CN106037725A - Bio-electrical brain signal analysis method based on wavelet packet transformation - Google Patents
Bio-electrical brain signal analysis method based on wavelet packet transformation Download PDFInfo
- Publication number
- CN106037725A CN106037725A CN201610487800.XA CN201610487800A CN106037725A CN 106037725 A CN106037725 A CN 106037725A CN 201610487800 A CN201610487800 A CN 201610487800A CN 106037725 A CN106037725 A CN 106037725A
- Authority
- CN
- China
- Prior art keywords
- wavelet
- signal
- band
- method based
- analysis method
- 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
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
-
- 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
- A61B5/7221—Determining signal validity, reliability or quality
-
- 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
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
-
- 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
- A61B5/7235—Details of waveform analysis
-
- 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
- A61B5/7235—Details of waveform analysis
- A61B5/7253—Details of waveform analysis characterised by using transforms
-
- 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
- A61B5/7271—Specific aspects of physiological measurement analysis
-
- G06F19/34—
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Physics & Mathematics (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Public Health (AREA)
- Signal Processing (AREA)
- Animal Behavior & Ethology (AREA)
- Psychiatry (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Power Engineering (AREA)
- Psychology (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
According to the scheme, the invention relates to a bio-electrical brain signal analysis method based on wavelet packet transformation. The method comprises the following steps: pre-processing collected brain signals, so that pre-processed signals are obtained; in accordance with the pre-processed signals, conducting wavelet transformation decomposition by virtue of a wavelet basis function which has the maximum matching degree, so that characteristic signals of an interested frequency band or frequency sub-band are obtained; conducting status division on the characteristic signals by establishing a threshold T, so that a status code of the interested frequency band or frequency sub-band is finally obtained. According to the scheme, the interested frequency band or frequency sub-band can be extracted more precisely through the wavelet packet transformation, status judgment is conducted, and patient statuses are depicted in the combination with different frequency band statuses; therefore, the method can guide analysis on the patient statuses; and the method has status specificity and self-adaptability.
Description
Technical field
The present invention relates to the data processing method of a kind of bio electricity brain signal, particularly to one based on wavelet package transforms
Brain field potential signal analysis method.
Background technology
Accurately determining for clinical analysis of patient's states is particularly significant, but the most a lot of physiological status is the most objective
The most dyskinetic degree of standard evaluated, the grade of chronic pain, the generally determination to these states rely primarily on doctor's warp
Test and the experience of patient self, there is the biggest subjectivity.Research for human-body biological pcs signal finds cerebral activity
Level with a lot of physiological statuss be correlated with, these signals can apply to the qualitative assessment of patient's states, and guides controlling of patient
Treat.
Bio electricity brain signal contains substantial amounts of information, but the only information of component frequency band and particular state phase
Close, it is therefore desirable to frequency band information is extracted, and obtains the relation between these information and patient's states.But existing
In technology, more single to the extracting method of these information, the way of substantially similar clean cut, i.e. set a power-threshold
Value, exceeding this threshold value is a kind of state, not less than be another kind of state, this rough extracting method is obviously follow-up shape
State analysis provides higher erroneous judgement information.
Summary of the invention
The deficiency existed for prior art, this case provides a kind of bio electricity brain signal based on wavelet package transforms and divides
Analysis method.Wavelet package transforms provides a kind of finer data classification method, and it can the most effectively extract arrowband letter
Number, and can be adaptive selected according to demand frequency band then according to frequecy characteristic select suitable wavelet function obtain accurately
Band information.When what multiple frequency bands carried out time relevant to patient's states that Combined expression can be finer is carried out patient's states
Portray.
For achieving the above object, this case is achieved through the following technical solutions:
A kind of bio electricity brain signal analysis method based on wavelet package transforms, comprising:
Step 1) pretreatment:
The initial data of the brain signal collected is screened, picks out useful signal；
Described useful signal is calculated differential signal；
Described differential signal is removed Hz noise and baseline drift, finally gives preprocessed signal；
Step 2) feature extraction:
For described preprocessed signal, the wavelet basis function selecting matching degree the highest carries out wavelet package transforms decomposition, it is thus achieved that
Band of interest or the characteristic signal of frequency subband；
Step 3) condition discrimination:
By setting up threshold value T, described characteristic signal is carried out state demarcation, finally give band of interest or frequency
The state encoding of band.
