CN110647870A - Method for calculating approximate entropy of resting state fMRI data based on sliding window - Google Patents
Method for calculating approximate entropy of resting state fMRI data based on sliding window Download PDFInfo
- Publication number
- CN110647870A CN110647870A CN201911057728.7A CN201911057728A CN110647870A CN 110647870 A CN110647870 A CN 110647870A CN 201911057728 A CN201911057728 A CN 201911057728A CN 110647870 A CN110647870 A CN 110647870A
- Authority
- CN
- China
- Prior art keywords
- sliding
- approximate entropy
- sliding window
- calculating
- brain region
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a method for calculating approximate entropy of resting state fMRI data based on a sliding window, which comprises the following steps: obtaining fMRI data; preprocessing data; determining an optimal sliding step length of the sliding window based on the average mutual information amount; dividing a target brain region time sequence by using a sliding window; calculating approximate entropy of the time series in each sliding window; and calculating the average value of the approximate entropy of the time series in each sliding window, and taking the average value as the approximate entropy of the time series of the brain region. According to the method, the approximate entropy of the time sequence of the target brain region is calculated based on the sliding window thought according to the stationary (pseudo-stationary) attribute of the brain signal in a short time interval, the sliding step length is determined by utilizing the average mutual information quantity, and the influence of the correlation degree of the time sequence in each sliding window on the approximate entropy is reduced. Compared with the traditional method for calculating the approximate entropy of the resting state fMRI data, the method can reflect the complexity of the time sequence of the target brain region more accurately and truly.
Description
Technical Field
The invention belongs to the technical field of biomedical image pattern recognition, and particularly relates to a method for calculating the approximate entropy of resting state fMRI data based on a sliding window.
Background
The human brain is an extremely complex network system, and a large number of neuronal cells exist in the human brain, and the neuronal cells are interconnected through synapses. Functional magnetic resonance imaging (fMRI) techniques reflect the physiological activity of neuronal cells in the brain by detecting a Blood Oxygen Level Dependent (BOLD) signal. With the increase of methods for analyzing and researching fMRI, people find that the brain is a nonlinear system, and the extracted fMRI signal is a nonlinear signal, so that people apply the nonlinear dynamics theory to the fMRI data analysis of the resting state, and calculate the correlation dimension, Lyapunov index, approximate entropy and the like of the fMRI data of the resting state in various physiological states.
Approximate entropy (ApEn) is a nonlinear parameter that measures sequence complexity and statistics, and distinguishes various processes by an effective statistical means, namely, distribution of edge probability. The complexity of a time sequence is represented by a non-negative number, the self-similarity degree of the time sequence on a mode is reflected, and the more complex time sequence corresponds to larger approximate entropy. The traditional method for calculating the approximate entropy of the fMRI data in the resting state is to directly calculate the approximate entropy of a target brain region time sequence, and easily ignore the property of whether signals are stable or not. Relevant research shows that brain signals are of a stationary (pseudo-stationary) property in a short time interval, so how to obtain approximate entropy capable of accurately and truly reflecting the complexity of acquired signals becomes a problem to be solved urgently.
Disclosure of Invention
Aiming at the technical defects in the prior art, the invention provides a method for calculating the approximate entropy of the fMRI data in the resting state based on a sliding window, aiming at calculating the approximate entropy of the fMRI data in the resting state by utilizing the sliding window idea, determining the optimal sliding step length based on the average mutual information quantity and reducing the influence of the correlation degree of a time sequence in each sliding window on the approximate entropy. Compared with the traditional method for calculating the approximate entropy of the resting state fMRI data, the method can reflect the complexity of the time sequence of the target brain region more accurately and truly.
