CN113040789B - Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition - Google Patents

Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition Download PDF

Info

Publication number
CN113040789B
CN113040789B CN202110286597.0A CN202110286597A CN113040789B CN 113040789 B CN113040789 B CN 113040789B CN 202110286597 A CN202110286597 A CN 202110286597A CN 113040789 B CN113040789 B CN 113040789B
Authority
CN
China
Prior art keywords
signal
template
eeg
artifact
artifacts
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110286597.0A
Other languages
Chinese (zh)
Other versions
CN113040789A (en
Inventor
张剑豪
俞祝良
于晓婧
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
South China University of Technology SCUT
Original Assignee
South China University of Technology SCUT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by South China University of Technology SCUT filed Critical South China University of Technology SCUT
Priority to CN202110286597.0A priority Critical patent/CN113040789B/en
Publication of CN113040789A publication Critical patent/CN113040789A/en
Application granted granted Critical
Publication of CN113040789B publication Critical patent/CN113040789B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves 
    • A61B5/055—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves  involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A61B5/7217—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal of noise originating from a therapeutic or surgical apparatus, e.g. from a pacemaker

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Molecular Biology (AREA)
  • Veterinary Medicine (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Psychiatry (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Physiology (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Radiology & Medical Imaging (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

The invention discloses a method for synchronizing EEG-fMRI dataThe online removal method of the nuclear magnetic artifact during collection, 1) the pulse artifact of the collected EEG signal containing noise is filtered by a low-pass filter with designed parameters; 2) up-sampling the filtered signals in the step 1), synchronizing the EEG signals with the marks sent by the fMRI equipment, namely synchronizing the EEG signals with a synchronization box to obtain a signal S h (ii) a 3) The electroencephalogram signal is a random signal, the gradient noise takes a slice scanning time T as a period, and the shape of the noise in each period is similar, so that a sliding window is constructed by N slices to construct a gradient noise template; compared with the OBS denoising based on PCA, the denoising method disclosed by the invention has the advantages that the artifacts are removed, meanwhile, the effective electroencephalogram signals are better kept, and the accuracy is higher.