Preferably, described bio electricity brain signal analysis method based on wavelet package transforms, wherein, in step 1) in
Also include described differential signal is carried out low-pass filtering.
Preferably, described bio electricity brain signal analysis method based on wavelet package transforms, wherein, in step 1) in
Also include described differential signal is carried out down-sampled process.
Preferably, described bio electricity brain signal analysis method based on wavelet package transforms, wherein, step 2) in,
The system of selection of described wavelet basis function is:
The requirement in terms of symmetry, orthogonality, compact sup-port of the characteristic signal according to band of interest or frequency subband, slightly
Selecting several wavelet basis functions with similarity degree, being calculated respectively if substituting in wavelet package transforms function the most respectively
The entropy of dry wavelet packet coefficient, is asserted, by the wavelet packet basis functions corresponding to minimum entropy, the wavelet basis letter that matching degree is the highest
Number.
Preferably, described bio electricity brain signal analysis method based on wavelet package transforms, wherein, band of interest
Or the characteristic signal of frequency subband be utilize wavelet package transforms after the corresponding band of interest that obtains or the small echo of frequency subband
Bag coefficient C characterizes.
Preferably, described bio electricity brain signal analysis method based on wavelet package transforms, wherein, step 3) in,
Described threshold value T is prepared by the following:
Take intermediate value to early than the wavelet packet coefficient C in a period of time of current data, obtain median (| C |)；
If σ=median (| C |)/0.6745；
σ is substituted into minimaxi function and obtains result of calculation t；
T is multiplied by threshold weights, obtains threshold value T corresponding to current data.
Preferably, described bio electricity brain signal analysis method based on wavelet package transforms, wherein, in step 3)
In, the method for state demarcation is:
It is critical with threshold value T, each wavelet packet coefficient C is compared with threshold value T respectively, thus by the feelings of all C >=T
The situation of shape and all C ＜ T distinguishes.
Preferably, described bio electricity brain signal analysis method based on wavelet package transforms, wherein, described state is compiled
Code uses binary coding.
The invention has the beneficial effects as follows: it is more accurate to frequency band interested or frequency that this case utilizes wavelet package transforms
Subband extracts, and carries out condition discrimination, combines different band condition and portrays patient's states, thus instructs trouble
The analysis of person's state；This method has state specificity and adaptivity, selects wavelet package transforms extract the most subtly and suffer from
The Bio-computer signal characteristic that person's state is relevant, and provide the selection scheme of parameter in solution wavelet package transforms, effectively
The effectiveness of the wavelet package transforms improved and accuracy；And this case selects threshold model to differentiate patient's states, to shape
State encodes, it is possible to finer divides morbid state.
Accompanying drawing explanation
Fig. 1 is to realize, based on Bio-computer signal, the model framework figure that patient's states differentiates.
Fig. 2 is the pretreatment schematic diagram in patient's states distinguished number model.
Fig. 3 is the feature extraction schematic diagram in patient's states distinguished number model.
Fig. 4 is the schematic diagram that the feature extracted carries out condition discrimination.
Fig. 5 is theta and the alpha band activity schematic diagram utilizing wavelet package transforms to extract.
Fig. 6 is that different wavelet basis function is at the wavelet packet coefficient entropy extracting theta and alpha band activity.
Fig. 7 is the alpha frequency band analogue signal generated, and its mid frequency is 10.5Hz, and signal to noise ratio is-5dB, and envelope is week
Phase is the trapezoidal wave figure of 4s.
Fig. 8 is the comparison diagram utilizing algorithm model that analogue signal is tested result and the notional result obtained.
Fig. 9 obtains theta and alpha frequency for utilizing this case model that true chronic pain signal carries out signal analysis extraction
Band wavelet packet coefficient and each rhythmic activity state is divided.
Figure 10 is that patient's states is encoded by the state utilizing theta and alpha rhythmic activity.