A method for calculating approximate entropy of resting state fMRI data based on a sliding window comprises the following specific steps:
(1) recording a blood oxygen level dependent signal in a resting state process of a human brain by using a magnetic resonance scanner to obtain fMRI data;
(2) preprocessing the fMRI data;
(3) extracting a time sequence (1 x T vector) with a target brain region time point of T, giving a sliding window length L (1 is equal to or more than L and is equal to or less than T), setting a sliding step length as D (1 is equal to or more than D and is equal to or less than L), dividing the target brain region time sequence (1 x T vector) into [ T/L ] section time sequences (1 x L vector), changing the sliding step length, solving the sum of average mutual information quantity of the time sequences in two sliding windows, taking the sliding step length corresponding to the minimum value as the optimal sliding step length, wherein [ ] represents a downward integer function, [ T/L ] is the number of the sliding windows, and the formula for calculating the average mutual information quantity is as follows:
wherein IABRepresenting the average mutual information content, PAB(ai，bk) Means that the time series A and B are measured simultaneously, and the result is aiAnd bkProbability of (P)A(ai) And PB(bk) Respectively shows that the measurement is carried out on a time series A, B, and the measurement result is aiAnd bkThe average mutual information quantity statistically represents the degree of correlation between a and B;
(4) dividing a target brain region time sequence according to the step (3) by using the optimal sliding step length;
(5) calculating approximate entropy of the time series in each sliding window, wherein the approximate entropy algorithm comprises the following steps:
given a one-dimensional time sequence with the length of N, determining a mode dimension m and a similarity volume r, wherein m is 2, and r is 0.1-0.25 SDx(SDxStandard deviation of original time series), the series { xiThe m-dimensional vectors are composed in order, i.e.:
X(i)＝[x(i)，x(i+1)，…，x(i+m-1)]，i＝1，2，…N-m+1；
(ii) defining the distance d [ x (i) between x (i) and x (j) (j ═ 1, 2, … N-m +1, j ≠ i), where x (j) is the largest one of the two corresponding elements, as follows:
d[X(i)，X(j)]＝max|x(i+k)-x(j+k)|，k＝0，1，…m-1
namely, the maximum value of the absolute value of the difference between the corresponding elements of the two vectors is the distance between the two vectors;
thirdly, according to a given threshold r (r is more than 0), d [ X (X) ((r)) is counted for each value of ii)，X(j)]The number < r and the ratio of this number to the total number of vectors N-m +1 are denoted as
Fourthly, will
Taking the logarithm, and averaging the logarithm with all i, and recording as phim(r)：
Increasing m by 1, and repeating the steps from the first step to the fourth step;
sixthly, bym(r)，φm+1(r) finding the approximate entropy of this sequence as:
ApEn(m，r)＝limN→∞φm(r)-φm+1(r)
in practical calculation, N cannot be ∞, so when N takes a finite value, the approximate entropy can be estimated by statistics:
ApEn(m，r，N)＝φm(r)-φm+1(r)。
(6) and calculating the approximate entropy average value of the time series in each sliding window to serve as the approximate entropy of the time series of the brain region.
Further, as described above, the calculation method of the number of sliding windows in step (3) is [ T/L ].
Further, in the method as described above, the rule for determining the optimal sliding step in step (3) is that the sum of the average mutual information amounts of the time series in two time windows is the minimum.
Further, according to the method described above, the length of the sliding window in step (3) can be flexibly selected.
The invention has the beneficial effects that:
the invention provides a method for calculating approximate entropy of resting state fMRI data based on a sliding window, wherein the sliding step length of the sliding window is determined by the sum of average mutual information. The method utilizes the steady (pseudo-steady) property of brain signals in a short time interval, calculates the approximate entropy of the resting state fMRI data based on the sliding window idea, has the characteristics of high robustness and strong stability, can more accurately and truly reflect the complexity of a target brain area time sequence compared with the traditional method for calculating the approximate entropy of the resting state fMRI data, and provides a new effective technology for related research.
Drawings
FIG. 1 is a flowchart of a method for calculating the approximate entropy of resting fMRI data based on a sliding window according to the present invention.
Fig. 2 is a time series of target brain region extraction.
Fig. 3 is a graph of the sum of the average mutual information amount of the time series in the sliding window of every two target brain areas and the sliding step length.
Fig. 4 is a schematic diagram of the target brain region dividing time sequence according to the determined sliding step size.
Fig. 5 shows the approximate entropy of the time series in each sliding window of the target brain region and the average thereof.
Detailed Description
In order that the detailed description of the present invention will be further described below in conjunction with the accompanying drawings and examples, the following examples are provided only for illustrating the present invention and are not intended to limit the scope of the present invention.
