CN111175272B - Performance-adjustable neuron calcium fluorescent signal peak detection system - Google Patents
Performance-adjustable neuron calcium fluorescent signal peak detection system Download PDFInfo
- Publication number
- CN111175272B CN111175272B CN202010194078.7A CN202010194078A CN111175272B CN 111175272 B CN111175272 B CN 111175272B CN 202010194078 A CN202010194078 A CN 202010194078A CN 111175272 B CN111175272 B CN 111175272B
- Authority
- CN
- China
- Prior art keywords
- peak
- suspected
- signal
- discriminator
- sensitivity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/64—Fluorescence; Phosphorescence
- G01N21/6402—Atomic fluorescence; Laser induced fluorescence
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/64—Fluorescence; Phosphorescence
- G01N21/6486—Measuring fluorescence of biological material, e.g. DNA, RNA, cells
Landscapes
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Pathology (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Engineering & Computer Science (AREA)
- Optics & Photonics (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention discloses a performance-adjustable neuron calcium fluorescent signal spike detection system, which comprises the steps of firstly obtaining all suspected spike moments on a neuron calcium signal through coarse detection, then respectively using a high-sensitivity discriminator and a high-specificity discriminator to discriminate the suspected spike in a fine detection stage, wherein the adopted discriminator is obtained by performing supervised learning on the multidimensional features of a pre-selected signal or fluorescent video, finally setting and adjusting confidence weights of output results of the two discriminators according to user preference, and outputting the finally identified spike moments in combination with an acceptance threshold of a user. The neuron fluorescence signal peak detection system is accurate in detection, wide in application range and adjustable in performance, and can effectively reduce the burden of later-stage manual calibration.
Description
Technical Field
The invention relates to the field of signal processing, in particular to a system for detecting peaks on neuron fluorescence signals recorded by a two-photon microscope.
Background
The vital activity is closely related to the behavior of neurons. In order to explore how a complex network of neurons is connected and interacted under specific behaviors, it is often necessary to dye the calcium ions of the neurons by a dye, then observe and analyze the calcium fluorescent signals emitted by the neurons under a two-photon microscope, and the peaks on the signals are closely related to the release of action potentials of the neurons, so that accurately detecting the peaks is a key for finding the rules of the nervous system and uncovering the mystery of brain science.
The two-photon microscope has the advantages that the two-photon microscope has very high spatial resolution, fluorescent signals issued by a plurality of neurons can be observed in a recorded video at the same time, however, due to factors such as the granularity of light, electronic thermal noise and the like, the signal-to-noise ratio of the signals is often lower, and the peak detection is difficult; in addition, because the respective electric signals of each neuron cannot be recorded while two-photon observation is performed, namely, the peak on the neuron calcium fluorescence signal does not have an absolute reliable standard, the peak detection task has great subjectivity, and the full-automatic detection system is difficult to meet the demands of scientific researchers.
At present, the most direct solution for detecting the peak of the fluorescent signal of the neuron is a threshold value discrimination method, namely, a signal locus with the fluorescent intensity higher than a base line to a certain extent is set as peak moment, however, the method ignores useful information such as peak morphology and the like, is only suitable for detecting the peak with higher fluorescent intensity, and has high false positive rate when detecting the peak with weak fluorescent intensity, and the obtained result still needs to be manually deleted in a large amount; the application range of the dynamic modeling method based on fluorescence characteristics is very limited, the method is only effective on peaks with specific forms, and changes in peak forms can be caused by changing the coloring agent and changing the observed cerebral cortex positions, so that the method is not applicable any more; in addition, different experimental designs have different requirements for peak detection, such as long-time recording often only focuses on more obvious peaks, and as many large and small peaks as possible need to be detected under certain short-time tasks. Therefore, a neuron fluorescence signal peak detection system with accurate detection, wide application range and adjustable performance is needed to reduce the burden of manual calibration in the later period.
