EEG decoding method based on a non-negative CP decomposition model
US11937934B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Nov 27, 2020 |
| Grant date | Mar 26, 2024 |
| Priority date | — |
| Expiry date | Jan 4, 2043 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06N20/10
- WIPO fieldMedical technology
- WIPO sectorInstruments
Abstract
This disclosure provides an EEG decoding method based on a non-negative CP decomposition model. The method extracts time component characteristics of the EEG of the different subjects in the boundary avoidance task, optimizes a characteristic dimension by using a 2-DPCA, and takes classification by using a support vector machine, so that differences of the EEG of subjects in different states can be reflected, and the EEG classification of the single subject has a great accuracy. The time component characteristics of the EEG can be obtained by using the channel components and the frequency components based on the non-negative CP decomposition model and by means of the interaction between the EEG modes. The characteristics of the obtained EEG time components have good separability, and the dimensions of the characteristics are optimized, so that the EEG of left and right hand movements in the boundary avoidance tasks can be effectively decoded.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.