扩展的卡尔曼滤波在近红外光谱提取脑血流信号中的研究

doi:10.3964/j.issn.1000-0593(2020)07-2048-06

Abstract
References
Related Citation (15)

Download: PDF (2812 KB)
Export: BibTeX | EndNote (RIS)

Abstract There are two types of hemoglobin in the cerebral blood stream: oxygenated hemoglobin (HbO₂) and reduced hemoglobin (HbR). The changes in the concentration of these two hemoglobins in the cerebral blood flow can reflect the neural activity in the brain. Extracting the signals of concentration changes can provide basis and reference for the diagnosis and treatment of related diseases such as epilepsy focus localization and depression. At present, algorithms for extracting cerebral blood flow signals using near-infrared spectroscopy include the EEMD-ICA method principal component analysis (PCA) , independent component analysis (ICA), the coherent averaging method, Adaptive filtering, etc. The above algorithms have their own characteristics and advantages in the extraction of near-infrared brain neural activity signals. However, the above methods all pay attention to various physiological interferences such as respiration and eye movement and ignore measurement interferences that conform to Gaussian distribution during measurements, such as instrument precision and crosstalk in signal transmission. In order to extract signals of changes in the concentration of oxygenated hemoglobin (HbO₂) and reduced hemoglobin (HbR) in cerebral blood flow, a functional near infrared spectroscopy (fNIRS) cerebral blood flow parameter acq uisition device is designed in this article. In the device, a light source Diode near-infrared light sources with wavelengths of 750 and 830 nm were selected to collect brain blood flow changes. The extended Klaman Filter (EKF) algorithm was used to establish a corresponding mathematical model of physiological interference and measurement interference. Perform recursive calculation with the minimum principle, and combine the initial state estimation of the system at the next moment with the measured feedback to obtain a state estimate of infinitely close to the real value at that moment.), The change of the optical density signal is converted into a signal of change in oxygenated hemoglobin (HbO₂) and reduced hemoglobin (HbR) concentration. The results show that the method proposed in this paper can effectively remove the measurement interference that conforms to the Gaussian distribution. In the Valsava experiment and the visual evoked experiment, the curve of changes in the concentration of oxygenated hemoglobin (HbO₂) and reduced hemoglobin (HbR) in the cerebral blood flow can be extracted. Compared with the mainstream EEMD algorithm for extracting brain signals, its RMSE value is increased by 0.96%, and r value is increased by 0.6%, which indicates that the proposed method has certain advantages. The method proposed in this paper provides an effective method for detecting neural activity in related brain diseases.

Key words：EKF algorithm; fNIRS; Valsava experiment; Visual induction; Hemoglobin

Received: 2020-01-07 Accepted: 2020-04-20

ZTFLH:

TP274

Corresponding Authors: QIU Ji-qing E-mail: songyangl15@mails.jlu.edu.cn

Cite this article:

URL:

https://www.gpxygpfx.com/EN/10.3964/j.issn.1000-0593(2020)07-2048-06 OR https://www.gpxygpfx.com/EN/Y2020/V40/I07/2048