| 제목 |
An Artificial Neural-Network Approach for Motor Hotspot localization Based on Electroencephalography |
| 소속 |
Korea University, Departmet of Electronics and Information Engineering1, Berlin Institute of Technology, Machine Learning Group2, Seoul National University Bundang Hospital, Department of Rehabilitation Medicine3 |
| 저자 |
Ga-Young Choi1, Chang-Hee Han2, Hyung-Tak Lee1, Nam-Jong Paik3, Won-Seok Kim3*†, Han-Jeong Hwang1† |
Objective: To apply transcranial electrical stimulation to the motor cortex, motor hotspot is generally identified using motor evoked potential by transcranial magnetic stimulation (TMS). The goal of this study is to validate the feasibility of a novel electroencephalography (EEG)-based motor-hotspot-localization approach using machine learning technique as a potential alternative to TMS.
Methods : EEG data were measured while thirty subjects performed a finger tapping task (Figure 1). Power spectral densities of the EEG data were extracted, and they were used to train and test an artificial neural network for motor hotspot localization.
Results: A minimum error distance between the motor hotspot locations identified by TMS and our proposed approach was 0.26 ± 0.19 cm (Figure 2 and 3), and which did not significantly increase even using a quarter of EEG channels for the dominant hand.
Conclusion: We demonstrated the feasibility of our novel EEG-based motor-hotspot-localization method. |
|
Experimental paradigm
3D coordinate information of motor hotspots identified TMS (red) and EEG PSD features (blue) based on EEG channel locations (green).
Mean hotspot detection error distance for each frequency band
|