Detailed description of the invention
The present invention is described in further detail below in conjunction with the accompanying drawings, to make those skilled in the art with reference to description literary composition
Word can be implemented according to this.
The bio electricity brain signal analysis method based on wavelet package transforms of an embodiment is listed in this case, comprising:
Step 1) pretreatment:
The initial data of the brain signal collected is screened, picks out useful signal；
Useful signal is calculated differential signal；
Differential signal is gone Hz noise and baseline drift, finally gives preprocessed signal；
Step 2) feature extraction:
For preprocessed signal, the wavelet basis function selecting matching degree the highest carries out wavelet transformation decomposition, it is thus achieved that interested
Frequency band or the characteristic signal of frequency subband；
Step 3) condition discrimination:
By setting up threshold value T, characteristic signal is carried out state demarcation, finally give band of interest or frequency subband
State encoding.
Wherein, in step 1) in also include differential signal is carried out low-pass filtering.
Wherein, in step 1) in also include differential signal is carried out down-sampled process.
Wherein, step 2) in, the system of selection of wavelet basis function is:
The requirement in terms of symmetry, orthogonality, compact sup-port of the characteristic signal according to band of interest or frequency subband, slightly
Selecting several wavelet basis functions with similarity degree, being calculated respectively if substituting in wavelet package transforms function the most respectively
The entropy of dry wavelet packet coefficient, is asserted, by the wavelet basis function corresponding to minimum entropy, the wavelet packet basis letter that matching degree is the highest
Number.
Wherein, the characteristic signal of band of interest or frequency subband is that after utilizing wavelet package transforms, the corresponding sense that obtains is emerging
The wavelet packet coefficient C of interest frequency band or frequency subband characterizes.
Wherein, step 3) in, threshold value T is prepared by the following:
Take intermediate value to early than the wavelet packet coefficient C in a period of time of current data, obtain median (| C |)；
If σ=median (| C |)/0.6745；
σ is substituted into minimaxi function and obtains result of calculation t；
T is multiplied by threshold weights, obtains threshold value T corresponding to current data.
Wherein, in step 3) in, the method for state demarcation is:
It is critical with threshold value T, each wavelet packet coefficient C is compared with threshold value T respectively, thus by the feelings of all C >=T
The situation of shape and all C ＜ T distinguishes.
Wherein, state encoding preferably employs binary coding.The band of interest selected or the number of frequency subband will certainly
Delimit the organizational structure the figure place of code, when the most only 1 band of interest or frequency subband, be divided into two states: more than or equal to the shape of threshold value
State is " 1 ", and is " 0 " less than the state of threshold value；If having two core groups to have 1 and 2 frequency-of-interest subbands respectively, then to suffer from
Person's state will have 8 kinds of integrated modes: 000,001,010,011,100,101,110 and 111.This case is not limited to binary system and compiles
Code, other coding form can also.
Determining band of interest and frequency subband according to practical situation, wherein practical situation refers to: such as: parkinson state with
The beta frequency range (15-30Hz) of STN core group is closely related, and research finds that pain status rolls into a ball neural activity with PAG and VPL core for another example
Theta with alpha frequency band or frequency subband relevant etc..This case may be used for any frequency band, and example bands includes δ band, α band, β
Band and γ carry and the frequency subband of these frequency bands.Frequency subband has the band width narrower than frequency band, and therefore frequency band can be by many
Individual frequency subband is defined.
The different frequency bands of bio electricity brain signal is associated with different conditions.Illustrate that frequency band divides in the following table 1 one
Individual example:
Table 1, frequency band divide
| Frequency band | Scope (Hz) |
| δ frequency band | 1.5-4Hz |
| θ frequency band | 4-8Hz |
| α frequency band | 8-13Hz |
| β frequency band | 13-30Hz |
| γ frequency band | 30-90Hz |
But, frequency band range is not limited only to the scope in above table, can include that multiple frequency subband (such as, can be by
Multiple frequency subbands form), each frequency subband has the width more narrower than frequency band, the frequency subband of frequency band can have identical or
Different width.