Description

Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition
Technical Field
The invention belongs to the technical field of brain signal processing, and particularly relates to an online removal method of nuclear magnetic artifacts in synchronous EEG-fMRI data acquisition.
Background
With the growing interest in hemodynamic measurements in combined physiologic electroencephalography and functional neuroimaging, techniques for data acquisition of synchronized EEG-fMRI have also evolved steadily. The electroencephalogram signal has the characteristics of high time resolution and low spatial resolution, and the nuclear magnetic resonance imaging has the characteristics of low time resolution and high spatial resolution, so that the combination of the electroencephalogram signal and the nuclear magnetic resonance imaging has important significance for researching the pathogenesis of the brain and mental diseases and cognition.
However, in the synchronous EEG-fMRI experiments, the acquired EEG signals are swamped by huge noise due to the presence of high intensity magnetic field transitions in the magnetic resonance machine. These noises greatly hinder the extraction of the electroencephalogram signal characteristics and also greatly influence the research on the electroencephalogram signal. The main noise artifacts affecting the brain electrical signals can be classified into the following three types: (1) impulse artifacts caused by the rf impulse signal, such artifacts having a frequency in the megahertz range that does not substantially overlap with the frequency of the EEG active signal and are therefore removed by the low pass filter; (2) the gradient artifact is caused by the change of a gradient magnetic field during scanning, has larger amplitude, occupies the main component of the artifact, and is also the artifact which needs to be considered in the artifact removing process; (3) electrocardio artifacts, artifacts brought by the motion of heartbeat and arterial blood flow under the condition of a high-intensity magnetic field.
In order to extract effective electroencephalogram signals from polluted data, an effective algorithm needs to be designed to remove gradient artifacts and electrocardio artifacts. Meanwhile, on the basis of ensuring artifact removal, the calculation amount of the algorithm cannot be too large in order to meet the real-time requirement. The currently used technology does not achieve a satisfactory result in the real-time denoising process of the online experiment.
1. The invention with the application number of CN 201310117717.X aims to simultaneously acquire electroencephalogram signals and a noise template by using a differential amplifier and filter noise by a noise processing algorithm;
2. the invention with application number CN 201520599490.1 is to shield noise signals in nuclear magnetic environment by means of a shielding case.
3. The invention with the application number of CN 201610918267.8 aims to remove gradient artifacts and power frequency interference by a two-stage self-adaptive noise cancellation method;
4. the invention with the application number of CN 201710089297 is used for off-line denoising of polluted EEG signals by a method for removing nuclear magnetic artifacts in the EEG signals based on automatic ICA;
in the existing technology for removing the nuclear magnetic artifacts, the implemented functions are only to shield noise signals through hardware or to perform offline denoising through an algorithm, and the synchronous real-time denoising through a signal denoising algorithm is not performed. The synchronous real-time denoising algorithm in the invention has the advantages of simple equipment and better effect of artifact removal. In addition, on the basis of an artifact removing algorithm, for the electrocardio artifact which is difficult to process, a self-adaptive SVD filter is provided, so that the electrocardio artifact is effectively removed, and the signal-to-noise ratio is improved.
Disclosure of Invention
The invention aims to provide a nuclear magnetic artifact removing method in synchronous EEG-fMRI data acquisition aiming at the defects of the prior art, which can realize acquisition of an fMRI image and acquisition of an electroencephalogram signal in a nuclear magnetic resonance environment, and can obtain an electroencephalogram signal with higher signal-to-noise ratio through a real-time online artifact removing algorithm. After the gradient artifacts are removed, the morphological problem of the electrical artifacts in the nuclear magnetic artifacts center is further considered, the morphological problem is used as a standard of feature classification, feature extraction and classification are carried out on the artifact components after singular value decomposition, and then the electroencephalogram signals with the electrocardio artifacts removed are obtained through regression analysis.
