Hardware architecture for classifying action potentials in brain-computer interface

Abstract

The invention discloses hardware architecture for classifying action potentials in a brain-computer interface. The architecture comprises a main action potential classification module, a neural network classification module and a control module. The main action potential classification module includes functions of action potential detection, alignment, feature value extraction, dimensionality reduction and classification and is used for completing primary classification of action potentials in original data, wherein the classification result is applied to the neural network classification module. The neural network classification module classifies the original data at the same time; and with the action potential classification result as training samples, two kinds of classification resultsare integrated by the control module to obtain a final classification result. The architecture can be applied to classification of action potentials in a brain-computer interface to realize unsupervised action potential classification with higher accuracy.

Description

technical field [0001] The invention relates to a hardware architecture for action potential classification in a brain-computer interface, which belongs to the field of hardware design, and more specifically relates to a high-accuracy unsupervised action potential classification hardware architecture in a brain-computer interface. Background technique [0002] The first research paper on brain-computer interface was published in 1973. In this paper, brain-computer interface was used for the first time to describe the direct information transmission path between the brain and the outside world, and the prototype of the system framework of brain-computer interface was proposed. The first step in the brain-computer interface is to collect and transmit brain activity in real time through various sensors. Extracellular recording technology in neuroscience is a technique in which electrodes are implanted into the extracellular tissue of neurons in the brain to record the activity ...

AI Technical Summary

Problems solved by technology

Various related algorithms have been proposed over the years, but so far are incomplete

Images