Unsupervised spike sorting based on discriminative subspace learning.

Objective. Spike sorting is a fundamental preprocessing step for many neuroscience studies which rely on the analysis of spike trains. Most of the feature extraction and dimensionality reduction techniques that have been used for spike sorting give a projection subspace which is not necessarily the most discriminative one. Therefore, the clusters which appear inherently separable in some discriminative subspace may overlap if projected using conventional feature extraction approaches leading to a poor sorting accuracy especially when the noise level is high. In this paper, we propose a noise-robust and unsupervised spike sorting algorithm based on learning discriminative spike features for clustering. Approach. The proposed algorithm uses discriminative subspace learning to extract low dimensional and most discriminative features from the spike waveforms and perform clustering with automatic detection of the number of the clusters. The core part of the algorithm involves iterative subspace selection using linear discriminant analysis and clustering using Gaussian mixture model with outlier detection. A statistical test in the discriminative subspace is proposed to automatically detect the number of the clusters. Main results. Comparative results on publicly available simulated and real in vivo datasets demonstrate that our algorithm achieves substantially improved cluster distinction leading to higher sorting accuracy and more reliable detection of clusters which are highly overlapping and not detectable using conventional feature extraction techniques such as principal component analysis or wavelets. Significance. By providing more accurate information about the activity of more number of individual neurons with high robustness to neural noise and outliers, the proposed unsupervised spike sorting algorithm facilitates more detailed and accurate analysis of single- and multi-unit activities in neuroscience and brain machine interface studies.

Discriminative sparse subspace learning and its application to unsupervised feature selection

In traditional machine learning algorithms, the classification models are learned on the training data (source domain) to reuse for labelling the test data (target domain) where the training and test samples are from the same distributions. However in nowadays applications, the existence of distribution shift across the source and target doamins degrades the model performance, significantly. Domain adaptation methods have been proposed to compensate domain shift problem by aligning the distributions across the source and target domains under various adaptation strategies. This paper addresses the robust image classification problem for unsupervised domain adaptation. Specifically, following three methods are proposed: Discriminative Subspace Learning (DSL), Joint Geometrical and Statistical Distribution Adaptation (GSDA), and Joint Subspace and Distribution Adaptation (DSL-GSDA). DSL is a subspace centric method that aligns the specific and shared features across domains. Indeed, DSL finds two projections to map the source and target data into independent subspaces by aligning the discriminant and global structures of domains. GSDA trends to find an adaptive classifier through statistical and geometrical distribution alignment and minimizes the prediction error. DSL-GSDA, as a combination of DSL and GSDA, consists of two subspace and distribution adaptation levels. DSL-GSDA uses DSL to build two aligned subspaces of source and target domains. The distributions of source and target data in new subspaces is adapted via GSDA. The proposed methods are evaluated on benchmark visual datasets for object, digit and face recongnition tasks. Visual datasets consist of image domains that have been captured under various real-world conditions where the domain shift is unavoidable. The experiment results show that DSL, GSDA and DSL-GSDA outperform other state-of-the-art domain adaptation methods by 6.19%, 1.48% and 1.99% improvement, respectively. Our source code is available at https://github.com/jtahmores/DSLGSDA (https://github.com/jtahmores/DSLGSDA).

This study introduces a new spike sorting method that classifies spike waveforms from multiunit recordings into spike trains of individual neurons. In particular, we develop a method to sort a spike mixture generated by a heterogeneous neural population. Such a spike sorting has a significant practical value, but was previously difficult. The method combines a feature extraction method, which we may term “multimodality-weighted principal component analysis” (mPCA), and a clustering method by variational Bayes for Student's t mixture model (SVB). The performance of the proposed method was compared with that of other conventional methods for simulated and experimental data sets. We found that the mPCA efficiently extracts highly informative features as clusters clearly separable in a relatively low-dimensional feature space. The SVB was implemented explicitly without relying on Maximum-A-Posterior (MAP) inference for the “degree of freedom” parameters. The explicit SVB is faster than the conventional SVB derived with MAP inference and works more reliably over various data sets that include spiking patterns difficult to sort. For instance, spikes of a single bursting neuron may be separated incorrectly into multiple clusters, whereas those of a sparsely firing neuron tend to be merged into clusters for other neurons. Our method showed significantly improved performance in spike sorting of these “difficult” neurons. A parallelized implementation of the proposed algorithm (EToS version 3) is available as open-source code at http://etos.sourceforge.net/.

