WO2023173770A1 - 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法 - Google Patents

基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法 Download PDF

Info

Publication number
WO2023173770A1
WO2023173770A1 PCT/CN2022/131368 CN2022131368W WO2023173770A1 WO 2023173770 A1 WO2023173770 A1 WO 2023173770A1 CN 2022131368 W CN2022131368 W CN 2022131368W WO 2023173770 A1 WO2023173770 A1 WO 2023173770A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
laplacian
calcium signal
dimensionality reduction
throughput
Prior art date
Application number
PCT/CN2022/131368
Other languages
English (en)
French (fr)
Inventor
张韶岷
张奕玮
刘腾俊
陈卫东
Original Assignee
浙江大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 浙江大学 filed Critical 浙江大学
Publication of WO2023173770A1 publication Critical patent/WO2023173770A1/zh

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/145Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
    • A61B5/14546Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue for measuring analytes not otherwise provided for, e.g. ions, cytochromes
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/24Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316Modalities, i.e. specific diagnostic methods
    • A61B5/388Nerve conduction study, e.g. detecting action potential of peripheral nerves
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Measuring for diagnostic purposes; Identification of persons
    • A61B5/72Signal processing specially adapted for physiological signals or for diagnostic purposes
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/904Browsing; Visualisation therefor
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10Complex mathematical operations
    • G06F17/16Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10Complex mathematical operations
    • G06F17/18Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis

