Neuronal classification from network connectivity via adjacency spectral embedding.
Neuronal classification from network connectivity via adjacency spectral embedding.
复制标题
DOI：
10.1162/netn_a_00195
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Ascoli GA
中科院分区：
文献类型：
--
作者：
Mehta K;Goldin RF;Marchette D;Vogelstein JT;Priebe CE;Ascoli GA
这项工作提出了一种新的神经元分类策略，由有向图的节点表示，基于它们的电路（边缘连接）。我们假设一个随机块模型（SBM），如果神经元按照相同的概率分布连接到其他组的神经元，那么它们就属于同一组。在SBM图的邻接谱嵌入之后，我们导出了类的数量，并使用基于高斯混合模型的期望最大化（EM）聚类算法将每个神经元分配到一个类。为了提高精度，我们引入了一个简单的变量，使用随机分层聚集聚类初始化EM算法，并在多次EM重启中选择最佳解决方案。我们在一个大型（≈212-215个神经元）、稀疏、具有8个神经元类别的生物启发连接组上测试了该过程。仿真结果表明，该方法对嵌入维数的选择具有广泛的稳定性，并且随着网络中神经元数量的增加具有很好的扩展性。聚类精度对模型参数的变化具有鲁棒性，对模拟实验噪声具有很高的耐受性，可以在高达40%的交换边情况下实现完美的分类。因此，这种方法可能有助于从潜在细胞成分的角度分析和解释大规模脑连接组学数据。
This work presents a novel strategy for classifying neurons, represented by nodes of a directed graph, based on their circuitry (edge connectivity). We assume a stochastic block model (SBM) in which neurons belong together if they connect to neurons of other groups according to the same probability distributions. Following adjacency spectral embedding of the SBM graph, we derive the number of classes and assign each neuron to a class with a Gaussian mixture model-based expectation maximization (EM) clustering algorithm. To improve accuracy, we introduce a simple variation using random hierarchical agglomerative clustering to initialize the EM algorithm and picking the best solution over multiple EM restarts. We test this procedure on a large (≈212–215 neurons), sparse, biologically inspired connectome with eight neuron classes. The simulation results demonstrate that the proposed approach is broadly stable to the choice of embedding dimension, and scales extremely well as the number of neurons in the network increases. Clustering accuracy is robust to variations in model parameters and highly tolerant to simulated experimental noise, achieving perfect classifications with up to 40% of swapped edges. Thus, this approach may be useful to analyze and interpret large-scale brain connectomics data in terms of underlying cellular components.
登录
查看更多内容
影响因子：
64.8
作者：
Bock, Davi D.;Lee, Wei-Chung Allen;Kerlin, Aaron M.;Andermann, Mark L.;Hood, Greg;Wetzel, Arthur W.;Yurgenson, Sergey;Soucy, Edward R.;Kim, Hyon Suk;Reid, R. Clay
通讯作者：
Reid, R. Clay
影响因子：
48
作者：
Craddock, R. Cameron;Jbabdi, Saad;Yan, Chao-Gan;Vogelstein, Joshua T.;Castellanos, F. Xavier;Di Martino, Adriana;Kelly, Clare;Heberlein, Keith;Colcombe, Stan;Milham, Michael P.
通讯作者：
Milham, Michael P.
影响因子：
48
作者：
Chung, Kwanghun;Deisseroth, Karl
通讯作者：
Deisseroth, Karl
影响因子：
5.7
作者：
Magnain C;Augustinack JC;Reuter M;Wachinger C;Frosch MP;Ragan T;Akkin T;Wedeen VJ;Boas DA;Fischl B
通讯作者：
Fischl B
影响因子：
34.7
作者：
Ascoli, Giorgio A.;Alonso-Nanclares, Lidia;Anderson, Stewart A.;Barrionuevo, German;Benavides-Piccione, Ruth;Burkhalter, Andreas;Buzsaki, Gyoergy;Cauli, Bruno;DeFelipe, Javier;Fairen, Alfonso;Feldmeyer, Dirk;Fishell, Gord;Fregnac, Yves;Freund, Tamas F.;Gardner, Daniel;Gardner, Esther P.;Goldberg, Jesse H.;Helmstaedter, Moritz;Hestrin, Shaul;Karube, Fuyuki;Kisvarday, Zoltan F.;Lambolez, Bertrand;Lewis, David A.;Marin, Oscar;Markram, Henry;Munoz, Alberto;Packer, Adam;Petersen, Carl C. H.;Rockland, Kathleen S.;Rossier, Jean;Rudy, Bernardo;Somogyi, Peter;Staiger, Jochen F.;Tamas, Gabor;Thomson, Alex M.;Toledo-Rodriguez, Maria;Wang, Yun;West, David C.;Yuste, Rafael
通讯作者：
Yuste, Rafael