In order to achieve the purpose, the invention provides the following technical scheme: an online removal method of nuclear magnetic artifacts in synchronous EEG-fMRI data acquisition mainly comprises the following steps:
1) filtering pulse artifacts of the collected electroencephalogram signals containing noise by a low-pass filter with designed parameters;
2) up-sampling the filtered signals in the step 1), synchronizing the EEG signals with the marks sent by the fMRI equipment, namely synchronizing the EEG signals with a synchronization box to obtain a signal S h ；
3) Because the electroencephalogram signal is a random signal, the gradient noise takes a slice scanning time T as a period, and the shape of the noise in each period is similar, a sliding window is constructed by the length of N slices to construct a gradient noise template, and preliminary denoising is carried out on the basis of the template;
4) QRS peak detection is carried out on the electrocardiosignal ECG of the 32 nd channel, and the position RR of each R peak on the ECG is recorded i And calculating the distance between adjacent R peaks, i.e. the period T of the electrocardio-artifact i ，i＝1,2,...,n；
5) Because the ECG delays Ts on the EEG signal EEG, the R peak in the ECG is recorded and reflected on the EEG signal S e In (2) position R i 1,2, ·, n; constructing matrix D by taking position of R peak as center m*n To matrix D m*n Singular Value Decomposition (SVD) is used, and artifact components are selected according to the mean value of the right singular value and the correlation coefficient of the left singular value and the ECG template ECGtemp respectively; constructing an artifact template by using the artifact component, and subtracting the artifact template from the original signal to obtain a clean electroencephalogram signal;
6) the signal segment obtained in the step 5) is processed
Figure GDA0003714817510000033
According to the position R of the R peak i Reconstructing the electroencephalogram signal to obtain a clean signal; with R i As the center, take 0.5 × T i-1 I.e. R i And R i-1 Taking 0.5 x T as the first half period pre i As the last half-cycle post; if pre>0.75*R m Then let pre equal 0.75R m If post>0.75*R m If the post is 0.75R m (ii) a Copying the clean signal with the corresponding range centered on the R peak to (R) i -pre,R i + post) to obtain a clean electroencephalogram signal S without electrocardio artifacts c 。
Preferably, the specific steps of step 3) are as follows: 3.1) for the signal S in the sliding window h By subtracting with an average template, i.e. in accordance with
Figure GDA0003714817510000031
Obtaining a gradient noise template A τ Wherein
Figure GDA0003714817510000032
A slice data segment in the window; selecting a section data segment in the window, then obtaining a fitting parameter by least square fitting with the gradient noise template, and obtaining a fitting parameter from S h Subtracting the product of the fitting parameter and the gradient noise template to obtain a signal S with residual artifacts l ；
3.2) for the signal S l Decomposing the sample matrix M constructed according to the period T by using principal component analysis, selecting the maximum T eigenvalues, and constructing an optimal basis by using the corresponding eigenvectors; slicing the signal to be processed with the optimal basis
Figure GDA0003714817510000049
Performing a least square to obtain a fitting coefficient, i.e.
Figure GDA0003714817510000041
Slave signal
Figure GDA0003714817510000042
Subtracting the product of the fitting coefficient and the optimal basis to obtain the electroencephalogram signal S without the gradient artifacts e 。
Preferably, the specific steps of step 5) are as follows: 5.1) for each EEG, take T i Median value of R m For the location RR of the ECG i Taking RR i ±0.75*R m The signal segments of i ═ 1, 2., n are superposed and averaged to obtain an electrocardiogram template ECGtemp;
5.2) for each R i Taking R i ±0.75*R m Signal segment of 1,2, 1
Figure GDA0003714817510000043
N, constructing a sample matrix D m*n Wherein m is the dimension of one sample, i.e. 1.5R m N is the number of samples;
5.3) matrix of pairs of samples D m*n The singular value decomposition SVD is performed,
namely, it is
Figure GDA0003714817510000044
Wherein the left singular value vector U m*m Each column of (a) represents the time domain feature of each component, the vector of right singular values V n*n Each column of (a) represents a magnitude distribution corresponding to a time domain feature; sorting the components according to the absolute value of the vector mean value of the right singular value, and selecting the largest f components; then calculating the correlation coefficient of the left singular value and the ECG template ECGtemp, sorting the components according to the magnitude of the correlation degree, and taking the largest f components;