Definitions

  • the invention belongs to the field of dimensionality reduction processing of neuroscience calcium signals, and specifically relates to a dimensionality reduction and visualization method of high-throughput calcium signals based on Laplacian Eigenmaps (LE), which is recorded by neuroscience experimental calcium imaging technology.
  • L Laplacian Eigenmaps
  • the dynamic representation of neural activity provides an effective dimensionality reduction and visualization method.
  • neural cluster calcium signals In neuroscience research, calcium imaging technology is widely used in in vivo or in vitro experiments, which can simultaneously record changes in calcium ions in hundreds or thousands of neurons to monitor neuron clusters. activity. Due to the large number of recorded neurons, neural cluster calcium signals have the characteristics of high throughput, large amount of information, and complex content. Choosing an appropriate processing and analysis method for neural cluster calcium signals can lay a solid foundation for subsequent research. At present, some existing dimensionality reduction methods for neuronal electrical signals often perform poorly when processing calcium signals.
  • the present invention provides a method for dimensionality reduction and visualization of high-throughput calcium signals based on Laplacian eigenmapping, which reduces the dimensionality of neuron calcium signals through Laplacian eigenmapping, extracts effective information, and displays them in a low-dimensional space.
  • Draw dynamic images to intuitively represent the activity patterns of neural clusters.
  • S k (t) represents the value of the original data at time t
  • N represents the smoothing width
  • LE dimensionality reduction algorithm Apply Laplacian eigenmap (LE) to reduce the dimensionality of the processed data to obtain a low-dimensional representation of the high-dimensional calcium signal.
  • the smoothed data can use the K nearest neighbor algorithm to treat the firing data of each neuron in one trial as a data point, calculate its K nearest neighbors, and obtain the adjacency graph G.
  • the adjacency graph G calculates the weight matrix W between neural data.
  • the calculation formula is:
  • n is the number of neurons.
  • Visualization Visualize the dimensionality reduction results in a low-dimensional space, analyze the low-dimensional data obtained by dimensionality reduction, determine the value range of the coordinate axis based on the maximum and minimum values of each dimension, and establish a coordinate system. Draw the data points in the low-dimensional space coordinate system, combine the data labels of the experimental conditions, and color the data points to show the distribution of the data under different experimental conditions in the low-dimensional space after dimensionality reduction, and you can use points of multiple colors Characterizing distinct activity patterns of neuronal clusters.
  • the present invention can process larger and noisier multi-channel neural signals, and therefore has significant advantages in the dynamic visualization of high-throughput calcium nerve signals.
  • the multidimensional neural signal dimensionality reduction and visualization method based on Laplacian eigenmap is also better than other dimensionality reduction methods. Considering that Laplacian eigenmap has lower computational complexity Therefore, the present invention also has significant advantages in terms of running time.
  • the present invention designs a visualization method based on Laplacian eigenmap for high-throughput neuronal calcium signals.
  • Laplacian eigenmap is a dimensionality reduction algorithm that reflects the internal structure of a manifold. It can be used from high-dimensional Effective information is efficiently extracted from the original signal to facilitate researchers to analyze the distribution and dynamic characteristics of calcium signals. It provides a new signal processing and analysis tool for calcium imaging technology to study neural cluster encoding and decoding and neural circuit working mechanisms.
  • Figure 1 is a flow chart of the method of the present invention.
  • Figure 2 is the flow chart of the Laplacian feature mapping algorithm.
  • W is the weight matrix
  • is the Euclidean distance between the data points Xi and
  • the parameter of n is the number of neurons, D is the degree matrix, and L is the Laplacian matrix.
  • Figure 3 shows the results of visualizing the calcium signal in two-dimensional space after dimensionality reduction through Laplacian in the monkey cursor control experiment.
  • Example 1 A high-throughput calcium signal visualization method based on Laplacian eigenmapping
  • the method of the present invention includes the following steps:
  • S k (t) represents the value of the original data at time t
  • N represents the smoothing width
  • n is the number of neurons.
  • FIG. 1 is the flow chart of the Laplacian feature mapping method, step (2).
  • Visualization Visualize the dimensionality reduction results in a low-dimensional space, analyze the low-dimensional data obtained by dimensionality reduction, determine the value range of the coordinate axis based on the maximum and minimum values of each dimension, and establish a coordinate system. Draw the data points in the low-dimensional space coordinate system, combine the data labels of the experimental conditions, and color the data points to show the distribution of the data under different experimental conditions in the low-dimensional space after dimensionality reduction, and you can use points of multiple colors Characterizing distinct activity patterns of neuronal clusters.
  • This example combines the calcium signals of neural clusters in the motor cortex M1 and PMd brain areas recorded in the monkey cursor control experiment to describe in detail the dimensionality reduction and visualization method of high-throughput calcium signals based on Laplacian eigenmapping.
  • the Laplacian-based method is used to process up to M-dimensional neural data, visualize it in three-dimensional space, and draw the dynamic distribution of neural clusters.
  • the main steps are shown in Figure 1.
  • the specific implementation steps are as follows:
  • n is the number of neurons, because the data comes from the discharge signals of M neurons in the monkey motor cortex M1 and PMd brain areas, so n is set to M.
  • the dimensionality reduction and visualization method of high-throughput calcium signals based on Laplacian feature mapping constructed by the present invention provides dynamic visualization of hundreds of neuron cluster activities recorded in experiments such as neuroelectrophysiology and calcium signal imaging. result.
  • the neural activity patterns of the monkey moving the cursor to hit the same target are relatively close, while the neural activity patterns of hitting different targets are relatively far away.
  • the visual data well demonstrates the different dynamic characteristics of neural activity under the cursor control task.
  • the inter-class distance of data points is relatively large, the intra-class distance is relatively small, and the aggregation effect is good.
  • the present invention only took 0.04 seconds to process M-dimensional high-throughput calcium signal data. Compared with existing methods, it has significant advantages in running speed.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • Surgery (AREA)
  • Pathology (AREA)
  • Veterinary Medicine (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Molecular Biology (AREA)
  • Medical Informatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Algebra (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Optics & Photonics (AREA)
  • Operations Research (AREA)
  • Neurosurgery (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Neurology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physiology (AREA)
  • Psychiatry (AREA)

Abstract

本发明提供基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法,通过拉普拉斯特征映射对神经光学成像中记录到的高维钙信号进行降维,提取有效信息,并在低维空间中绘制出动态图像,实现对钙成像技术记录的神经集群活动模式的直观表征。与传统的方法例如广义线性回归模型相比,本发明可处理规模更大、信噪比更低的高通量钙信号,因此在处理大规模神经元集群的钙信号及其动态可视化方面具有明显优势。在聚类准确度和可视化图形区分度方面,基于拉普拉斯特征映射的高通量钙信号降维及可视化方法也优于其余降维方法,考虑到拉普拉斯特征映射有较低的计算复杂度,因此本发明在运行耗时方面也具有显著优势,可用于实时在线降维和可视化处理。