Disclosure of Invention
The invention mainly aims to meet the requirement of detecting neuron fluorescence signal peaks and provides a neuron calcium fluorescence signal peak detection system with adjustable performance. The technical scheme of the invention is as follows:
an automatic detection system for the peak of the calcium fluorescent signal of neuron is composed of local peak detector for converting the calcium fluorescent signal C of neuron into a sequence consisting of n signals suspected to be peakThe fine detection module consists of a multidimensional feature extractor, a high-sensitivity peak discriminator and a high-specificity peak discriminator and is responsible for T 0 Probability sequence converted into suspected peak and distinguished as correct peak under high sensitivity and high specificity respectively +.>And P is taken up 1 、P 2 Delivering to a preference setting module; the preference setting module is responsible for setting the probability sequence P according to the preference of the user on the sensitivity and the specificity 1 、P 2 Conversion to the final discriminant probability sequence->Then according to the acceptance probability threshold p of the user for discrimination th Will T 0 Conversion to the final automatically recognized sequence of peak positions +.>
The local peak detector obtains the first-order difference of the calcium fluorescent signal C and makes the absolute value of the difference higher than a preset coarse detection threshold A th All n positions with opposite signs of the left and right first order differential values are marked as a position sequence T of suspected peaks 0 。
The high-sensitivity peak discriminator and the high-specificity peak discriminator in the fine detection module respectively calculate a high-sensitivity feature matrix M according to the calcium fluorescence signal C or the calcium fluorescence video V by the multidimensional feature extractor 1 And a high specificity feature matrix M 2 Will T 0 Probability sequence P converted into suspected peak and distinguished as correct peak under high sensitivity and high specificity 1 、P 2 。
The preference setting module outputs the confidence weight omega of the result according to the high-sensitivity peak discriminator given by the user 1 And confidence weight omega for high specificity peak discriminator output result 2 The weighted sum result omega 1 ×P 1 +ω 2 ×P 2 As T 0 Final predicted probability sequence P for suspected spikes in the middle 3 Then according to the acceptance threshold p given by the user th In sequence P 3 In which the final prediction probability is greater than p th M elements of (2) to reserve T 0 M suspected peak moments in the sequence, the rest n-m suspected peak moments are removed, and a final automatically-identified peak position sequence T is output 1 。
The multidimensional feature extractor is used for extracting k in a plurality of dimensions according to the pre-selection 1 Features sensitive to spikes, k 2 Sequences T of peak-specific features and n times suspected of being peaks 0 The input calcium fluorescence signal C or calcium fluorescence video V is converted into a high-sensitivity characteristic matrix M 1 And a high specificity feature matrix M 2 Respectively transmitting the signals to a high-sensitivity peak discriminator and a high-specificity peak discriminator; m is M 1 And M 2 The following formula is shown:
wherein, the liquid crystal display device comprises a liquid crystal display device,is k 1 The peak-sensitive features are corresponding to the signal segments of the calcium fluorescence signal C or the feature values corresponding to the video segments of the calcium fluorescence video V at n suspected peak moments, < >>Is k 2 The peak-specific features correspond to feature values corresponding to signal segments of the calcium fluorescence signal C or video segments of the calcium fluorescence video V at n suspected peak moments.
The high-sensitivity peak discriminator is a peak classifier obtained by performing supervised learning on high-sensitivity features obtained by extracting multidimensional features in advance, namely, k sensitive to peaks is selected in multiple dimensions in advance 1 Features, for manually markingA batch of correct peaks and incorrect peaks are correctly distinguished, so that a classifier with distinguishing capability is obtained, and the method is realized according to a high-sensitivity characteristic matrix M 1 K of n suspected peaks 1 Pairs of characteristic values T 0 Discrimination is performed, and a sequence P consisting of probability that each suspected peak is predicted to be a correct peak by a high-sensitivity peak discriminator under high sensitivity is output 1 。
The high specificity peak discriminator is a peak classifier obtained by performing supervised learning on the high specificity features obtained by extracting multidimensional features in advance, namely, k specific to the peak is selected in multiple dimensions in advance 2 The feature is that a batch of correct peaks and error peaks marked by manpower are correctly distinguished, thus obtaining a classifier with distinguishing capability, and realizing the characteristic matrix M according to high specificity 2 K of n suspected peaks 2 Pairs of characteristic values T 0 Discrimination is carried out, and each suspected peak is H under high specificity 2 Sequence P predicted as probability composition of correct spike 2 。
In order to ensure that the fine detection module can execute discrimination according to expert experience and meet the requirement of a user on the required peak form, the adopted high-sensitivity peak discriminator and the high-specificity peak discriminator are peak classifiers obtained by performing supervised learning through a linear kernel SVM in advance, namely according to k selected in advance 1 Peak sensitive features and k 2 And (3) accurately distinguishing a group of correct peaks and incorrect peaks marked manually by the peak-specific features, so as to obtain the classifier with discrimination capability.