5.4) combining the results of the two sequencing to obtain t components, and taking the left singular value of the t components as the optimal base theta of the electrocardio artifact j J 1.. n, which is fit with least squares
Figure GDA0003714817510000045
j 1.. n, obtaining fitting parameters
Figure GDA0003714817510000046
From
Figure GDA0003714817510000047
Subtracting the artifact template theta j *b j Obtaining signals without electrocardio-artifacts
Figure GDA0003714817510000048
Preferably, the low-pass filter parameter in step 1) is controlled at 1000-1250 Hz.
Compared with the prior art, the invention has the beneficial effects that: the artifact removing algorithm in the invention has small calculation amount and meets the requirement of data processing speed of online experiments.
According to the method, self-adaptive SVD denoising is used, the electrocardio artifact is further removed on the basis of removing the gradient artifact, and the influence of the electrocardio artifact on data is reduced to a great extent.
The self-adaptive SVD denoising method takes morphological characteristics of the electrocardio artifact into consideration, and improves the signal-to-noise ratio of the collected signals better.
Compared with the method for removing the electrocardio artifact by using the OBS based on the PCA, the denoising method disclosed by the invention has the advantages that the artifact is removed, the effective electroencephalogram signal is better kept, and the accuracy is higher.
Drawings
FIG. 1 is a flow chart of a method of cardiac artifact removal;
FIG. 2 is an ECG template constructed in an embodiment of the invention;
FIG. 3 is a diagram of the mean distribution of the right singular value vectors and the distribution of the correlation coefficients of the left singular value vectors and the ECG template according to an embodiment of the present invention;
FIG. 4 shows 4 selected components in an embodiment of the present invention;
FIG. 5 is a P300 waveform of P3 lead and P4 lead obtained by denoising according to the method of the present invention in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
The present invention will be further described with reference to the following specific examples and drawings, but the embodiments of the present invention are not limited thereto.
The method for removing the nuclear magnetic artifact in the synchronous EEG-fMRI data acquisition provided by the embodiment can be applied to fMRI and EEG mixed brain-computer interface platforms, and comprises the following steps:
1) designing a program framework, wherein a program mainly comprises two threads: a data acquisition thread and a data processing thread; in order to realize parallel processing of each channel data, a thread needs to be created for each de-noised channel;
2) and filtering out pulse artifacts by low-pass filtering of 1250Hz on the acquired electroencephalogram signals containing noise.
3) Up-sampling the filtered signal in the step 2), synchronizing the EEG signal with the mark sent by the fMRI device, namely, synchronizing the EEG signal with a synchronization box to obtain a signal S h ；
4) Because the electroencephalogram signal is a random signal, the gradient noise takes a slice scanning time T as a period, and the shape of the noise in each period is similar, a sliding window is constructed by the length of N slices to construct a gradient noise template, and preliminary denoising is carried out based on the template, and the method specifically comprises the following steps:
4.1) for the signal S in the sliding window h By subtraction using an average template, i.e. in accordance with
Figure GDA0003714817510000061
Obtaining a gradient noise template A τ Wherein
Figure GDA0003714817510000062
In a windowA slice data segment of (a); selecting a section data segment in the window, then obtaining a fitting parameter by least square fitting with the gradient noise template, and obtaining a fitting parameter S h Subtracting the product of the fitting parameter and the gradient noise template to obtain a signal S with residual artifacts l ；
4.2) for the signal S l Decomposing the sample matrix M constructed according to the period T by using principal component analysis, selecting the maximum T eigenvalues, and constructing an optimal basis by using the corresponding eigenvectors; slicing the signal to be processed with the optimal basis
Figure GDA0003714817510000063
Performing a least square to obtain a fitting coefficient, i.e.
Figure GDA0003714817510000064
Slave signal
Figure GDA0003714817510000065
Subtracting the product of the fitting coefficient and the optimal basis to obtain the electroencephalogram signal S without the gradient artifacts e ；