Description

基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法 技术领域
本发明属于神经科学钙信号的降维处理领域,具体涉及一种基于拉普拉斯特征映射(Laplacian Eigenmaps,LE)的高通量钙信号的降维和可视化方法,为神经科学实验钙成像技术记录的神经活动的动态表征提供一种有效的降维及可视化方法。
背景技术
在神经科学研究中,在体(in vivo)或者离体(in vitro)的实验,广泛应用了钙成像技术,能够同时记录成百上千个神经元内钙离子的变化,从而监测神经元集群的活动。由于记录的神经元数量庞大,神经集群钙信号具有高通量、大信息量以及内容复杂的特点,选择合适的神经集群钙信号的处理分析方法,能够为之后的研究奠定坚实的基础。目前,已有的一些针对神经元电信号的降维方法在处理钙信号时往往表现欠佳,例如通过测量峰电位计数向量出现的概率,形成离散的概率分布来表征集群活动,这种方法有着显著的缺陷:随着记录神经元数量的增加,峰电位计数向量呈指数增长,所以它仅仅适用于少量神经元,并且涉及大量的数据分析,并不适用于高通量的钙信号。因此,急需一种行之有效的信息约简方法,从高通量的、有噪声的原始数据最中,提取出有意义的成分,并用合适的可视化方法进行表示。
发明内容
本发明提供一种基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法,通过拉普拉斯特征映射对神经元钙信号进行降维,提取有效信息,并在低维空间中绘制出动态图像,实现对神经集群活动模式的直观表征。
本方法通过以下步骤实现:
(1)数据预处理:对钙信号进行预处理,采用均值平滑的方法,对每个时间区间的信号,根据相邻时间区间取均值。平滑后某个时间点数据的数学表达式为:
F k(t)=mean(S k(t-N+1),S k(t-N+2),…,S k(t))
式中S k(t)表示t时刻原始数据的值,N表示平滑宽度。
(2)LE降维算法:应用拉普拉斯特征映射(LE)对处理后的数据进行降维,得到高维钙信号的低维表示。
平滑过的数据可使用K近邻算法将每个神经元在一次试验中的发放数据作为一个数据点,计算它的K个最近邻,得到邻接图G。
其中邻接图G计算神经数据之间的权值矩阵W,计算公式为:
Figure PCTCN2022131368-appb-000001
其中||X i-X j||是数据点X i和X j之间的欧式距离,X i表示第i个神经元的活动的特征数据,σ是指数核函数的参数。
根据权值矩阵W计算相应的度矩阵D,计算公式为:
Figure PCTCN2022131368-appb-000002
其中n的值为神经元的个数。
计算拉普拉斯矩阵L,公式为L=D-W。其中D为度矩阵,W为权值矩阵。
对拉普拉斯矩阵L进行特征分解,截取最小的m(m为降维目标维度,为了达到可视化的目的,m不大于4)个非零特征值{λ 12,...,λ m}对应的特征向量[υ (1)(2),...,υ (m)]作为降维后的结果输出。
(3)可视化:将降维结果在低维空间可视化,分析降维得到的低维数据,根据每一维的最大值和最小值确定坐标轴的取值范围,建立坐标系。在低维空间坐标系中画出数据点,结合实验条件的数据标签,给数据点染色,展现降维后的不同实验条件下的数据在低维空间的分布,并可以用多种颜色的点表征神经元集群的不同活动模式。
本发明在与传统的方法例如广义线性回归模型(GLM)等方法相比,可以处理更大规模更嘈杂的多通道神经信号,因此在高通量神经钙信号的动态可视化方面具有较为显著的优势。在聚类准确度和可视化图形区分度方面,基于拉普拉斯特征映射的多维神经信号降维及可视化方法也优于其余降维方法,考虑到拉普拉斯特征映射有较低的计算复杂度,因此本发明在运行耗时方面也具有显著优势。
本发明为高通量的神经元钙信号设计了一种基于拉普拉斯特征映射的可视化方法,拉普拉斯特征映射是一种反映流形内部结构的降维算法,它可以从高维原始信号中高效的提取有效信息,以便于研究人员分析钙信号的分布特性和动态特性,为钙成像技术研究神经集群编解码和神经环路工作机制提供了一种崭新的信号处理和分析工具。
附图说明
图1为本发明方法流程图。
图2为拉普拉斯特征映射算法流程图。其中,W为权值矩阵,||X i-X j||是数据点X i和X j之间的欧式距离,X i表示第i个神经元的活动的特征数据,σ是指数核函数的参数,n的值为神经元的个数,D为度矩阵,L为拉普拉斯矩阵。
图3为猴子光标控制实验中,钙信号通过拉普拉斯降维后,在二维空间可视化结果。
具体实施方式
下面结合附图和实施例,对本发明基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法做详细描述。
实施例1 一种基于拉普拉斯特征映射的高通量钙信号可视化方法