Activity in the special frequency band of the bio electricity brain signal detected be likely to be due to the state of patient change and
Change, it is also possible to the reason that patient's states changes, therefore the bioelectric in special frequency band is close with patient's states
Relevant, can be used to indicate patient's states.
Comparing the state of a single band signal, the active state of multiple frequency subbands of multiple cores group can be more preferable
Patient's states is divided, the most multiple frequency subbands can divide the different brackets of pain, from one-level to ten grade.Based on
One or more band of interest of one or more cores group or the wavelet packet coefficient of frequency subband calculate threshold value division state can
So that patient's states is analyzed.
The present embodiment is as a example by chronic pain conditions, but in other examples, in addition to chronic pain, and this algorithm model
Can be applicable to the analysis of other patient's states, such as, but be not limited to, the condition adjudgement such as dyskinesia.
Although run through the bio electricity brain signal of this case mainly as a example by human deep field potential signal (LFP), but this case
The algorithm provided can be additionally used in such as, but is not limited to, scalp brain electricity (EEG), magneticencephalogram, functional MRI data etc. and patient
The bio electricity brain signal that state analysis is relevant.
This algorithm model is broadly divided into 3 modules, its structural representation as shown in Figure 1: pretreatment module, feature extraction mould
Block and condition discrimination module.
The specific implementation of algorithm model is as follows:
In order to make result more accurate, the initial data collected is carried out Preprocessing, as in figure 2 it is shown, mainly wrap
Include following steps:
1) signal that signal quality is the best is rejected, judge whether the signal received comes from including utilizing signal
The cerebral nucleus group being pre-selected.
2) the four-way signal obtained from the four of recording electrode contacts calculates differential signal；
3) signal is carried out 50Hz notch filter and removes Hz noise, due to the local field potentials master relevant to patient's states
Below 90Hz to be concentrated on, therefore can do 90Hz low-pass filtering and remove high-frequency signal, then do 3Hz high-pass filtering and get rid of base
Line drift about, last signal be downsampled to sample rate be 500Hz to reduce amount of calculation, improve calculating speed.
Pretreated data are carried out feature extraction, concrete steps characteristic extracting module as shown in Figure 3.Utilize small echo
Packet transform decomposition data, extracts band of interest or the signal of frequency subband, utilizes wavelet packet coefficient to characterize these band signals
Movable situation.
Feature extraction is carried out in real time, and the characteristic extracting module signal to receiving carries out the shortest time window and processes, example
As often received 0.2s signal processing once, it is also possible to select other relatively short period of time windows.This algorithm model has real-time, its
It is achieved in that each module the most only processes the latest data of very short time, and the process window length of each module is the most identical,
Confirm as theta (6-9Hz) frequency band being 0.33s after process window length is chosen in this example, and alpha (9-12Hz) frequency band is
0.29s.The priori data of condition discrimination module also has real-time simultaneously, and its mode is when discrimination model completes a time window
Just data in window are brought in priori data after the differentiation of interior data, and abandon the number that the afterbody in original priori data is isometric
According to, the most both kept the length of prior information, be updated the most simultaneously.
Utilize wavelet package transforms that local field potentials signal is decomposed in calculating window, and extract characteristic spectra, obtain
Corresponding wavelet packet coefficient.First have to carry out parameter selection before doing wavelet package transforms, comprise determining that suitable wavelet basis function
Extract the specific signal rhythm and pace of moving things, and Decomposition order when decomposing, entropy, and extract the isoparametric setting of node.But owing to decomposing
The number of plies, entropy, and the selection extracting these parameters of node are well known in the art general under current state, therefore, and these ginsengs
Number is equivalent to be exactly known, and what unique needs selected is exactly wavelet basis function.
Above-mentioned for the system of selection of wavelet basis function in wavelet package transforms be: combine need to extract with patient's states phase
The characteristic of the prosodic feature closed and the characteristic of wavelet basis function itself, if symmetry, orthogonality and compact sup-port are to wavelet basis function
Carry out coarse sizing and obtain one group of reliable wavelet basis function group.Be compared with that these wavelet basis functions extract after feature is little
The entropy of ripple bag coefficient, the most applicable current feature extraction of wavelet basis function of the least correspondence of entropy, thus obtain for this rhythm and pace of moving things special
The Optimum wavelet basic function levied.Fig. 6 is to utilize the result organizing the more different wavelet basis function of actual signal more, and statistics finds
Bior3.7 wavelet basis has good performance at extraction alpha frequency band entropy relative to other, and for theta frequency band rbio3.7
Preferably.