Referring to fig. 1, a method for calculating approximate entropy of fMRI data in a resting state based on a sliding window includes the following steps:
(1) recording a blood oxygen level dependent signal in a brain resting state process of a tested person by using a magnetic resonance scanner to obtain original fMRI data of a target tested person;
(2) the method comprises the following steps of preprocessing raw fMRI data, wherein the main operations comprise: format conversion, converting the format of the original data from the format of DIM to NIFTI (hdr and img) format; deleting the first 20 time points; performing time layer correction by taking the 32 nd layer as a reference layer; head motion correction, wherein the tested data are screened according to a standard that the translation does not exceed 3 mm and the rotation does not exceed 3 degrees; image space normalization, normalizing the data to MNI (montreal neurology institute) human brain standard space according to the EPI template; image smoothing, using a gaussian function of full width at half maximum (FWHM) of 4 smoothing kernels; removing the linear drift; band-pass filtering, which is carried out in a frequency band of 0.01-0.08 Hz; removing covariates including cranial motor parameters, white matter signals and cerebrospinal fluid signals; extracting a time sequence of a brain region, and adopting an anatomical automatic labeling template (AAL); thus obtaining the preprocessed data.
(3) Extracting a time sequence (1 x 780 vector) with the tested target brain region time point of 780, setting the sliding step length to be D (1 is not less than D and not more than 130) as a result shown in figure 2, dividing the time sequence (1 x 780 vector) of the brain region into 6 time sequences (1 x 130 vector), changing the sliding step length, solving the sum of the average mutual information quantity of the time sequences in two sliding windows, taking the sliding step length corresponding to the minimum value as the optimal sliding step length, and obtaining the optimal sliding step length to be 93 according to figure 3;
(4) dividing the time sequence of the target brain region according to the step (3) by using the optimal sliding step length, wherein a schematic division diagram is shown in FIG. 4;
(5) calculating approximate entropy of the time series in each sliding window, wherein the mode dimension m and the similarity volume r are respectively 2 and 0.6290;
(6) the average value of the approximate entropy of the time series in each sliding window is calculated and is used as the approximate entropy of the time series of the brain region, namely the approximate entropy of the brain region is 0.5246, and the result is shown in fig. 5.
The method calculates the approximate entropy of the fMRI data in the resting state based on the sliding window idea, determines the optimal sliding step length by calculating the sum of average mutual information quantity, ensures that the correlation degree of time sequences in different sliding windows is minimum, and further ensures that the obtained approximate entropy is more accurate. Compared with the traditional method for directly calculating the approximate entropy of the resting state fMRI data, the method can more accurately and truly reflect the complexity of the time sequence of the target brain region.
Finally, it should be noted that: the above examples are intended only to illustrate the technical solution of the present invention more intuitively and not to limit it; the technical scheme and the characteristics can be changed, and the equivalent transformation and the improvement on the technical basis of the invention are not excluded from the protection scope of the invention.
Claims (4)
1. A method for calculating approximate entropy of fMRI data in a resting state based on a sliding window is characterized by comprising the following steps:
(1) recording a blood oxygen level dependent signal in a resting state process of a human brain by using a magnetic resonance scanner to obtain fMRI data;
(2) preprocessing the fMRI data;
(3) extracting a time sequence (1 x T vector) with a target brain region time point of T, giving a sliding window length L (L is more than or equal to 1 and less than or equal to T), setting a sliding step length as D (D is more than or equal to 1 and less than or equal to L), dividing the target brain region time sequence (1 x T vector) into [ T/L ] section time sequences (1 x L vectors), changing the sliding step length, solving the sum of the average mutual information quantity of the time sequences in every two sliding windows, and taking the sliding step length corresponding to the minimum value as the optimal sliding step length, wherein [ ] represents a downward integer taking function, and [ T/L ] is the number of the sliding windows;
(4) dividing a target brain region time sequence according to the step (3) by using the optimal sliding step length;
(5) calculating approximate entropy of the time series in each sliding window;
(6) and calculating the average value of the approximate entropy of the time series in each sliding window, and taking the average value as the approximate entropy of the time series of the brain region.
2. The method of claim 1, wherein the number of sliding windows in step (3) is calculated as [ T/L ].
3. The method of claim 1, wherein the rule of determining the optimal sliding step in step (3) is that the sum of the average mutual information amounts of the time series in two sliding windows is minimum.