Objective. This paper discusses a novel method for automating the curation of neural spike events detected from neural recordings using spike sorting methods. Spike sorting seeks to identify isolated neural events from extracellular recordings. This is critical for interpretation of electrophysiology recordings in neuroscience studies. Spike sorting analysis is vulnerable to errors because of non-neural events, such as experimental artifacts or electrical interference. To improve the specificity of spike sorting results, a manual postprocessing curation is typically used to examine the detected events and identify neural spikes based on their specific features. However, this manual curation process is subjective, prone to human errors and not scalable, especially for large datasets.Approach. To address these challenges, we introduce AECuration, a novel automatic curation method based on an autoencoder model trained on features of simulated extracellular spike waveforms. Using reconstruction error as a performance metric, our method classifies neural and non-neural events in experimental electrophysiology datasets.Main results. This paper demonstrates that AECuration can classify neural events with 97.46% accuracy on synthetic datasets. Moreover, our method can improve the sensitivity of different spike sorting pipelines on datasets with ground-truth recordings by up to 20%. The ratio of clustered units with low interspike interval violation rates is improved from 55.3% to 85.5% as demonstrated using our in-house experimental dataset.Significance. AEcuration is a time-domain evaluation method that automates the analysis of extracellular recordings based on learned time-domain features. Once trained on a synthetic dataset, this method can be applied to real extracellular datasets without the need for re-training. This highlights the generalizability of AECuration. It can be readily integrated with existing spike sorting pipelines as a preprocessing filtering or a postprocessing curation step to improve the overall accuracy and efficiency.

Subspace learning has been widely utilized to extract discriminative features for classification task, such as face recognition, even when facial images are occluded or corrupted. However, the performance of most existing methods would be degraded significantly in the scenario of that data being contaminated with severe noise, especially when the magnitude of the gross corruption can be arbitrarily large. To this end, in this paper, a novel discriminative subspace learning method is proposed based on the well-known low-rank representation (LRR). Specifically, a discriminant low-rank representation and the projecting subspace are learned simultaneously, in a supervised way. To avoid the deviation from the original solution by using some relaxation, we adopt the Schatten [Formula: see text]-norm and [Formula: see text]-norm, instead of the nuclear norm and [Formula: see text]-norm, respectively. Experimental results on two famous databases, i.e. PIE and ORL, demonstrate that the proposed method achieves better classification scores than the state-of-the-art approaches.

Improving Spike Sorting Efficiency with Separability Index and Spectral Clustering

Objective. Detection and sorting (classification) of action potentials from extracellular recordings are two important pre-processing steps for brain–computer interfaces (BCIs) and some neuroscientific studies. Traditional approaches perform these two steps serially, but using shapes of action potential waveforms during detection, i.e. combining the two steps, may lead to better performance, especially during high noise. We propose a hidden Markov model (HMM) based method for combined detecting and sorting of spikes, with the aim of improving the final decoding accuracy of BCIs. Approach. The states of the HMM indicate whether there is a spike, what unit a spike belongs to, and the time course within a waveform. The HMM outputs probabilities of spike detection, and from this we can calculate expectations of spike counts in time bins, which can replace integer spike counts as input to BCI decoders. We evaluate the HMM method on simulated spiking data. We then examine the impact of using this method on decoding real neural data recorded from primary motor cortex of two Rhesus monkeys. Main results. Our comparisons on simulated data to detection-then-sorting approaches and combined detection-and-sorting algorithms indicate that the HMM method performs more accurately at detection and sorting (0.93 versus 0.73 spike count correlation, 0.73 versus 0.49 adjusted mutual information). On real neural data, the HMM method led to higher adjusted mutual information between spike counts and kinematics (monkey K: 0.034 versus 0.027; monkey M: 0.033 versus 0.022) and better neuron encoding model predictions (K: 0.016 dB improvement; M: 0.056 dB improvement). Lastly, the HMM method facilitated higher offline decoding accuracy (Kalman filter, K: 8.5% mean squared error reduction, M: 18.6% reduction). Significance. The HMM spike detection and sorting method offers a new approach to spike pre-processing for BCIs and neuroscientific studies.

Spike sorting is a fundamental step in extracting single-unit activity from neural ensemble recordings, which play an important role in basic neuroscience and neurotechnologies. A few algorithms have been applied in spike sorting. However, when noise level or waveform similarity becomes relatively high, their robustness still faces a big challenge. In this study, we propose a spike sorting method combining Linear Discriminant Analysis (LDA) and Density Peaks (DP) for feature extraction and clustering. Relying on the joint optimization of LDA and DP: DP provides more accurate classification labels for LDA, LDA extracts more discriminative features to cluster for DP, and the algorithm achieves high performance after iteration. We first compared the proposed LDA-DP algorithm with several algorithms on one publicly available simulated dataset and one real rodent neural dataset with different noise levels. We further demonstrated the performance of the LDA-DP method on a real neural dataset from non-human primates with more complex distribution characteristics. The results show that our LDA-DP algorithm extracts a more discriminative feature subspace and achieves better cluster quality than previously established methods in both simulated and real data. Especially in the neural recordings with high noise levels or waveform similarity, the LDA-DP still yields a robust performance with automatic detection of the number of clusters. The proposed LDA-DP algorithm achieved high sorting accuracy and robustness to noise, which offers a promising tool for spike sorting and facilitates the following analysis of neural population activity.