Above-mentioned do WAVELET PACKET DECOMPOSITION extract feature time Decomposition order and extract node determination method: need according to feature save
The frequency range of rule changes the sample rate of primary signal, enables complete the extracting from WAVELET PACKET DECOMPOSITION tree of the required rhythm and pace of moving things.Example
If the frequency range of the feature rhythm and pace of moving things is 9-12Hz, then need primary signal resampling to make sample rate become 384Hz, so according to small echo
Bag decomposes law, and 9-12Hz is present in node [6 2], and therefore Decomposition order is 6 layers, and extracting node is [6 2].
In the parameter of above-mentioned wavelet package transforms, entropy uses Shannon entropy.
The feature extracted is carried out condition discrimination, and as shown in Figure 4, condition discrimination module is based on the small echo extracting feature for method
Bag coefficient utilizes threshold model to calculate threshold value, more than or equal to threshold value corresponding states " 1 ", less than threshold value corresponding states " 0 ".This right
The coding of state is to use binary coding, and this case can also select other coding forms.
Above-mentioned threshold model builds based on threshold denoising, the small echo of its theoretical signal rhythm and pace of moving things for having pattern feature
Bag coefficient is greater than equal to the wavelet packet coefficient under random manner.Its concrete steps include:
1) obtain noise level is carried out estimating σ, for median (| Cj,k|)/0.6745, wherein Cj,kFor wavelet packet coefficient.
2) select " minimaxi " threshold estimation method estimation threshold value t；
3) threshold value t obtains final threshold value T being multiplied by threshold weights, and in this example, threshold weights is set to 1.
Further, feature decision module has predictive ability, uses the priori data of regular length to calculate threshold value conduct
Prediction to current data threshold value.
In order to verify that the effectiveness experiment of this case method generates the analogue signal (Fig. 8) of the 9-12Hz rhythm and pace of moving things, its center frequency
Rate is the signal of 10.5Hz and adds-5dB noise signal, its envelope be the cycle be the trapezoidal wave of 4s, utilize algorithm model to calculate
Showing that the wavelet basis function of wavelet package transforms selects bior3.7, Decomposition order is 6 layers, and extracting node is [6 2], and in theory
Calculate and open and close the time stimulated.Fig. 8 also includes utilize this case the calculated dynamic threshold of algorithm model and
State demarcation, and compare with theory state, result display algorithm has higher precision and accuracy, correctly opening of algorithm
The rate of opening is 95.14%, and correct closedown rate is 98.28%.It should be noted that the signal to noise ratio of analogue signal has relatively for accuracy rate
Big impact.
In this example, utilize this algorithm that the field potential signal of chronic pain patient is carried out state demarcation, band of interest
Being respectively theta (6-9Hz) and alpha frequency band (9-12Hz) (Fig. 9), wherein signal is essentially from PVAG and sensation thalamic nuclei
Group.Experiment finds to exist in data by four kinds of patterns (Figure 10) of two kinds of frequency band combinations simultaneously: " 00 ", " 01 ", " 10 ",
" 11 ", and the percentage ratio of these four pattern is relevant with the state of patient.
The experiment of analog data and true pain field potential data demonstrates algorithm model and can effectively realize patient's states
Differentiation, overcome the most simple use power spectrumanalysis as the tolerance relevant to patient's states, while also solve standard
The problem really extracting the rhythm and pace of moving things.
Although embodiment of the present invention are disclosed as above, but it is not restricted in description and embodiment listed
Using, it can be applied to various applicable the field of the invention completely, for those skilled in the art, and can be easily
Realizing other amendment, therefore under the general concept limited without departing substantially from claim and equivalency range, the present invention does not limit
In specific details with shown here as the legend with description.
