CN110495879A - Brain wave patterns time-frequency characteristics extracting method based on information gain - Google Patents
Brain wave patterns time-frequency characteristics extracting method based on information gain Download PDFInfo
- Publication number
- CN110495879A CN110495879A CN201910694554.9A CN201910694554A CN110495879A CN 110495879 A CN110495879 A CN 110495879A CN 201910694554 A CN201910694554 A CN 201910694554A CN 110495879 A CN110495879 A CN 110495879A
- Authority
- CN
- China
- Prior art keywords
- sample
- feature
- brain wave
- wave patterns
- wavelet field
- 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
- 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
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- Psychiatry (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Physiology (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Fuzzy Systems (AREA)
- Psychology (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The present invention provides a kind of brain wave patterns time-frequency characteristics extracting methods based on information gain in brain wave patterns identification technology field.Brain wave patterns are the basic elements in electroencephalogram, and all kinds of brain wave patterns must be first identified when carrying out medical diagnosis, and the present invention is the feature extracting method in automatic identification brain wave patterns technology.The method of the present invention includes wavelet transformation, discretization and information gains to select three key steps.Each time domain samples is transformed to wavelet field sample by wavelet transformation, and transformation results are according to the sequential arrangement in frequency from high frequency to low frequency from front to back on the time.Each wavelet field sample is changed into discrete value vector from successive value vector by discretization.Information gain selection, which is chosen from small echo characteristic of field on the wavelet field sample basis of discretization according to information gain, acts on biggish feature to classification.The present invention has the advantages that highlighting the mutual difference of all kinds of brain wave patterns, the recognition efficiency of brain wave patterns is improved.
Description
Technical field
The present invention relates to brain wave patterns identification technology fields, and it is special to refer in particular to a kind of brain wave patterns time-frequency based on information gain
Levy extracting method.
Background technique
When brain activity, the synchronous postsynaptic potential variation occurred of a large amount of neurons can be detected, referred to as on scalp
For electroencephalogram (Electroencephalogram, EEG).Measurement electroencephalogram is a kind of electro physiology means for monitoring brain activity.
Brain wave patterns are the basic elements in electroencephalogram, and all kinds of brain wave patterns must be first identified when carrying out medical diagnosis.
Traditionally, after collecting electroencephalogram using electroencephalograph, all kinds of brain wave patterns are identified by way of artificial observation, are deposited
It can not effectively be distinguished in the mutual difference of the defect of inefficiency, and all kinds of brain wave patterns.
Therefore, how a kind of brain wave patterns feature extracting method is provided, realizes the recognition efficiency for promoting brain wave patterns, it is prominent
The mutual difference of all kinds of brain wave patterns becomes a urgent problem to be solved.
Summary of the invention
The technical problem to be solved in the present invention is to provide a kind of brain wave patterns time-frequency characteristics extraction based on information gain
Method realizes the mutual difference of prominent all kinds of brain wave patterns, promotes the recognition efficiency of brain wave patterns.
The invention is realized in this way the brain wave patterns time-frequency characteristics extracting method based on information gain includes following step
It is rapid:
Step S1: obtaining brain wave patterns from brain power supply signal, intercepts each brain wave patterns by setting length and obtains brain wave
The sample vector of shape, the sample set of length, classification information and sample vector the building brain wave patterns according to the brain wave patterns
It closes；
Step S2: the sample set is handled using wavelet transform and obtains wavelet field sample set；
Step S3: for wavelet field sample vector each dimension execute discretization process, each wavelet field sample from
Successive value vector is changed into discrete value vector；
Step S4: to the wavelet field sample set execution information gain selection course of the discrete value vector, output is advantageous
In the feature of classification；
Step S5: the identification according to the feature construction feature vector, for brain wave patterns；
Discretization process in the step S3 specifically includes:
Step S31: value of all wavelet field sample vectors in the dimension is taken out, is divided into not according to the classification of brain wave patterns
Same set；
Step S32: the mean value and variance of each set are calculated；
Step S33: being respectively that each set draws a normal distribution curve according to the mean value and variance,；
Step S34: finding out the intersection point of adjacent normal distribution curve, and number axis is divided into multiple value areas using the intersection point
Between, and give one, each section label；
Step S35: successive value of the wavelet field sample vector in the dimension is replaced with corresponding interval mark；
Information gain selection course in the step S4 specifically includes:
Step S41: setting information gain threshold is G；Check [i]=false, i=1 ..., M is set；With node structure
Variable T is created, T.cur=0, T.feature [i]=0, i=1 ..., M are set, T.sample [i]=true, i=is set
1,…,N；Queue queue is created, T is inserted into queue；
Step S42: if queue is sky, feature corresponding to check [i]=true is the selected feature of this process, knot
Beam information gain selection course, returns to main step；Otherwise, S43 is entered step；
Step S43: variable assignments is taken out to T from queue team head；T.feature [i] is deleted from F, i=1 ...,