In order to ensure that the high-sensitivity peak discriminator has the characteristic of high-sensitivity discrimination on signal peaks, k is selected in the process of supervised training of the high-sensitivity peak discriminator and during discrimination of suspected peaks 1 The multidimensional features are characterized by a calcium fluorescence signal C and a transform domain signal C trans Peak amplitude, peak width, peak left slope, peak right slope composition of local signal segment at suspected peak time, i.e. k 1 =8; wherein the transform domain signal C trans The calcium fluorescence signal C is regarded as the amplitude of a certain minimum phase signalSpectrum obtained by obtaining a minimum phase signal by taking a real cepstrum reconstruction of the magnitude spectrum and then differentiating the phase spectrum of the minimum phase signal, the transformation process highlighting fluctuations present on the original signal, the resulting transformed domain signal C trans With peaks that are the same as the suspected peaks on the calcium fluorescence signal C but are enhanced in magnitude.
In order to ensure that the high specificity peak discriminator has the characteristic of high specificity when discriminating the suspected peak of the signal, the method is used for the supervision training H 2 K selected in process and during suspected peak discrimination 2 The multidimensional features are composed of a calcium fluorescent signal C and a space-time consistency signal C corr Peak amplitude, peak width, peak left slope, peak right slope composition of local signal segment at suspected peak time, i.e. k 2 =8; wherein the space-time consistency signal C corr The method is that the brightness change curve of all pixel points in the neuron cell body region in the fluorescence video V is calculated to obtain the pearson correlation coefficient in the time window with the specified length, the curve representing the consistency of the change of the fluorescence intensity of each part in the neuron cell body region along with the time is obtained, and the time-space consistency signal C is obtained because the change of the calcium ion concentration with stronger synchronism can only appear in the whole cell body region when the neuron emits action potential corr The upper spike indicates a spike that is more consistent with the onset of action potential by the neuron.
The invention has the following beneficial effects:
1) According to the method, the suspected peak is combined with expert experience to carry out secondary screening through a double-scale detection strategy combining thickness and fineness, the obtained detection result is closer to manual marking, and the result is more accurate.
2) According to the method, the suspected peak is judged through the characteristics selected in multiple dimensions, no fixed model assumption exists, the fluorescent characteristic of the coloring agent is not limited, the peak can be selected according to the peak form in a specific scene, and the application range is wider.
3) According to the invention, the free adjustment of the detection performance is realized according to the confidence weights of the user on the high-sensitivity discriminant and the high-specificity discriminant and the setting of the receiving threshold value, so that the burden of manual modification in the later period is reduced.
Drawings
FIG. 1 is a flow chart of a process of spike detection of neuronal fluorescent signals by the system of the present invention;
FIG. 2 is a graph showing the effect of spike detection on a neuronal calcium fluorescent signal in an exemplary embodiment;
fig. 3 is a diagram of a transform domain signal and a space-time consistency signal corresponding to a calcein signal when extracting multidimensional features of suspected peaks of a neuron in an embodiment.