Claims (8)
1. a bio electricity brain signal analysis method based on wavelet package transforms, it is characterised in that including:
Step 1) pretreatment:
The initial data of the brain signal collected is screened, picks out useful signal；
Described useful signal is calculated differential signal；
Described differential signal is removed Hz noise and baseline drift, finally gives preprocessed signal；
Step 2) feature extraction:
For described preprocessed signal, the wavelet basis function selecting matching degree the highest carries out wavelet package transforms decomposition, it is thus achieved that feel emerging
Interest frequency band or the characteristic signal of frequency subband；
Step 3) condition discrimination:
By setting up threshold value T, described characteristic signal is carried out state demarcation, finally give band of interest or frequency subband
State encoding.
2. bio electricity brain signal analysis method based on wavelet package transforms as claimed in claim 1, it is characterised in that in step
Rapid 1) also include in described differential signal is carried out low-pass filtering.
3. bio electricity brain signal analysis method based on wavelet package transforms as claimed in claim 1, it is characterised in that in step
Rapid 1) also include in described differential signal is carried out down-sampled process.
4. bio electricity brain signal analysis method based on wavelet package transforms as claimed in claim 1, it is characterised in that step
2), in, the system of selection of described wavelet basis function is:
The requirement in terms of symmetry, orthogonality, compact sup-port of the characteristic signal according to band of interest or frequency subband, roughly selects out
Several have the wavelet basis function of similarity degree with characteristic signal, substitute in wavelet package transforms function the most respectively and calculate respectively
Obtain the entropy of several wavelet packet coefficients, the wavelet packet basis functions corresponding to minimum entropy is asserted the small echo that matching degree is the highest
Basic function.
5. bio electricity brain signal analysis method based on wavelet package transforms as claimed in claim 1, it is characterised in that feel emerging
Interest frequency band or the characteristic signal of frequency subband be utilize wavelet package transforms after the corresponding band of interest that obtains or frequency subband
Wavelet packet coefficient C characterize.
6. bio electricity brain signal analysis method based on wavelet package transforms as claimed in claim 1, it is characterised in that step
3), in, described threshold value T is prepared by the following:
Take intermediate value to early than the wavelet packet coefficient C in a period of time of current data, obtain median (| C |)；
If σ=median (| C |)/0.6745；
σ is substituted into minimaxi threshold calculations function and obtains result of calculation t；
T is multiplied by threshold weights, obtains threshold value T corresponding to current data.
7. bio electricity brain signal analysis method based on wavelet package transforms as claimed in claim 1, it is characterised in that in step
Rapid 3), in, the method for state demarcation is:
Be critical with threshold value T, each wavelet packet coefficient C compared with threshold value T respectively, thus by the situation of all C >=T and
The situation of all C ＜ T distinguishes.
8. bio electricity brain signal analysis method based on wavelet package transforms as claimed in claim 1, it is characterised in that described
State encoding uses binary coding.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610487800.XA CN106037725B (en) | 2016-06-28 | 2016-06-28 | Bioelectricity brain signal analysis method based on wavelet package transforms |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610487800.XA CN106037725B (en) | 2016-06-28 | 2016-06-28 | Bioelectricity brain signal analysis method based on wavelet package transforms |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106037725A true CN106037725A (en) | 2016-10-26 |
| CN106037725B CN106037725B (en) | 2019-08-27 |
Family
ID=57166737
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610487800.XA Active CN106037725B (en) | 2016-06-28 | 2016-06-28 | Bioelectricity brain signal analysis method based on wavelet package transforms |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106037725B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109222906A (en) * | 2018-09-13 | 2019-01-18 | 复旦大学 | A method of pain status prediction model is constructed based on brain electrical signals |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100114813A1 (en) * | 2008-10-20 | 2010-05-06 | Zalay Osbert C | Method and rhythm extractor for detecting and isolating rhythmic signal features from an input signal using the wavelet packet transform |