T.cur is assigned to subF；
Step S44: a subset D is filtered out from wavelet field sample set using T.sample [i]=true as condition；According to
According to the sample class information in D, information is calculated according to sample class information and the mark value in section for features all in subF
Gain, it is gMax that wherein information gain is maximum, and corresponding signature identification is fMax；
Step S45: if gMax > G, enters step S46；Otherwise, check [T.feature [i]]=true, i=
1 ..., T.cur is transferred to step S42；
Step S46: according to the number of distinct values K of the fMax feature of wavelet field sample, it is directed to i=1 ..., K respectively, uses
NewT [i] .feature=T.feature, newT [i] .cur=T.cur+1, newT is arranged in node Structure Creating newT [i]
[i] .feature [newT [i] .cur]=fMax, newT [i] .sample=T.sample, if the fMax feature of sample
The corresponding entry of newT [i] .sample is not then changed to false for i-th of value, newT [i] is inserted into queue queue；It is transferred to step
Rapid S42；
Wherein, M is the Characteristic Number of wavelet field sample, and N is the number of sample in wavelet field sample set；Use integer
1 ..., M identification characteristics, F are the set of all signature identifications；The member cur that node structure includes is integer type, and cur is used for
The feature quantity that record algorithm has been selected in current location；Member feature [M] is integer array, its preceding cur are for remembering
The signature identification that record algorithm has been selected in current location；Sample [N] member is Boolean type array, for recording algorithm current
Which sample position has weeded out, and sample [i], which is equal to false, indicates that sample i is weeded out；Check [M] is Boolean type number
Group, for recording the selected feature of algorithm, check [i], which is equal to true, indicates that ith feature is chosen by algorithm.
Further, in the step S1, classification information include at least sinusoidal sample wave, bow wave, spike, sharp wave, slow wave,
Bi-phase wave, triphasic wave, polyphasic wave and complex wave.
Further, the step S2 specifically:
Handle each brain wave patterns sample in the sample set using wavelet transform, by transformation results according to
Time, order of the frequency from high frequency to low frequency was arranged from front to back, was formed and brain wave patterns sample length and time time
The same wavelet field sample of sequence constructs wavelet field sample set according to the wavelet field sample.
The present invention has the advantages that being converted to wavelet field sample by the change commanders sample set of brain wave patterns of discrete wavelet transformer
Set, and discretization is carried out to it, and then execution information gain selects to obtain the feature for being conducive to classification, construction feature vector is used
In the identification of brain wave patterns, it is greatly improved the recognition efficiency of brain wave patterns, convenient for the phase mutual deviation of all kinds of brain wave patterns of protrusion
It is different.
Detailed description of the invention
The present invention is further illustrated in conjunction with the embodiments with reference to the accompanying drawings.
Fig. 1 is the flow chart of the brain wave patterns time-frequency characteristics extracting method the present invention is based on information gain.
Fig. 2 is the schematic diagram of the brain wave patterns time-frequency characteristics extracting method the present invention is based on information gain.
Fig. 3 is wavelet field sample set discretization schematic diagram of the present invention.
Specific embodiment
Overall procedure of the invention is as shown in Figure 1, main processing steps and data flow are as shown in Figure 2.The present invention is based on letters
The preferred embodiment for ceasing the brain wave patterns time-frequency characteristics extracting method of gain, includes the following steps:
Step S1: obtaining brain wave patterns from brain power supply signal, intercepts each brain wave patterns by setting length and obtains brain wave
The sample vector of shape, the sample set of length, classification information and sample vector the building brain wave patterns according to the brain wave patterns
It closes；
Step S2: the sample set is handled using wavelet transform and obtains wavelet field sample set；
Step S3: for wavelet field sample vector each dimension execute discretization process, each wavelet field sample from
Successive value vector is changed into discrete value vector；
Step S4: to the wavelet field sample set execution information gain selection course of the discrete value vector, output is advantageous
In the feature of classification；
Step S5: the identification according to the feature construction feature vector, for brain wave patterns；
Discretization process in the step S3 specifically includes:
Step S31: value of all wavelet field sample vectors in the dimension is taken out, is divided into not according to the classification of brain wave patterns
Same set；
Step S32: the mean value and variance of each set are calculated；
Step S33: being respectively that each set draws a normal distribution curve according to the mean value and variance, as shown in Figure 3；
Step S34: finding out the intersection point of adjacent normal distribution curve, and number axis is divided into multiple value areas using the intersection point
Between, and give one, each section label；
Step S35: successive value of the wavelet field sample vector in the dimension is replaced with corresponding interval mark；
Information gain selection course in the step S4 specifically includes:
Step S41: setting information gain threshold is G；Check [i]=false, i=1 ..., M is set；With node structure
Variable T is created, T.cur=0, T.feature [i]=0, i=1 ..., M are set, T.sample [i]=true, i=is set
1,…,N；Queue queue is created, T is inserted into queue；Whether array check [i] is conducive to classify for marker characteristic, when taking
It indicates to be conducive to classify when value is true, indicates to be unfavorable for classifying when value is false；Array T.feature [i] is for marking
Remember which feature has been selected；Which wavelet field sample array T.sample [i] for marking processed.