Detailed Description
The automatic detection system of the neuron calcium fluorescent signal peak of the invention, as shown in figure 1, consists of a local peak detector, a fine detection module and a preference setting module,
1. two-photon fluorescence video and neuron original calcium signal acquisition
The nerve cells of the primary motor cortex of the brain of a living mouse are stained by using a gene coding calcium ion indicator GCaMP6f, a new generation miniature two-photon fluorescence microscope of Beijing super-dimensional scene biotechnology limited company is adopted for video acquisition, the video is used as a fluorescence video of an input detection system, the fluorescence signals emitted by 276 neurons in total in the field of view, namely 276 neuron original calcium fluorescence signals, the sampling rate of the signals is 10Hz, the acquisition time length is 300 seconds, and the original calcium fluorescence signals of one neuron are shown as (a) in fig. 2.
The method comprises the steps of sequentially carrying out signal filtering, baseline drift removal and normalization on an obtained neuron original calcium fluorescent signal, firstly adopting a nonlinear median filter to process the neuron original calcium fluorescent signal, setting the window width of the median filter to be 5 sampling points so as to remove jump caused by an abnormal value in the signal, then adopting a Savitzky-Golay filter based on least square fitting to filter the signal, selecting the fitting order to be 2, and selecting the fitting window width to be 21 so as to effectively filter high-frequency noise in the signal on the premise of retaining the peak shape; and then removing baseline drift of the signals, adaptively decomposing the signals into a plurality of eigenmode functions by adopting empirical mode decomposition, removing residual steady-state quantity from the signals, removing the baseline drift, finally carrying out normalization processing on the signals, linearly mapping the maximum value and the minimum value of each group of fluorescent signals into the range of [0,1], thereby adjusting the dynamic range of the signals, and obtaining preprocessed calcium fluorescent signals which are used as neuron calcium signals input into a detection system, as shown in (b) of fig. 2.
2. Local peak coarse detector
The local peak detector calculates 23184 peak positions of the pre-processed 276 group neuron fluorescence signals, namely positions with local derivative absolute values higher than 0.005 and opposite left and right derivative sign values, as peak rough detection results, wherein some peak positions meet the user requirements, some peak positions are false peak positions, the peak rough detection result of one neuron is shown in fig. 2 (c), and a hollow black circle marked on the signal represents the position of the suspected peak obtained through rough detection.
3. Fine detection module
The multidimensional feature extractor extracts multidimensional features of the peaks obtained by the coarse detection, provides criteria for subsequent fine detection of the peaks, the extracted high-sensitivity features consist of 8 features of peak amplitude, peak width, peak left slope and peak right slope of the preprocessed calcium fluorescent signals and the transformation domain signals thereof at peak positions, and the extracted high-specificity features consist of 8 features of peak amplitude, peak width, peak left slope and peak right slope of the preprocessed calcium fluorescent signals and corresponding space-time consistency signals at peak positions, and the preprocessed calcium fluorescent signals and the corresponding transformation domain signals and space-time consistency signals thereof of one neuron are shown in fig. 3. The above 8 high-sensitivity characteristics and 8 high-specificity characteristics are respectively selected to obtain a high-sensitivity characteristic matrix and a high-specificity characteristic matrix with the sizes of 23184 multiplied by 8, wherein the first dimension of the matrix is the number of suspected peaks, and the second dimension is the selected characteristic number for judging the suspected peaks.
Firstly, inviting the expert in the biology field to manually mark peaks of calcium fluorescent signals of 5 neurons, and performing supervised learning on a high-sensitivity peak discriminator and a high-specificity peak discriminator in advance before fine detection. The neurons obtain a total of 471 suspected peaks obtained by coarse detection, 244 peaks which accord with expert labeling are added, the high-sensitivity feature matrix of the expert labeling peaks and the non-expert labeling peaks are sent into a linear kernel SVM to be distinguished to the greatest extent, a high-sensitivity peak discriminator capable of giving out the probability that the suspected peaks accord with the expert labeling peaks according to the 8 high-sensitivity features is obtained, the high-specificity feature matrix of the expert labeling peaks and the non-expert labeling peaks are sent into the linear kernel SVM to be distinguished to the greatest extent, a high-specificity peak discriminator capable of giving out the suspected peaks according to the 8 high-specificity features to accord with the expert labeling peaks is obtained, the high-sensitivity feature matrix and the high-specificity feature matrix of the 23184 suspected peaks which accord with the expert labeling peaks are respectively distinguished by the two discriminators, and a sequence consisting of the probability that each suspected peak is predicted to be a correct peak by the discriminator under the high sensitivity and a sequence consisting of the probability that each suspected peak is predicted to be a correct peak by the discriminator under the high specificity are obtained.