参见图1该方法的流程图,本发明方法包括以下步骤:
(1)对神经信号进行预处理,采用均值平滑的方法,对每个时间区间的信号,根据相邻时间区间取均值。平滑后某个时间点数据的数学表达式为:
F k(t)=mean(S k(t-N+1),S k(t-N+2),…,S k(t))
式中S k(t)表示t时刻原始数据的值,N表示平滑宽度。
(2)应用拉普拉斯特征映射(LE)对处理后的数据进行降维,得到高维钙信号的低维表示。
在数据上构建邻接图G,其中数据之间的权值W,计算公式为:
Figure PCTCN2022131368-appb-000003
||X i-X j||是数据点X i和X j之间的欧式距离,X i表示第i个神经元的活动的特征数据,σ是指数核函数的参数。
根据权值矩阵W计算相应的度矩阵D,计算公式为:
Figure PCTCN2022131368-appb-000004
其中n的值为神经元的个数。
计算拉普拉斯矩阵L,公式为L=D-W。其中D为度矩阵,W为权值矩阵。
对拉普拉斯矩阵L进行特征分解,截取最小的m(m为降维目标维度,为了达到可视化的目的,m不大于4)个非零特征值{λ 12,...,λ m}对应的特征向量[υ (1)(2),...,υ (m)]作为降维后的结果输出。图2为拉普拉斯特征映射方法,即步骤(2)的流程图。
(3)可视化:将降维结果在低维空间可视化,分析降维得到的低维数据,根据每一维的最大值和最小值确定坐标轴的取值范围,建立坐标系。在低维空间坐标系中画出数据点,结合实验条件的数据标签,给数据点染色,展现降维后的不同实验条件下的数据在低维空间的分布,并可以用多种颜色的点表征神经元集群的不同活动模式。
实施例2 猴子光标控制实验中记录到的运动皮层神经集群钙信号的降维
本实例结合猴子光标控制实验中记录到的运动皮层M1和PMd脑区神经集群钙信号,对基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法做详细描述。利用基于拉普拉 斯的方法,处理多达M维的神经数据,将其在三维空间可视化,绘制出神经集群动态分布。其主要步骤参见图1。具体实施步骤如下:
(1)光标控制实验中,猴子移动光标分别击中位于左右两个不同位置的目标,同时记录到猴子运动皮层M1和PMd脑区的M个神经元的活动(例如M=639),原始数据可表示为M个神经元在N次试验中的钙信号(例如N=364)。
(2)在实验数据点之间构建邻接图G,可使用K近邻算法,如果K取值12,即将每个神经元在一次试验中的发放数据作为一个数据点,计算它的12个最近邻并与之相连,建立邻接图G。
(3)根据邻接图G,确定神经数据点之间的权值,得到权值矩阵W。G中相邻点之间权值的计算公式为:
Figure PCTCN2022131368-appb-000005
||X i-X j||是数据点X i和X j之间的欧式距离,X i表示第i个神经元的活动的特征数据,σ是指数核函数的参数,在实验中一般设置为1。
(4)根据权值矩阵W计算相应的度矩阵D,计算公式为:
Figure PCTCN2022131368-appb-000006
其中n的值为神经元的个数,因为数据来自猴子运动皮层M1和PMd脑区M个神经元的放电信号,n设置为M。
(5)计算拉普拉斯矩阵L,公式为:L=D-W,其中D为步骤(4)中计算出的度矩阵,W为步骤(3)中计算出的权值矩阵W。对拉普拉斯矩阵L进行特征分解,截取最小的2个非零特征值{λ 12}对应的特征向量[υ (1)(2)]作为降维后的结果输出。步骤(2)~(5)即为拉普拉斯特征映射的流程图,参见图2。
(6)将以上降维得到的2维数据,根据每一维的最大值和最小值确定坐标轴的取值范围,建立坐标系。在坐标系中画出每一个数据点,在猴子光标控制实验中,猴子会移动位于二维屏幕上的光标,击中分别位于左右两个不同位置的目标,我们用不同颜色(黑色、灰色)染色数据点,从而对不同目标下的神经活动加以区分。每一次光标控制中记录到的神经钙信号可以用二维空间的数据点加以表征。在图3中,我们利用本发明所构建的拉普拉斯特征映射方法对猴子N次光标控制试验过程中的记录到的M个神经集元活动的数据进行了降维处理,并在2维空间上做了动态可视化的展示。
本发明构建的基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法为神经电 生理、钙信号成像等实验中记录到的成百上千个神经元集群活动提供了动态的可视化结果。图3中猴子移动光标击中相同目标的神经活动模式相对靠近,击中不同目标的神经活动模式相对远离,可视化数据很好地展现了光标控制任务下神经活动不同的动态特性。降维后的数据点类间距离相对较大,类内距离相对较小,聚合效果较好。在猴子光标控制实验中,本发明处理M维的高通量钙信号数据仅耗时0.04秒,相比于现有方法,在运行速率上有着显著优势。