4. The method of claim 1, wherein the length of the sliding window in step (3) can be flexibly selected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911057728.7A CN110647870B (en) | 2019-11-01 | 2019-11-01 | Method for calculating approximate entropy of resting state fMRI data based on sliding window |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911057728.7A CN110647870B (en) | 2019-11-01 | 2019-11-01 | Method for calculating approximate entropy of resting state fMRI data based on sliding window |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110647870A true CN110647870A (en) | 2020-01-03 |
| CN110647870B CN110647870B (en) | 2022-11-29 |
Family
ID=68995490
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911057728.7A Active CN110647870B (en) | 2019-11-01 | 2019-11-01 | Method for calculating approximate entropy of resting state fMRI data based on sliding window |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110647870B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114200879A (en) * | 2021-12-08 | 2022-03-18 | 长江存储科技有限责任公司 | Gas leakage monitoring method and device and computer storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020082495A1 (en) * | 2000-03-29 | 2002-06-27 | Biswal Bharat B. | Method for determining the reliability of fMRI parameters |
| CN106037741A (en) * | 2016-07-04 | 2016-10-26 | 电子科技大学 | Data method for detecting dynamic brain spontaneous activity based on fMRI |
| CN109522894A (en) * | 2018-11-12 | 2019-03-26 | 电子科技大学 | A method of detection fMRI brain network dynamic covariant |
-
2019
- 2019-11-01 CN CN201911057728.7A patent/CN110647870B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020082495A1 (en) * | 2000-03-29 | 2002-06-27 | Biswal Bharat B. | Method for determining the reliability of fMRI parameters |
| CN106037741A (en) * | 2016-07-04 | 2016-10-26 | 电子科技大学 | Data method for detecting dynamic brain spontaneous activity based on fMRI |
| CN109522894A (en) * | 2018-11-12 | 2019-03-26 | 电子科技大学 | A method of detection fMRI brain network dynamic covariant |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114200879A (en) * | 2021-12-08 | 2022-03-18 | 长江存储科技有限责任公司 | Gas leakage monitoring method and device and computer storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110647870B (en) | 2022-11-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109522894B (en) | Method for detecting dynamic covariation of fMRI brain network | |
| CN108537751B (en) | Thyroid ultrasound image automatic segmentation method based on radial basis function neural network | |
| CN110175649A (en) | It is a kind of about the quick multiscale estimatiL method for tracking target detected again | |
| CN115841434B (en) | Infrared image enhancement method for gas concentration analysis | |
| CN112068056B (en) | Method for determining FMRI dynamic brain function time window | |
| CN112465905A (en) | Characteristic brain region positioning method of magnetic resonance imaging data based on deep learning | |
| CN111754548A (en) | Multi-scale correlation filtering target tracking method and device based on response discrimination | |
| Setyawan et al. | MRI image segmentation using morphological enhancement and noise removal based on fuzzy C-means | |
| Jin et al. | Perceptual Gradient Similarity Deviation for Full Reference Image Quality Assessment. | |
| CN110647870B (en) | Method for calculating approximate entropy of resting state fMRI data based on sliding window | |
| CN110718301A (en) | Alzheimer disease auxiliary diagnosis device and method based on dynamic brain function network | |
| CN117764864A (en) | nuclear magnetic resonance tumor visual detection method based on image denoising | |
| CN108846407B (en) | Magnetic resonance image classification method based on independent component high-order uncertain brain network | |
| CN106709921B (en) | Color image segmentation method based on space Dirichlet mixed model | |
| CN111259914B (en) | Hyperspectral extraction method for characteristic information of tea leaves | |
| CN111640126A (en) | Artificial intelligence diagnosis auxiliary method based on medical image | |
| CN110927644A (en) | Method for measuring mechanical asynchrony in left ventricle based on cardiac nuclear magnetic resonance imaging | |
| CN115995282B (en) | Expiratory flow data processing system based on knowledge graph | |
| CN116091489B (en) | Electrocardiogram image recognition method and system based on deep learning | |
| CN117974530B (en) | Method and system for optimally reading fingerprint module information of liquid crystal display screen | |
| CN110703915B (en) | Particle method and particle analysis method for synchronously recognizing gesture and grasping force | |
| CN117745808B (en) | Electrocardiogram image positioning comparison method based on photogrammetry | |
| CN115965626B (en) | Medical image processing method and system based on artificial intelligence algorithm | |
| CN117275130B (en) | Intelligent access control verification system based on face recognition | |
| CN117115186B (en) | Cardiovascular segmentation method based on region growth |
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 |