |
Research Abstract |
Source localization in the brain from EEG or EMG is a kind of inverse problem. It is very important and useful to solve the inverse problem for brain diagnoses and analyzes. Particularly, the recent increase of aged patients needs a real time brain diagnosis system at bed side. Biofeedback system to analyze brain functions also requires real time localization. The objective of this reserach is to realize high speed and high accuracy localization of brain sources from EEG.BP neural networks are used here for this purpose. The following subjects are investigated here. 1. Learning method and optium structure of neural networks. 2. Relation between localization accuracy and electrode arrangement. 3. Relation between localization accuracy and brain models. 4. Single and double source localization. 5. Applications to clinical data, e.g., epilepsy. As a result, brain sources are localized with high accuracy of 2 to 4 % and with short time within 5 ms. Therefore, the developed system has been very useful for high speed and high accuracy localization of brain sources from EEG.
|