5) QRS peak detection is carried out on the electrocardiosignal ECG of the 32 nd channel, and the position RR of each R peak on the ECG is recorded i And calculating the distance between adjacent R peaks, i.e. the period T of the electrocardio-artifact i ，i＝1,2,...,n；
6) Because the ECG is delayed on the EEG signal by Ts (standard delay of 210ms), the R peak in the ECG is recorded and reflected on the EEG signal S e Position R in (1) i 1,2, ·, n; constructing matrix D by taking position of R peak as center m*n To matrix D m*n Singular Value Decomposition (SVD) is used, and artifact components are selected according to the mean value of the right singular value and the correlation coefficient of the left singular value and the ECG template ECGtemp respectively; constructing an artifact template by using the artifact component, and subtracting the artifact template from the original signal to obtain a clean electroencephalogram signal; the method comprises the following specific steps:
6.1) for each EEG, take T i Median value of R m For the location RR of the ECG i Taking RR i ±0.75*R m The signal segments of i ═ 1, 2., n are superposed and averaged to obtain an electrocardiogram template ECGtemp, and the waveform of the ECGtemp can be shown in fig. 2;
6.2) for each R i Taking R i ±0.75*R m Signal section of 1,2
Figure GDA0003714817510000071
N, constructing a sample matrix D m*n Wherein m is the dimension of one sample, namely 1.5R m N is the number of samples;
6.3) Pair sample matrix D m*n The singular value decomposition SVD is performed,
namely, it is
Figure GDA0003714817510000072
Wherein the left singular value vector U m*m Each column of (a) represents the time domain feature of each component, the vector of right singular values V n*n Each column of (a) represents a magnitude distribution corresponding to a time domain feature; sorting the components according to the absolute value of the vector mean value of the right singular value, and selecting the largest f components; then calculating the correlation coefficient of the left singular value and the ECG template ECGtemp, sorting the components according to the magnitude of the correlation degree, and taking the largest f components, wherein the specific distribution condition can be shown in FIG. 3;
6.4) combining the results of the two sequencing to obtain t components, and taking the left singular value of the t components as the optimal base theta of the electrocardio artifact j J 1.. n, which is fit with least squares
Figure GDA0003714817510000073
j 1.. n, obtaining fitting parameters
Figure GDA0003714817510000074
From
Figure GDA0003714817510000075
Subtracting the artifact template theta j *b j Obtaining signals without electrocardio-artifacts
Figure GDA0003714817510000076
7) The signal segment obtained in the step 6) is processed
Figure GDA0003714817510000077
According to the position R of the R peak i And reconstructing the electroencephalogram signal to obtain a clean signal. With R i As center, take 0.5 × T i-1 I.e. R i And R i-1 Taking 0.5 x T as the first half period pre i As the last half-cycle post; if pre>0.75*R m Then let pre be 0.75R m If post>0.75*R m Then let post be 0.75R m . Copying the clean signal with the corresponding range centered on the R peak to (R) i -pre,R i + post) to obtain a clean electroencephalogram signal S without the electrocardio-artifacts c ；
On-line experiments are designed according to the method and the PCA-based OBS denoising method, and experimental results are compared.
In a nuclear magnetic environment, the method and the OBS denoising algorithm are respectively applied to carry out experiments through a P300 character spelling paradigm. When constructing the optimal basis, the feature vectors of the first 4 components are selected as the optimal basis calculation artifact template of the method of the invention, and the selected 4 components are shown in fig. 4. In the experiment, 20 groups of data are collected as a training set to be trained to obtain a classifier, the classifier is used for testing 30 characters on line to obtain the testing accuracy, and the average value of multiple groups of accuracy is obtained. And the overlapped P300 signal processed by the denoising method is drawn through Matlab simulation, and it can be seen that the P300 signal in the data processed by the method is more obvious, and specifically shown in FIG. 2. The results of the experiment are shown in table 1.
Method Average accuracy
OBS 75.34％
The method of the invention 95.34％
TABLE 1
As can be seen from Table 1, compared with the conventional OBS, the method provided by the invention can remove artifacts and simultaneously reserve complete electroencephalogram signals to the greatest extent, so that the identification accuracy is higher.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (2)