Claims (7)

  1. 一种基于拉普拉斯特征映射的高通量钙信号降维和可视化方法,其特征在于,通过以下步骤实现:
    (1)数据预处理:对记录到高维度钙信号进行预处理,采用均值平滑的方法,对每个时间区间的信号,根据相邻时间区间取均值。平滑后某个时间点数据的数学表达式为:
    F k(t)=mean(S k(t-N+1),S k(t-N+2),…,S k(t))
    式中S k(t)表示t时刻原始数据的值,N表示平滑宽度。
    (2)拉普拉斯特征映射降维算法:应用拉普拉斯特征映射(LE)对处理后的神经数据进行降维,得到原始神经数据的低维表示;
    (3)可视化:结合实验条件的数据标签,绘制出其在低维空间的数据点,展现降维后的数据在低维空间的动态变化,并可以用多种颜色表征神经元集群在不同实验条件下的不同活动模式。
  2. 根据权利要求1所述的一种基于拉普拉斯特征映射的高通量钙信号可视化方法,其特征在于,步骤(2)平滑过的钙信号可使用K近邻算法将每个神经元在一次试验中的发放数据作为一个数据点,计算它的K个最近邻,得到邻接图G。
  3. 根据权利要求1所述的一种基于拉普拉斯特征映射的高通量钙信号可视化方法,其特征在于,步骤(2)中邻接图G计算神经数据之间的权值矩阵W,计算公式为
    Figure PCTCN2022131368-appb-100001
    其中||X i-X j||是数据点X i和X j之间的欧式距离,X i表示第i个神经元的活动的特征数据,σ是指数核函数的参数。
  4. 根据权利要求1所述的一种基于拉普拉斯特征映射的高通量钙信号可视化方法,其特征在于,步骤(2)中根据权值矩阵W计算相应的度矩阵D,计算公式为:
    Figure PCTCN2022131368-appb-100002
    其中n的值为神经元的个数。
  5. 根据权利要求1所述的一种基于拉普拉斯特征映射的高通量钙信号可视化方法,其特征在于,步骤(2)中计算拉普拉斯矩阵L,公式为L=D-W,其中D为度矩阵,W为权值矩阵。
  6. 根据权利要求5所述的一种基于拉普拉斯特征映射的多维神经信号可视化方法,其特征在于,其中对拉普拉斯矩阵L进行特征分解,截取最小的m个非零特征值{λ 12,...,λ m}对应的特征向量[υ (1)(2),...,υ (m)]作为降维后的结果输出。m为降维目标维度,为了达到可视化的目的,m不大于4。
  7. 根据权利要求1所述的一种基于拉普拉斯特征映射的高通量钙信号可视化方法,其特征在于,将降维结果在低维空间可视化:分析降维得到的低维数据,根据每一维的最大值和最小值确定坐标轴的取值范围,建立坐标系,在低维空间坐标系中画出数据点,结合实验条件的数据标签,给数据点染色,展现降维后的不同实验条件下的数据在低维空间的分布,并可以用多种颜色的点表征神经元集群的不同活动模式。
PCT/CN2022/131368 2022-03-15 2022-11-11 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法 WO2023173770A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210254488.5 2022-03-15
CN202210254488.5A CN115034253A (zh) 2022-03-15 2022-03-15 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法