| CN101835420A (en) * | 2007-08-23 | 2010-09-15 | 麻省理工学院 | Be used for reducing method and apparatus based on the number of active lanes of the epileptic seizure detector of EEG |
| CN105395192A (en) * | 2015-12-09 | 2016-03-16 | 恒爱高科(北京)科技有限公司 | Wearable emotion recognition method and system based on electroencephalogram |
| CN105559777A (en) * | 2016-03-17 | 2016-05-11 | 北京工业大学 | Electroencephalographic identification method based on wavelet packet and LSTM-type RNN neural network |
-
2016
- 2016-06-28 CN CN201610487800.XA patent/CN106037725B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101835420A (en) * | 2007-08-23 | 2010-09-15 | 麻省理工学院 | Be used for reducing method and apparatus based on the number of active lanes of the epileptic seizure detector of EEG |
| US20100114813A1 (en) * | 2008-10-20 | 2010-05-06 | Zalay Osbert C | Method and rhythm extractor for detecting and isolating rhythmic signal features from an input signal using the wavelet packet transform |
| CN105395192A (en) * | 2015-12-09 | 2016-03-16 | 恒爱高科(北京)科技有限公司 | Wearable emotion recognition method and system based on electroencephalogram |
| CN105559777A (en) * | 2016-03-17 | 2016-05-11 | 北京工业大学 | Electroencephalographic identification method based on wavelet packet and LSTM-type RNN neural network |
Non-Patent Citations (1)
| Title |
|---|
| 冯志燕等: "基于互样本熵的脑电功能网络构建及其应用", 《电子设计工程》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109222906A (en) * | 2018-09-13 | 2019-01-18 | 复旦大学 | A method of pain status prediction model is constructed based on brain electrical signals |
| CN109222906B (en) * | 2018-09-13 | 2021-06-04 | 复旦大学 | Method for constructing pain state prediction model based on brain electrical signals |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106037725B (en) | 2019-08-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pachori | Discrimination between ictal and seizure-free EEG signals using empirical mode decomposition. | |
| Kumari et al. | Seizure detection in EEG using time frequency analysis and SVM | |
| CN110781945A (en) | Electroencephalogram signal emotion recognition method and system integrating multiple features | |
| CN111657889A (en) | Non-contact driver fatigue detection method based on millimeter wave radar | |
| CN104586387A (en) | Method for extracting and fusing time, frequency and space domain multi-parameter electroencephalogram characters | |
| CN109480872B (en) | Driving fatigue detection method based on electroencephalogram signal frequency band energy ratio characteristics | |
| CN109222906B (en) | Method for constructing pain state prediction model based on brain electrical signals | |
| CN106236080B (en) | The removing method of myoelectricity noise in EEG signals based on multichannel | |
| CN108143409A (en) | Sleep stage method and device by stages | |
| CN112957055B (en) | EEMD-PCA-based method and device for removing motion artifacts in EEG signals | |
| Ramos-Aguilar et al. | Analysis of EEG signal processing techniques based on spectrograms | |
| CN113208613A (en) | Multi-mode BCI (binary coded decimal) timing optimization method based on FHLS (FHLS) feature selection | |
| CN110403602A (en) | Improvement public space pattern feature extracting method for EEG signals sentiment analysis | |
| Azami et al. | Automatic signal segmentation based on singular spectrum analysis and imperialist competitive algorithm | |
| CN106037725A (en) | Bio-electrical brain signal analysis method based on wavelet packet transformation | |
| Jiao et al. | Change-point detection using spectral PCA for multivariate time series | |
| CN112784686A (en) | Adaptive signal analysis method based on continuous multivariable variational modal decomposition | |
| Easwaramoorthy et al. | Fuzzy generalized fractal dimensions using inter-heartbeat interval dynamics in ECG signals for age related discrimination | |
| JP3933568B2 (en) | Brain function measuring device | |
| Mehboob et al. | Information quantification of empirical mode decomposition and applications to field potentials | |
| Soleymanpour et al. | Entropy-based EEG time interval selection for improving motor imagery classification | |
| Li et al. | Improved GP algorithm for the analysis of sleep stages based on grey model | |
| Janjarasjitt et al. | Comparison of complexity measures using two complex system analysis methods applied to the epileptic ECoG | |
| CN107657278B (en) | Optimal sample number sampling method for multi-classification of electroencephalogram signal modes | |
| Świderski et al. | Epileptic seizure prediction using Lyapunov exponents and support vector machine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |