Energy feature extraction method of composite lower limb imaginary movement EEG

Abstract

The invention belongs to the fields of biomedicalengineering and computers, relating to an energy feature extraction method of a composite lower limb imaginary movement EEG, comprising the following steps: (1) preprocessing an EEG signal; (2) electrolyzing the composite lower limb imaginary movement EEG into each natural oscillation mode component from high to low of frequency through an empirical mode decomposition method; (3) analyzing spectral distribution feature of each natural oscillation mode through a power spectral density analytical method to determine a feature oscillation mode reflecting an EEG feature rhythm; (4) carrying out Hilbert transform on the feature oscillation mode to obtain desynchronization features of the energy change; and (5) identifying mode: taking the energy change desynchronization feature of the feature oscillation mode as a classifier to input to realize the mode identification of the lower limb imaginary movement. The invention fully considers nonstationarity of the EEG signals; the highest recognition rate is 87.8% and has obvious improvement compared with 82.3% of the traditional method.

Description

technical field [0001] The invention belongs to the field of biomedical engineering and computer, and relates to an energy feature extraction method of compound lower limb imagination action EEG. Background technique [0002] Brain-computer interface (Brain-Computer Interface, BCI) is to establish a direct information exchange and control channel between the human brain and computers or other electronic devices that does not depend on conventional brain output pathways (peripheral nerves and muscle tissue). A new human-computer interaction system. The earliest EEG signals applied to the brain-computer interface system are mainly spontaneous EEG signals, such as alpha (α) waves in EEG. However, this type of EEG signal mode is single, and it is impossible to truly achieve "consciousness control action", which seriously restricts the development of brain-computer interface systems. In recent years, scholars from various countries have gradually carried out research on EEG sig...

AI Technical Summary

Problems solved by technology

[0005] But so far, the pattern extraction of lower extremity imaginary action potentials has been slow, and the accuracy of judgment is difficult to improve

The main reason is that the functional area of ​​the cerebral cortex mapped by the movement of the lower extremities is a relatively narrow area in the parietal sulcus, and the distinction of its spatial structure is already very limited. In addition, the EEG signals extracted by the scalp electrodes are extremely diffuse and inconsistent. Aliasing, very unfavorable for source signal acquisition and identification

This key factor leads to the limited applicability of the feature extraction algorithm applied to the imagery EEG of the upper limbs in the feature extraction of the EEG of the lower limbs.

Images