Publications (1)

Publication Number Publication Date
WO2023173770A1 true WO2023173770A1 (zh) 2023-09-21

Family

ID=83119927

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/131368 WO2023173770A1 (zh) 2022-03-15 2022-11-11 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法

Country Status (2)

Country Link
CN (1) CN115034253A (zh)
WO (1) WO2023173770A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115034253A (zh) * 2022-03-15 2022-09-09 浙江大学 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109033415A (zh) * 2018-08-06 2018-12-18 浙江大学 一种基于拉普拉斯特征映射的多维神经信号的降维和可视化方法
CN112763225A (zh) * 2020-12-25 2021-05-07 哈尔滨工业大学 一种基于拉普拉斯特征映射算法的传感器误差识别及消除方法
CN115034253A (zh) * 2022-03-15 2022-09-09 浙江大学 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109033415A (zh) * 2018-08-06 2018-12-18 浙江大学 一种基于拉普拉斯特征映射的多维神经信号的降维和可视化方法
CN112763225A (zh) * 2020-12-25 2021-05-07 哈尔滨工业大学 一种基于拉普拉斯特征映射算法的传感器误差识别及消除方法
CN115034253A (zh) * 2022-03-15 2022-09-09 浙江大学 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
SUN, GUANGHAO ET AL.: "Effective Dimensionality Reduction for Visualizing Neural Dynamics by Laplacian Eigenmaps", NEURAL COMPUTATION, vol. 31, no. 7, 1 July 2019 (2019-07-01), pages 1356 - 1379, XP009548787 *

Also Published As

Publication number Publication date
CN115034253A (zh) 2022-09-09

Similar Documents

Publication Publication Date Title
Fu et al. Improvement motor imagery EEG classification based on regularized linear discriminant analysis
YongHua et al. Study on the identification of the wood surface defects based on texture features
Gulhane et al. Diagnosis of diseases on cotton leaves using principal component analysis classifier
CN106096547B (zh) 一种面向识别的低分辨率人脸图像特征超分辨率重建方法
CN106778687B (zh) 基于局部评估和全局优化的注视点检测方法
Wang et al. GKFC-CNN: Modified Gaussian kernel fuzzy C-means and convolutional neural network for apple segmentation and recognition
WO2023173770A1 (zh) 基于拉普拉斯特征映射的高通量钙信号的降维和可视化方法
CN109033415A (zh) 一种基于拉普拉斯特征映射的多维神经信号的降维和可视化方法
CN106529441B (zh) 基于模糊边界分片的深度动作图人体行为识别方法
Islam et al. Analysis of CT scan images to predict lung cancer stages using image processing techniques
Liang et al. Automatic defect detection of texture surface with an efficient texture removal network
Dai et al. PPLC-Net: Neural network-based plant disease identification model supported by weather data augmentation and multi-level attention mechanism
CN107045624B (zh) 一种基于最大加权团的脑电信号预处理与分类方法
Huang et al. Retinal vessel segmentation using simple SPCNN model and line connector
Lin et al. Identification of pumpkin powdery mildew based on image processing PCA and machine learning
Ihsanuddin et al. Identification of Bacterial Leaf Blight and Brown Spot Disease In Rice Plants With Image Processing Approach
CN109241932B (zh) 一种基于运动方差图相位特征的热红外人体动作识别方法
CN105488798B (zh) 基于点集对比的sar图像相似性度量方法
Sivakamasundari et al. Classification of leaf diseases in apple using support vector machine
Qiao Application of Gabor image recognition technology in intelligent clothing design
CN111127407B (zh) 一种基于傅里叶变换的风格迁移伪造图像检测装置及方法
CN110070538B (zh) 基于形态优化深度特征的螺栓二维视觉结构聚类方法
Liang et al. Research on Improved K-means Image Segmentation Algorithm Based on HSV Space
CN110929734A (zh) 一种基于双侧非对称感受野机制的显著轮廓提取方法
Hou et al. Visual detection of Parkinson’s disease via facial features recognition

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22931783

Country of ref document: EP

Kind code of ref document: A1