4. Preference setting module and result output
The preference setting module sets a confidence weight omega for the high sensitivity peak discriminator according to the confidence weight omega given by the user 1 And confidence weight ω for high specificity spike discriminant 2 Calculating weighted sum probability sequence of two probability sequences output by the high sensitivity peak discriminator and the high specificity peak discriminator, and according to a given acceptance threshold p th And finding out elements exceeding the acceptance threshold from the weighted summation probability sequence, and reserving peak moments in the coarse detection results corresponding to the elements to obtain final detection results. The confidence weights set for the first time are omega respectively 1 =0.8 and ω 2 =0.2, accept threshold p th =0.9, i.e. the detection of the bias high specificity is performed, and the number of the finally obtained peaks is 12760; the confidence weights set for the second time are omega respectively 1 =0.3 and ω 2 =0.7, accept threshold p th =0.9, i.e. perform biasingDetecting the heavy high sensitivity, wherein the number of finally obtained peaks is 13816; the final detection results obtained for both preferences set on the same neuronal calcium signal are shown in fig. 2 (d) and fig. 2 (e).
In this example, the original calcium fluorescence signal emitted by primary motor cortex nerve cells of the brain of a living mouse stained with GCaMP was obtained, and the signal was first preprocessed, and then the peak detection was performed according to the system of the present invention, so as to obtain the detection results shown in fig. 2 and 3. FIG. 2 is an effect diagram of spike detection on a neuron, which is sequentially an original calcium fluorescent signal, a preprocessed calcium fluorescent signal, a spike time obtained by coarse detection, and a spike time obtained by fine detection according to two preference settings; FIG. 3 is a graph of a calcium fluorescent signal and its corresponding transform domain signal, spatiotemporal consensus signal, showing that the magnitude of some of the suspected peaks on the transform domain will be enhanced for discrimination of the suspected peaks at high sensitivity, whereas an insignificant peak is missing on the spatiotemporal consensus signal, the remaining peaks are biased against the more significant peaks indicative of neuronal activity for discrimination of the suspected peaks at high specificity.
Claims (3)
1. The automatic detection system for the peak of the neuron calcium fluorescent signal is characterized by comprising a local peak detector, a fine detection module and a preference setting module, wherein:
the local peak detector converts the input neuron calcium fluorescence signal C into a sequence formed by the moments suspected to be peaks on n signals
Sequence T formed by time points suspected to be peak by fine detection module 0 Probability sequence converted into suspected peak and distinguished as correct peak under high sensitivity and high specificityAnd P is taken up 1 、P 2 Delivery to preferencesThe fine detection module is composed of a multidimensional feature extractor, a high-sensitivity peak discriminator and a high-specificity peak discriminator, wherein: the multidimensional feature extractor is used for extracting k in a plurality of dimensions according to the pre-selection 1 Features sensitive to spikes, k 2 Sequences T of peak-specific features and n times suspected of being peaks 0 The input calcium fluorescence signal C or calcium fluorescence video V is converted into a high-sensitivity characteristic matrix M 1 And a high specificity feature matrix M 2 Respectively transmitting the signals to a high-sensitivity peak discriminator and a high-specificity peak discriminator; m is M 1 And M 2 The following formula is shown:
is k 1 Feature values of a spike-sensitive feature at n suspected spike times, the feature values consisting of a calcium fluorescent signal C and a transform domain signal C trans Peak amplitude, peak width, peak left slope, peak right slope composition of local signal segment at suspected peak time, i.e. k 1 =8, wherein the transform domain signal C trans Taking a calcium fluorescent signal C as an amplitude spectrum of a certain minimum phase signal, obtaining the minimum phase signal by solving a real cepstrum reconstruction of the amplitude spectrum, and then solving a difference of the phase spectrum of the minimum phase signal;