Step S42: if queue is sky, feature corresponding to check [i]=true is the selected feature of this process, knot
Beam information gain selection course, returns to main step；Otherwise, S43 is entered step；
Step S43: variable assignments is taken out to T from queue team head；T.feature [i] is deleted from F, i=1 ...,
T.cur is assigned to subF；
Step S44: a subset D is filtered out from wavelet field sample set using T.sample [i]=true as condition；According to
According to the sample class information in D, information is calculated according to sample class information and the mark value in section for features all in subF
Gain, it is gMax that wherein information gain is maximum, and corresponding signature identification is fMax；
Step S45: if gMax > G, enters step S46；Otherwise, check [T.feature [i]]=true, i=
1 ..., T.cur is transferred to step S42；
Step S46: according to the number of distinct values K of the fMax feature of wavelet field sample, it is directed to i=1 ..., K respectively, uses
NewT [i] .feature=T.feature, newT [i] .cur=T.cur+1, newT is arranged in node Structure Creating newT [i]
[i] .feature [newT [i] .cur]=fMax, newT [i] .sample=T.sample, if the fMax feature of sample
The corresponding entry of newT [i] .sample is not then changed to false for i-th of value, newT [i] is inserted into queue queue；It is transferred to step
Rapid S42；
Wherein, M is the Characteristic Number of wavelet field sample, and N is the number of sample in wavelet field sample set；Use integer
1 ..., M identification characteristics, F are the set of all signature identifications；The member cur that node structure includes is integer type, and cur is used for
The feature quantity that record algorithm has been selected in current location；Member feature [M] is integer array, its preceding cur are for remembering
The signature identification that record algorithm has been selected in current location；Sample [N] member is Boolean type array, for recording algorithm current
Which sample position has weeded out, and sample [i], which is equal to false, indicates that sample i is weeded out；Check [M] is Boolean type number
Group, for recording the selected feature of algorithm, check [i], which is equal to true, indicates that ith feature is chosen by algorithm；
Node structure can be described as follows:
In the step S1, classification information includes at least sinusoidal sample wave, bow wave, spike, sharp wave, slow wave, bi-phase wave, three
Phase wave, polyphasic wave and complex wave.
The step S2 specifically:
Handle each brain wave patterns sample in the sample set using wavelet transform, by transformation results according to
Time, order of the frequency from high frequency to low frequency was arranged from front to back, was formed and brain wave patterns sample length and time time
The same wavelet field sample of sequence constructs wavelet field sample set according to the wavelet field sample.
In conclusion the present invention has the advantages that being converted to by the change commanders sample set of brain wave patterns of discrete wavelet transformer
Wavelet field sample set, and discretization is carried out to it, and then execution information gain selects to obtain the feature for being conducive to classification, building
Feature vector is used for the identification for brain wave patterns, is greatly improved the recognition efficiency of brain wave patterns, convenient for prominent all kinds of brains
The mutual difference of electrical waveform.
Although specific embodiments of the present invention have been described above, those familiar with the art should be managed
Solution, we are merely exemplary described specific embodiment, rather than for the restriction to the scope of the present invention, it is familiar with this
The technical staff in field should be covered of the invention according to modification and variation equivalent made by spirit of the invention
In scope of the claimed protection.
