Brain wave signal causality detection method based on two-dimensional autoregression model parameter estimation

Abstract

The invention discloses a brain wave signal causality detection method based on two-dimensional autoregression model parameter estimation. The method comprises the steps that (1) order estimation is performed through an autoregression model; (2) windowing processing is performed on an original signal; (3) parameter estimation is performed on the autoregression model in combination with an OPS algorithm; and (4) the obtained autoregression model is applied to a Wiener-Granger causality method to detect the causality among brain wave signals. Through the method, disturbance items in autoregression model parameter estimation are reduced to the maximum extent, and the precision of a parameter estimation result is improved.

Description

technical field [0001] The invention belongs to the field of biomedicine, and in particular relates to a method for detecting causality between electroencephalogram signals based on two-dimensional autoregressive model parameter estimation. Background technique [0002] Due to the complexity of the brain system, it is difficult to accurately model the EEG signal using mathematical methods. Generally, the dynamic equation of the EEG signal is unknown. In order to analyze the EEG signal, it is necessary to first model the EEG signal. to identify. When studying the linear relationship between EEG signals, the brain system is regarded as a linear system, and any linear system that meets certain conditions can be represented by a linear autoregressive model. The autoregressive model contains both the past output information and the past input information, and requires fewer parameter sets. Using the autoregressive model can effectively reduce the computational workload of linear...

AI Technical Summary

Problems solved by technology

[0038] Purpose of the invention: In order to solve the difficulty that the AIC method cannot separately estimate the orders of different signals in the autoregressive model, and also in order to solve the problem that the OPS algorithm is not stable enough and not accurate enough to estimate the parameters of the autoregressive model, the present invention provides a An autoregressive model parameter estimation method based on the sliding window method to detect the causal relationship between EEG signals. This method minimizes the interference items in the autoregressive model parameter estimation and improves the accuracy of the parameter estimation results.

Images