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

Nguyen, Thanh; Khosravi, Abbas; Creighton, Douglas; Nahavandi, Saeid

doi:10.1016/j.jneumeth.2014.09.011

Cited by 33 publications

(32 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…That is, the i, jth entry of A (A ij ) was computed as the Euclidean distance between preprocessed data samples i and j (Nguyen et al. ). We then used the spectral clustering implementation in the Python package “scikit‐learn” to determine the cluster assignments.…”

Section: Methodsmentioning

confidence: 99%

“…After aligning the resulting samples, we calculated the Euclidean distance between all pairs of samples in the data set. That is, the i, jth entry of A (A ij ) was computed as the Euclidean distance between preprocessed data samples i and j (Nguyen et al 2014). We then used the spectral clustering implementation in the Python package "scikit-learn" to determine the cluster assignments.…”

Section: Clustering Analysesmentioning

confidence: 99%

See 1 more Smart Citation

Properties of neurons in the superficial laminae of trigeminal nucleus caudalis

et al. 2019

View full text Add to dashboard Cite

The trigeminal nucleus caudalis ( TN c) receives extensive afferent innervation from peripheral sensory neurons of the trigeminal ganglion (TG), and is the first central relay in the circuitry underpinning orofacial pain. Despite the initial characterization of the neurons in the superficial laminae, many questions remain. Here we report on electrophysiological properties of 535 superficial lamina I/ II TN c neurons. Based on their firing pattern, we assigned these cells to five main groups, including (1) tonic, (2) phasic, (3) delayed, (4) H‐current, and (5) tonic‐phasic neurons, groups that exhibit distinct intrinsic properties and share some similarity with groups identified in the spinal dorsal horn. Driving predominantly nociceptive TG primary afferents using optogenetic stimulation in TRPV 1/ChR2 animals, we found that tonic and H‐current cells are most likely to receive pure monosynaptic input, whereas delayed neurons are more likely to exhibit inputs that appear polysynaptic. Finally, for the first time in TN c neurons, we used unsupervised clustering analysis methods and found that the kinetics of the action potentials and other intrinsic properties of these groups differ significantly from one another. Unsupervised spectral clustering based solely on a single voltage response to rheobase current was sufficient to group cells with shared properties independent of action potential discharge pattern, indicating that this approach can be effectively applied to identify functional neuronal subclasses. Together, our data illustrate that cells in the TN c with distinct patterns of TRPV 1/ChR2 afferent innervation are physiologically diverse, but can be understood as a few major groups of cells having shared functional properties.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Clustering Analysesmentioning

confidence: 99%

Properties of neurons in the superficial laminae of trigeminal nucleus caudalis

et al. 2019

View full text Add to dashboard Cite

show abstract

“…The variation of accuracy with different template waveforms by fastDTW is the smallest. With fastDTW, the biphasic spikes demonstrate slightly higher R TP than the monophasic spikes (Nguyen et al (2014)). …”

Section: Resultes and Discussionmentioning

confidence: 97%

A real-time spike classification method based on dynamic time warping for extracellular enteric neural recording with large waveform variability

Cao

Rakhilin

Gordon

et al. 2016

Journal of Neuroscience Methods

View full text Add to dashboard Cite

Background Computationally efficient spike recognition methods are required for real-time analysis of extracellular neural recordings. The enteric nervous system (ENS) is important to human health but less well-understood with few appropriate spike recognition algorithms due to large waveform variability. New Method Here we present a method based on dynamic time warping (DTW) with high tolerance to variability in time and magnitude. Adaptive temporal gridding for fastDTW in similarity calculation significantly reduces the computational cost. The automated threshold selection allows for real-time classification for extracellular recordings. Results Our method is first evaluated on synthesized data at different noise levels, improving both classification accuracy and computational complexity over the conventional cross-correlation based template-matching method (CCTM) and PCA + k-means clustering without time warping. Our method is then applied to analyze the mouse enteric neural recording with mechanical and chemical stimuli. Successful classification of biphasic and monophasic spikes is achieved even when the spike variability is larger than millisecond in width and millivolt in magnitude. Comparison with Existing Method(s) In comparison with conventional template matching and clustering methods, the fastDTW method is computationally efficient with high tolerance to waveform variability. Conclusions We have developed an adaptive fastDTW algorithm for real-time spike classification of ENS recording with large waveform variability against colony motility, ambient changes and cellular heterogeneity.

show abstract

“…Whereas the simultaneous use of positive and negative thresholds reduces the number of misses by about a factor of two, it increases the number of false positives by about a factor of eight in the present data set (results not shown). Perhaps because of this reason, either the positive or the negative threshold alone is used in virtually all studies that use Wave clus for spike sorting [6,13,24]. Because a positive threshold may be used instead of a negative one simply by using the negative of x as the input data, the algorithms considered here will be explained under the assumption that the positive threshold is used for spike detection.…”

Section: Spike Detection Algorithm Of Wave Clusmentioning

confidence: 99%

“…The clustering can be performed in a variety of feature spaces spanned by features such as peak or valley amplitude, principal components, or wavelet coefficients. In addition to these commercial or widely used tools, new algorithms for performing spike sorting continue to be developed [5][6][7][8][9][10][11][12].…”

Section: Introductionmentioning

confidence: 99%

Taller Peaks: an improved spike detection algorithm that simultaneously reduces type I and type II errors for Wave{\_}clus

Okatan¹

2017

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

Abstract:Wave clus is an unsupervised spike detection and sorting algorithm that has been used in dozens of experimental studies as a spike sorting tool. It is often used as a benchmark for comparing the performance of new spike sorting algorithms. For these reasons, the spike detection performance of Wave clus is important for both experimental and computational studies that involve spike sorting. Two measures of spike detection performance are the number of false positive detections (type I error) and the number of missed spikes (type II error). Here, a new spike detection algorithm is proposed that reduces the number of misses and false positives of Wave clus in a widely used simulated data set across the entire range of commonly used detection thresholds. The algorithm accepts a spike if its amplitude is larger than the amplitude of its two immediate neighbors, where an immediate neighbor is the nearest peak of the same polarity within ± 1 refractory period. The simultaneous reduction that is achieved in the number of false positives and misses is important for experimental and computational studies that use Wave clus as a spike sorting tool or as a benchmark. A software patch that incorporates the algorithm into Wave clus as an optional spike detection algorithm is provided.

show abstract

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

Cited by 33 publications

References 27 publications

Properties of neurons in the superficial laminae of trigeminal nucleus caudalis

Properties of neurons in the superficial laminae of trigeminal nucleus caudalis

A real-time spike classification method based on dynamic time warping for extracellular enteric neural recording with large waveform variability

Taller Peaks: an improved spike detection algorithm that simultaneously reduces type I and type II errors for Wave{\_}clus

Contact Info

Product

Resources

About