Claims (3)
1. the brain wave patterns time-frequency characteristics extracting method based on information gain, which comprises the steps of:
Step S1: obtaining brain wave patterns from brain power supply signal, intercepts each brain wave patterns by setting length and obtains brain wave patterns
Sample vector, the sample set of length, classification information and sample vector the building brain wave patterns according to the brain wave patterns；
Step S2: the sample set is handled using wavelet transform and obtains wavelet field sample set；
Step S3: discretization process is executed for each dimension of wavelet field sample vector, each wavelet field sample from continuous
Value vector is changed into discrete value vector；
Step S4: to the wavelet field sample set execution information gain selection course of the discrete value vector, output is conducive to point
The feature of class；
Step S5: the identification according to the feature construction feature vector, for brain wave patterns；
Discretization process in the step S3 specifically includes:
Step S31: value of all wavelet field sample vectors in the dimension is taken out, is divided into according to the classification of brain wave patterns different
Set；
Step S32: the mean value and variance of each set are calculated；
Step S33: being respectively that each set draws a normal distribution curve according to the mean value and variance；
Step S34: finding out the intersection point of adjacent normal distribution curve, and number axis is divided into multiple value intervals using the intersection point, and
Give one, each section label；
Step S35: successive value of the wavelet field sample vector in the dimension is replaced with corresponding interval mark；
Information gain selection course in the step S4 specifically includes:
Step S41: setting information gain threshold is G；Check [i]=false, i=1 ..., M is set；With node Structure Creating
T.cur=0, T.feature [i]=0, i=1 ..., M is arranged in variable T, and T.sample [i]=true, i=1 ..., N is arranged；
Queue queue is created, T is inserted into queue；
Step S42: if queue is sky, feature corresponding to check [i]=true is the selected feature of this process, terminates letter
Gain selection course is ceased, main step is returned；Otherwise, S43 is entered step；
Step S43: variable assignments is taken out to T from queue team head；T.feature [i] is deleted from F, i=1 ..., T.cur are assigned
It is worth to subF；
Step S44: a subset D is filtered out from wavelet field sample set using T.sample [i]=true as condition；According to D
In sample class information, calculate information according to sample class information and the mark value in section for features all in subF and increase
Benefit, it is gMax that wherein information gain is maximum, and corresponding signature identification is fMax；
Step S45: if gMax > G, enters step S46；Otherwise, check [T.feature [i]]=true, i=1 ...,
T.cur is transferred to step S42；
Step S46: according to the number of distinct values K of the fMax feature of wavelet field sample, it is directed to i=1 ..., K respectively, uses node
NewT [i] .feature=T.feature, newT [i] .cur=T.cur+1, newT [i] is arranged in Structure Creating newT [i]
.feature [newT [i] .cur]=fMax, newT [i] .sample=T.sample, if the fMax feature of sample is not
The corresponding entry of newT [i] .sample is then changed to false by i-th of value, and newT [i] is inserted into queue queue；It is transferred to step
S42；
Wherein, M is the Characteristic Number of wavelet field sample, and N is the number of sample in wavelet field sample set；With integer 1 ..., M mark
Know feature, F is the set of all signature identifications；The member cur that node structure includes is integer type, and cur is for recording algorithm
In the feature quantity that current location has been selected；Member feature [M] is integer array, its preceding cur exist for recording algorithm
The signature identification that current location has been selected；Sample [N] member is Boolean type array, has been sieved for recording algorithm in current location
Which sample is removed, sample [i], which is equal to false, indicates that sample i is weeded out；Check [M] is Boolean type array, for remembering
The selected feature of algorithm is recorded, check [i], which is equal to true, indicates that ith feature is chosen by algorithm.
2. the brain wave patterns time-frequency characteristics extracting method based on information gain as described in claim 1, it is characterised in that: described
In step S1, classification information includes at least sinusoidal sample wave, bow wave, spike, sharp wave, slow wave, bi-phase wave, triphasic wave, polyphasic wave
And complex wave.