is k 2 Feature values of peak-specific features at n suspected peak moments, which are composed of a calcium fluorescent signal C and a spatiotemporal coincidence signal C corr Peak amplitude, peak width, peak left slope, peak right slope composition of local signal segment at suspected peak time, i.e. k 2 =8, wherein the space-timeConsistency signal C corr The method comprises the steps of obtaining a pearson correlation coefficient of brightness change curves of all pixel points in a neuron cell body region in a fluorescence video V in a time window with a specified length;
the high-sensitivity peak discriminator is based on the high-sensitivity characteristic matrix M 1 K of n suspected peaks 1 Pairs of characteristic values T 0 Discrimination is performed to output a sequence of probability components that each suspected spike is predicted to be a correct spike by the high-sensitivity spike discriminator under high sensitivity
The high specificity peak discriminator is based on the high specificity characteristic matrix M 2 K of n suspected peaks 2 Pairs of characteristic values T 0 Discrimination is carried out, and a sequence consisting of probability that each suspected peak is predicted to be a correct peak by a high specificity peak discriminator under high specificity is output
The preference setting module sets the probability sequence P according to the preference of the user on the sensitivity and the specificity 1 、P 2 Conversion to the final discriminant probability sequenceThen according to the acceptance probability threshold p of the user for discrimination th Will T 0 Conversion to the final automatically recognized sequence of peak positions +.>
2. The automatic detection system for spikes of calcium fluorescent signals of neurons according to claim 1, wherein the local spike detector obtains the first-order difference of the calcium fluorescent signals C and makes the absolute value of the difference higher than a preset coarse detection threshold A th All the first-order differential values on the left and right sides have opposite signsPosition sequence T with n positions marked as suspected peaks 0 。
3. The system for automatically detecting spikes in calcium-based signals of neurons according to claim 1, wherein the preference setting module is configured to determine the confidence weight ω given by the user to the output of the high sensitivity spike discriminator 1 And confidence weight omega for high specificity peak discriminator output result 2 The weighted sum result omega 1 ×P 1 +ω 2 ×P 2 As T 0 Final predicted probability sequence P for suspected spikes in the middle 3 Then according to the acceptance threshold p given by the user th In sequence P 3 In which the final prediction probability is greater than p th M elements of (2) to reserve T 0 M suspected peak moments in the sequence, the rest n-m suspected peak moments are removed, and a final automatically-identified peak position sequence T is output 1 。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010194078.7A CN111175272B (en) | 2020-03-19 | 2020-03-19 | Performance-adjustable neuron calcium fluorescent signal peak detection system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010194078.7A CN111175272B (en) | 2020-03-19 | 2020-03-19 | Performance-adjustable neuron calcium fluorescent signal peak detection system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111175272A CN111175272A (en) | 2020-05-19 |
| CN111175272B true CN111175272B (en) | 2023-07-25 |
Family
ID=70658560
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010194078.7A Active CN111175272B (en) | 2020-03-19 | 2020-03-19 | Performance-adjustable neuron calcium fluorescent signal peak detection system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111175272B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113092887A (en) * | 2021-04-07 | 2021-07-09 | 中国电子科技集团公司第五十八研究所 | ENG-oriented low-power-consumption dual-threshold peak detection processing method and detection circuit |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102047246A (en) * | 2008-05-28 | 2011-05-04 | 皇家飞利浦电子股份有限公司 | Method and system for determining a threshold for spike detection of electrophysiological signals |
| CN104318242A (en) * | 2014-10-08 | 2015-01-28 | 中国人民解放军空军工程大学 | High-efficiency SVM active half-supervision learning algorithm |