Discriminative sparse subspace learning with manifold regularization

In order to determine patterns of neural activity, spike signals recorded by extracellular electrodes have to be clustered (sorted) with the aim of ensuring that each cluster represents all the spikes generated by an individual neuron. Many methods for spike sorting have been proposed but few are easily applicable to recordings from polytrodes which may have 16 or more recording sites. As with tetrodes, these are spaced sufficiently closely that signals from single neurons will usually be recorded on several adjacent sites. Although this offers a better chance of distinguishing neurons with similarly shaped spikes, sorting is difficult in such cases because of the high dimensionality of the space in which the signals must be classified. This report details a method for spike sorting based on a divide and conquer approach. Clusters are initially formed by assigning each event to the channel on which it is largest. Each channel-based cluster is then sub-divided into as many distinct clusters as possible. These are then recombined on the basis of pairwise tests into a final set of clusters. Pairwise tests are also performed to establish how distinct each cluster is from the others. A modified gradient ascent clustering (GAC) algorithm is used to do the clustering. The method can sort spikes with minimal user input in times comparable to real time for recordings lasting up to 45 min. Our results illustrate some of the difficulties inherent in spike sorting, including changes in spike shape over time. We show that some physiologically distinct units may have very similar spike shapes. We show that RMS measures of spike shape similarity are not sensitive enough to discriminate clusters that can otherwise be separated by principal components analysis (PCA). Hence spike sorting based on least-squares matching to templates may be unreliable. Our methods should be applicable to tetrodes and scalable to larger multi-electrode arrays (MEAs).

Images, as high-dimensional data, usually embody large variabilities. To classify images for versatile applications, an effective algorithm is necessarily designed by systematically considering the data structure, similarity metric, discriminant subspace, and classifier. In this paper, we provide evidence that, besides the Fisher criterion, graph embedding, and tensorization used in many existing methods, the correlation-based similarity metric embodied in supervised multilinear discriminant subspace learning can additionally improve the classification performance. In particular, a novel discriminant subspace learning algorithm, called correlation tensor analysis (CTA), is designed to incorporate both graph-embedded correlational mapping and discriminant analysis in a Fisher type of learning manner. The correlation metric can estimate intrinsic angles and distances for the locally isometric embedding, which can deal with the case when Euclidean metric is incapable of capturing the intrinsic similarities between data points. CTA learns multiple interrelated subspaces to obtain a low-dimensional data representation reflecting both class label information and intrinsic geometric structure of the data distribution. Extensive comparisons with most popular subspace learning methods on face recognition evaluation demonstrate the effectiveness and superiority of CTA. Parameter analysis also reveals its robustness.

Human age estimation from facial images has become an active research topic in computer vision field because of various real-world applications. Temporal property of facial aging display sequential patterns that lie on the low-dimensional aging manifold. In this paper, we propose hidden factor analysis (HFA) model-based discriminative manifold learning method for age estimation. The hidden factor analysis decomposes facial features into independent age factor and identity factor. Various age invariant face recognition systems in the literature utilize identity factor for face recognition; however, the age factor remains unutilized. The age component of the hidden factor analysis model depends on the subject’s age. Thus it carries significant age-related information. In this paper, we demonstrate that such aging patterns can be effectively extracted from the HFA-based discriminant subspace learning algorithm. Next, we have applied multiple regression methods on low-dimensional aging features learned from the HFA model. Effect of reduced dimensionality on the accuracy has been evaluated by extensive experiments and compared with the state-of-the-art methods. Effectiveness and robustness in terms of MAE and CS of the proposed framework are demonstrated using experimental analysis on a large-scale aging database MORPH II. The accuracy of our method is found superior to the current state-of-the-art methods.

Recently, cross-view feature learning has been a hot topic in machine learning due to the wide applications of multiview data. Nevertheless, the distribution discrepancy between cross-views leads to the fact that instances of the different views from same class are farther than those within the same view but from different classes. To address this problem, in this paper, we develop a novel cross-view discriminative feature subspace learning method inspired by layered visual perception from human. Firstly, the proposed method utilizes a separable low-rank self-representation model to disentangle the class and view structure layers, respectively. Secondly, a local alignment is constructed with two designed graphs to guide the subspace decomposition in a pairwise way. Finally, the global discriminative constraint on distribution center in each view is designed for further alignment improvement. Extensive cross-view classification experiments on several public datasets prove that our proposed method is more effective than other existing feature learning methods.

Spike sorting using locality preserving projection with gap statistics and landmark-based spectral clustering.