3. the brain wave patterns time-frequency characteristics extracting method based on information gain as described in claim 1, it is characterised in that: described
Step S2 specifically:
Each brain wave patterns sample in the sample set is handled using wavelet transform, by transformation results according to the time
From front to back, order of the frequency from high frequency to low frequency is arranged, and is formed and brain wave patterns sample length and chronological order one
The wavelet field sample of sample constructs wavelet field sample set according to the wavelet field sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910694554.9A CN110495879B (en) | 2019-07-30 | 2019-07-30 | Brain waveform time-frequency feature extraction method based on information gain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910694554.9A CN110495879B (en) | 2019-07-30 | 2019-07-30 | Brain waveform time-frequency feature extraction method based on information gain |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110495879A true CN110495879A (en) | 2019-11-26 |
| CN110495879B CN110495879B (en) | 2021-11-09 |
Family
ID=68587845
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910694554.9A Active CN110495879B (en) | 2019-07-30 | 2019-07-30 | Brain waveform time-frequency feature extraction method based on information gain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110495879B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114391810A (en) * | 2022-01-06 | 2022-04-26 | 深圳元海恒鑫医疗科技有限公司 | Intelligent child anesthesia target control method and system based on electroencephalogram signal monitoring |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102542291A (en) * | 2011-12-23 | 2012-07-04 | 国网电力科学研究院 | Hyperspectral remote sensing image classification method based on binary decision tree |
| CN103092971A (en) * | 2013-01-24 | 2013-05-08 | 电子科技大学 | Classification method used in brain-computer interfaces |
| CN104899573A (en) * | 2015-06-18 | 2015-09-09 | 福州大学 | P300 feature extraction method based on wavelet transformation and Fisher criterion |
| CN109431497A (en) * | 2018-10-23 | 2019-03-08 | 南京医科大学 | A kind of brain-electrical signal processing method and epilepsy detection system |
-
2019
- 2019-07-30 CN CN201910694554.9A patent/CN110495879B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102542291A (en) * | 2011-12-23 | 2012-07-04 | 国网电力科学研究院 | Hyperspectral remote sensing image classification method based on binary decision tree |
| CN103092971A (en) * | 2013-01-24 | 2013-05-08 | 电子科技大学 | Classification method used in brain-computer interfaces |
| CN104899573A (en) * | 2015-06-18 | 2015-09-09 | 福州大学 | P300 feature extraction method based on wavelet transformation and Fisher criterion |
| CN109431497A (en) * | 2018-10-23 | 2019-03-08 | 南京医科大学 | A kind of brain-electrical signal processing method and epilepsy detection system |
Non-Patent Citations (4)
| Title |
|---|
| HADI RATHAM AL GHAYAB ET AL: "Epileptic seizures detection in EEGs blending frequency domain with information gain technique", 《SOFT COMPUTING》 * |
| R. UPADHYAY ET AL: "A comparative study of feature ranking techniques for A comparative study of feature ranking techniques for", 《COMPUTERS AND ELECTRICAL ENGINEERING》 * |
| 何帆等: "基于小波变换的投影寻踪快速图像分割", 《湖南理工学院学报(自然科学版)》 * |
| 薄洪健等: "猜拳博弈过程脑电的时频特征分析及基于优选特征的博弈决策预测", 《信号处理》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114391810A (en) * | 2022-01-06 | 2022-04-26 | 深圳元海恒鑫医疗科技有限公司 | Intelligent child anesthesia target control method and system based on electroencephalogram signal monitoring |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110495879B (en) | 2021-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104586383B (en) | A kind of ecg wave form sorting technique and device | |
| Othman et al. | Probabilistic neural network for brain tumor classification | |
| CN103714281B (en) | A kind of personal identification method based on electrocardiosignal | |
| CN105956623A (en) | Epilepsy electroencephalogram signal classification method based on fuzzy entropy | |
| CN104715261A (en) | FMRI dynamic brain function sub-network construction and parallel connection SVM weighted recognition method | |
| AU2008356919A1 (en) | A method and system for classifying brain signals in a BCI | |
| Zhu et al. | An efficient visibility graph similarity algorithm and its application on sleep stages classification | |
| Ji et al. | A deep multi-task learning approach for ECG data analysis | |
| David et al. | Improvement of framework for the grouping of CA diseases by investigating bigdata | |
| CN109684374A (en) | A kind of extracting method and device of the key-value pair of time series data | |
| CN109330613A (en) | Human body Emotion identification method based on real-time brain electricity | |
| CN105528072A (en) | Brain-computer interface speller by utilization of dynamic stop strategy | |
| CN110495879A (en) | Brain wave patterns time-frequency characteristics extracting method based on information gain | |
| CN115081486A (en) | Epileptic focus positioning system and method for intracranial electroencephalogram network in early stage of epileptic seizure | |
| CN113466634A (en) | Ground fault waveform identification method based on fault indicator | |
| Jasm et al. | Deep image mining for convolution neural network | |
| Briggman et al. | From crawling to cognition: analyzing the dynamical interactions among populations of neurons | |
| Toups et al. | Finding the event structure of neuronal spike trains | |
| Schwarz et al. | Semi-supervised spike sorting using pattern matching and a scaled Mahalanobis distance metric | |
| CN109472216A (en) | Radiation source feature extraction and individual discrimination method based on signal non-Gaussian feature | |
| CN110604572A (en) | Brain activity state identification method based on human brain characteristic map | |
| Wouters et al. | A data-driven spike sorting feature map for resolving spike overlap in the feature space | |
| CN106618552B (en) | A kind of detection model acquisition methods and device | |
| CN110110795A (en) | Image classification method and device | |
| Petrantonakis et al. | A novel and simple spike sorting implementation |
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 |