| CN109214318A (en) * | 2018-08-22 | 2019-01-15 | 北京天泽智云科技有限公司 | A method of finding the faint spike of unstable state time series |
| CN110090002A (en) * | 2018-06-21 | 2019-08-06 | 北京大学 | A kind of automatic testing method of mouse prefrontal lobe neuron two-photon fluorescence Ca2+ oscillations |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105518455B (en) * | 2013-09-09 | 2017-05-03 | 株式会社岛津制作所 | Peak detection method |
-
2020
- 2020-03-19 CN CN202010194078.7A patent/CN111175272B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102047246A (en) * | 2008-05-28 | 2011-05-04 | 皇家飞利浦电子股份有限公司 | Method and system for determining a threshold for spike detection of electrophysiological signals |
| CN104318242A (en) * | 2014-10-08 | 2015-01-28 | 中国人民解放军空军工程大学 | High-efficiency SVM active half-supervision learning algorithm |
| CN110090002A (en) * | 2018-06-21 | 2019-08-06 | 北京大学 | A kind of automatic testing method of mouse prefrontal lobe neuron two-photon fluorescence Ca2+ oscillations |
| CN109214318A (en) * | 2018-08-22 | 2019-01-15 | 北京天泽智云科技有限公司 | A method of finding the faint spike of unstable state time series |
Non-Patent Citations (3)
| Title |
|---|
| 基于BT-SVM模型组合的动态加权多分类算法研究;李涛，张景肖;《统计与信息论坛》;20190131;第34卷(第1期);20-24 * |
| 基于支持向量机和k-近邻分类器的多特征融合方法;陈丽，陈静;《计算机应用》;20090331;第29卷(第3期);833-835 * |
| 进化的加权随机SVM集群算法研究;王志刚，胥茜，毕夏安;《研究与开发》;20191130(第31期);21-24 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111175272A (en) | 2020-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111709902B (en) | Infrared and visible light image fusion method based on self-attention mechanism | |
| Johnson et al. | The spatial transformation of color in the primary visual cortex of the macaque monkey | |
| Clement et al. | A new colour vision system to quantify automatically foliar discolouration caused by insect pests feeding on leaf cells | |
| CN101561932B (en) | Method and device for detecting real-time movement target under dynamic and complicated background | |
| CN108960182A (en) | A kind of P300 event related potential classifying identification method based on deep learning | |
| CN109934154A (en) | A kind of remote sensing image variation detection method and detection device | |
| CN102421007A (en) | Image quality evaluating method based on multi-scale structure similarity weighted aggregate | |
| CN1317624C (en) | Method of extracting brain machine interface control signa based on instantaneous vision sense induced electric potential | |
| CN113662560B (en) | Method for detecting seizure-like discharge between attacks, storage medium and device | |
| CN111175272B (en) | Performance-adjustable neuron calcium fluorescent signal peak detection system | |
| CN102214290B (en) | License plate positioning method and license plate positioning template training method | |
| CN108957125A (en) | Smart frequency spectrum figure analysis method based on machine learning | |
| CN111067513B (en) | Sleep quality detection key brain area judgment method based on characteristic weight self-learning | |
| CN113010013A (en) | Wasserstein distance-based motor imagery electroencephalogram migration learning method | |
| CN101605273A (en) | A kind of method of evaluating colour saturation quality and subsystem | |
| CN109598284A (en) | A kind of hyperspectral image classification method based on large-spacing distribution and space characteristics | |
| CN113796873B (en) | Wearable dynamic electrocardiosignal classification method and system | |
| Jin et al. | Perceptual Gradient Similarity Deviation for Full Reference Image Quality Assessment. | |
| CN106073767B (en) | Phase synchronization measurement, coupling feature extraction and the signal recognition method of EEG signal | |
| Motoyoshi et al. | Differential roles of contrast polarity reveal two streams of second-order visual processing | |
| CN113109782B (en) | Classification method directly applied to radar radiation source amplitude sequence | |
| CN110251076B (en) | Method and device for detecting significance based on contrast and fusing visual attention | |
| CN111281382A (en) | Feature extraction and classification method based on electroencephalogram signals | |
| CN115508288A (en) | Tea soup turbidity determination method based on mask normal fitting | |
| Díaz-Huerta et al. | Image preprocessing to improve Acid-Fast Bacilli (AFB) detection in smear microscopy to diagnose pulmonary tuberculosis |
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 |