1. An online removing method of nuclear magnetic artifacts in synchronous EEG-fMRI data acquisition is characterized by comprising the following steps: the method mainly comprises the following steps:
1) filtering pulse artifacts of the collected electroencephalogram signals containing noise by a low-pass filter with designed parameters;
2) up-sampling the filtered signals in the step 1), synchronizing the EEG signals with marks sent by fMRI equipment, namely, synchronizing the EEG signals with a synchronization box to obtain a signal S h ；
3) Because the electroencephalogram signal is a random signal, the gradient noise takes a slice scanning time T as a period, and the shape of the noise in each period is similar, a sliding window is constructed by the length of N slices to construct a gradient noise template, and preliminary denoising is carried out on the basis of the template; the method comprises the following specific steps: 3.1) for the signal S in the sliding window h By subtraction using an average template, i.e. in accordance with
Figure FDA0003714817500000011
Obtaining a gradient noise template A τ Wherein
Figure FDA0003714817500000012
A slice data segment in the window; selecting a section data segment in the window, then obtaining a fitting parameter by least square fitting with the gradient noise template, and obtaining a fitting parameter from S h Subtracting the product of the fitting parameter and the gradient noise template to obtain a signal S with residual artifacts l ；
3.2) for the signal S l Decomposing the sample matrix M constructed according to the period T by using principal component analysis, selecting the maximum T eigenvalues, and constructing an optimal basis by using the corresponding eigenvectors; slicing the signal to be processed with the optimal basis
Figure FDA0003714817500000013
Performing a least squares to obtain a fitting coefficient, i.e.
Figure FDA0003714817500000014
Slave signal
Figure FDA0003714817500000015
Subtracting the product of the fitting coefficient and the optimal base to obtain the electroencephalogram signal S without the gradient artifacts e ；
4) QRS peak detection is carried out on the electrocardiosignal ECG of the 32 nd channel, and the position RR of each R peak on the ECG is recorded i And calculating the distance between adjacent R peaks, i.e. the period T of the electrocardio-artifact i ，i＝1,2,...,n；
5) Because the ECG delays Ts on the EEG signal EEG, the R peak in the ECG is recorded and reflected on the EEG signal S e In (2) position R i 1,2, ·, n; constructing matrix D by taking position of R peak as center m*n To matrix D m*n Singular Value Decomposition (SVD) is used, and artifact components are selected according to the mean value of the right singular value and the correlation coefficient of the left singular value and the ECG template ECGtemp respectively; construction of an artifact template using artifact components, followed by subtraction using the original signalObtaining a clean electroencephalogram signal after the artifact template;
the method comprises the following specific steps: 5.1) for each EEG, take T i Median value of R m For the location RR of the ECG i Taking RR i ±0.75*R m The signal segments of i ═ 1, 2., n are superposed and averaged to obtain an electrocardiogram template ECGtemp;
5.2) for each R i Taking R i ±0.75*R m Signal segment of 1,2, 1
Figure FDA0003714817500000027
Constructing a sample matrix D m*n Wherein m is the dimension of one sample, i.e. 1.5R m N is the number of samples;
5.3) matrix of pairs of samples D m*n The singular value decomposition SVD is performed,
namely, it is
Figure FDA0003714817500000021
Wherein the left singular value vector U m*m Each column of (a) represents the time domain feature of each component, the vector of right singular values V n*n Each column of (a) represents a magnitude distribution corresponding to a time domain feature; sorting the components according to the absolute value of the vector mean value of the right singular value, and selecting the largest f components; then calculating the correlation coefficient of the left singular value and the ECG template ECGtemp, sorting the components according to the magnitude of the correlation degree, and taking the largest f components;
5.4) combining the results of the two sorting to obtain t components, and taking the left singular value of the t components as the optimal base theta of the electrocardio artifact j J 1.. n, which is fit with least squares
Figure FDA0003714817500000022
Obtaining fitting parameters
Figure FDA0003714817500000023
From
Figure FDA0003714817500000024
Subtracting the artifact template theta j *b j Obtaining signals without electrocardio-artifacts
Figure FDA0003714817500000025
6) The signal segment obtained in the step 5) is processed
Figure FDA0003714817500000026
According to the position R of the R peak i Reconstructing the electroencephalogram signal to obtain a clean signal; with R i As the center, take 0.5 × T i-1 I.e. R i And R i-1 Taking 0.5 x T as the first half period pre i As the last half-cycle post; if pre>0.75*R m Then let pre be 0.75R m If post>0.75*R m If the post is 0.75R m (ii) a Copying the clean signal with the corresponding range centered on the R peak to (R) i -pre,R i + post) to obtain a clean electroencephalogram signal S without the electrocardio-artifacts c 。
2. The method of claim 1 for on-line removal of nuclear magnetic artifacts in synchronous EEG-fMRI data acquisition, characterized by: the parameters of the low-pass filter in the step 1) are controlled at 1000-1250 Hz.
CN202110286597.0A 2021-03-17 2021-03-17 Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition Active CN113040789B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110286597.0A CN113040789B (en) 2021-03-17 2021-03-17 Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110286597.0A CN113040789B (en) 2021-03-17 2021-03-17 Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition

Publications (2)

Publication Number Publication Date
CN113040789A CN113040789A (en) 2021-06-29
CN113040789B true CN113040789B (en) 2022-09-09

Family

ID=76513019

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110286597.0A Active CN113040789B (en) 2021-03-17 2021-03-17 Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition

Country Status (1)

Country Link
CN (1) CN113040789B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115251947B (en) * 2022-06-24 2024-09-10 四川大学 Method for removing myoelectric noise in electrocardiosignal based on singular value decomposition

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5474078A (en) * 1990-12-14 1995-12-12 Hutson; William H. Method and system for near real-time analysis and display of electrocardiographic signals
CN109567792A (en) * 2018-11-19 2019-04-05 北京工业大学 A kind of single channel abdomen record fetal electrocardiogram extracting method

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110028827A1 (en) * 2009-07-28 2011-02-03 Ranganatha Sitaram Spatiotemporal pattern classification of brain states
US9636019B2 (en) * 2010-10-07 2017-05-02 The Medical Research, Infrastructure, And Health Services Fund Of The Tel-Aviv Medical Center Device for use in electro-biological signal measurement in the presence of a magnetic field
KR20160044079A (en) * 2014-10-14 2016-04-25 고려대학교 산학협력단 Device and method for denoising of electroencephalography signal
KR20160044641A (en) * 2014-10-15 2016-04-26 고려대학교 산학협력단 Device and method for denoising noise of electroencephalogram signal
CN104905786A (en) * 2015-07-02 2015-09-16 北京师范大学 Electrocardiographic artifact on-line removal algorithm
KR101896273B1 (en) * 2016-11-15 2018-09-10 금오공과대학교 산학협력단 Noncontact bio-signal measurement method using SVD(Singular Value Decomposition)
CN106709244B (en) * 2016-12-12 2019-08-13 西北工业大学 A kind of tranquillization state synchronizes the brain function network modeling method of EEG-fMRI
CN106859641B (en) * 2017-02-20 2019-08-20 华南理工大学 A method of based on nuclear-magnetism artefact in automatic ICA removal EEG signal
KR102014597B1 (en) * 2017-08-23 2019-08-26 원광대학교산학협력단 Wearable multichannel photo plethysmography measuring device using singular value decomposition and method for removing noise from a signal using the same
CN107669244B (en) * 2017-10-27 2018-11-13 中国人民解放军国防科技大学 Epileptic abnormal discharge site positioning system based on EEG-fMRI
JP6884344B2 (en) * 2017-11-27 2021-06-09 株式会社国際電気通信基礎技術研究所 Brain network activity estimation system, brain network activity estimation method, brain network activity estimation program, and learned brain activity estimation model
CN107981862B (en) * 2017-11-30 2020-06-19 华南理工大学 Online denoising method for brain signals in nuclear magnetic resonance environment
CN109222965B (en) * 2018-09-21 2021-05-14 华南理工大学 Synchronous EEG-fMRI online artifact removing method
CN109820503A (en) * 2019-04-10 2019-05-31 合肥工业大学 The synchronous minimizing technology of a variety of artefacts in single channel EEG signals

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5474078A (en) * 1990-12-14 1995-12-12 Hutson; William H. Method and system for near real-time analysis and display of electrocardiographic signals
CN109567792A (en) * 2018-11-19 2019-04-05 北京工业大学 A kind of single channel abdomen record fetal electrocardiogram extracting method

Also Published As

Publication number Publication date
CN113040789A (en) 2021-06-29

Similar Documents

Publication Publication Date Title
Urigüen et al. EEG artifact removal—state-of-the-art and guidelines
CN108158573B (en) Electrocardiosignal noise reduction method based on adaptive threshold wavelet transformation
Roy et al. Image denoising by data adaptive and non-data adaptive transform domain denoising method using EEG signal
CN109222965B (en) Synchronous EEG-fMRI online artifact removing method
CN106485208B (en) The automatic removal method of eye electrical interference in single channel EEG signals
Navarro et al. ECG removal in preterm EEG combining empirical mode decomposition and adaptive filtering
Breuer et al. A constrained ICA approach for real-time cardiac artifact rejection in magnetoencephalography
CN107361764B (en) Method for rapidly extracting electrocardiosignal characteristic waveform R wave
Zou et al. Automatic EEG artifact removal based on ICA and Hierarchical Clustering
Raghavendra et al. Wavelet enhanced CCA for minimization of ocular and muscle artifacts in EEG
CN113040789B (en) Online removal method for nuclear magnetic artifact in synchronous EEG-fMRI data acquisition
JP2012030111A (en) Method and instrument for measuring brain wave, and recording medium
CN107981862B (en) Online denoising method for brain signals in nuclear magnetic resonance environment
Heydari et al. Adaptive wavelet technique for EEG de-noising
CN108926348B (en) A kind of extracting method and device of atrial fibrillation signal
Kaushal et al. Better approach for denoising EEG signals
Mayeli et al. An automatic ICA-based method for removing artifacts from EEG data acquired during fMRI in real time
abdullah et al. Separation enhancement of power line noise from human ECG signal based on stone technique
Wang et al. Automatic removal of ocular artifacts from electroencephalogram using Hilbert-Huang transform
CN116725550A (en) Method for removing ocular artifacts in multichannel electroencephalogram signals
Chu et al. An adaptive noise removal method for EEG signals
Rodríguez-González et al. High frequential resolution networks: Considerations on a new functional brain connectivity framework
Kiamini et al. A wavelet based algorithm for ocular artifact detection in the EEG signals
Wang et al. Modeling Strategies and Spatial Filters for Improving the Performance of P300-speller within and across Individuals
Rasheed et al. Attenuation of artifacts in EEG signals measured inside an MRI scanner using constrained independent component analysis

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant