The Study Of Brain-Computer Interface Based On Local Field Potentials

Brain-Computer Interface (BCI) is a kind of technology that enables people to communicate with computer or to control devices with brain signals. The study of BCI technology is significant to the development of neuroscience and has great application value in neural rehabilitation. Extraction and recognition of features of brain signal related to special movement behavior is one of the keys of BCI technology. Temporal characters and spectral characters of local field potentials (LFPs) in rats with different forelimb behavior mode have been studied in this paper, and this is important to the improvement of BCI technology.In first the experiment, rats were trained to press the paddle for water. Self-made multi-channel microelectrodes were implanted in forelimb representation region in primary motor cortex of rats. Deep EEG signal in primary motor cortex were collected and Local field potentials were segmented according to paddle-pressed pulse. Then LFP features of pressing behavior were extracted, and then the rat's press-paddle behavior were recognized based on these features. Second, an induced-behavior experiment in freely moving rats was developed. Rats'forelimbs were induced to do some special movement such as teasing rat's forelimb with a stick, pricking rat with a need, frightening rat with claps, etc. These movement patterns were also recognized from LFP signals with pattern discrimination methods, such as template..Research indicates that:local field potentials in primary motor cortex present high amplitude and low frequency wave when rats'limbs move, and the features of these wave are also relative to the recording position. Local field potentials present negative-positive wave when rats press the paddle. The press paddle behavior could be recognized correctly to 80% from the LFP signals with threshold criterion. In induced behavior experiment, the induced forelimb movements could be exceeded to 85% with template matching method. These Results indicate that the LFPs in a small region in primary motor cortex are distinctively relative to special limb's movement, and it maybe as they reflect the actions of motor muscles which control limb movements.