Limits...
Fast, scalable, Bayesian spike identification for multi-electrode arrays.

Prentice JS, Homann J, Simmons KD, Tkačik G, Balasubramanian V, Nelson PC - PLoS ONE (2011)

Bottom Line: Our method can distinguish large numbers of distinct neural units, even when spikes overlap, and accounts for intrinsic variability of spikes from each unit.Human interaction plays a key role in our method; but effort is minimized and streamlined via a graphical interface.We illustrate our method on data from guinea pig retinal ganglion cells and document its performance on simulated data consisting of spikes added to experimentally measured background noise.

View Article: PubMed Central - PubMed

Affiliation: Department of Physics and Astronomy, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America. jprentic@sas.upenn.edu

ABSTRACT
We present an algorithm to identify individual neural spikes observed on high-density multi-electrode arrays (MEAs). Our method can distinguish large numbers of distinct neural units, even when spikes overlap, and accounts for intrinsic variability of spikes from each unit. As MEAs grow larger, it is important to find spike-identification methods that are scalable, that is, the computational cost of spike fitting should scale well with the number of units observed. Our algorithm accomplishes this goal, and is fast, because it exploits the spatial locality of each unit and the basic biophysics of extracellular signal propagation. Human interaction plays a key role in our method; but effort is minimized and streamlined via a graphical interface. We illustrate our method on data from guinea pig retinal ganglion cells and document its performance on simulated data consisting of spikes added to experimentally measured background noise. We present several tests demonstrating that the algorithm is highly accurate: it exhibits low error rates on fits to synthetic data, low refractory violation rates, good receptive field coverage, and consistency across users.

Show MeSH

Related in: MedlinePlus

Template fitting to complex events.(A) Example of a single-spike event. Each subpanel shows the time course of electrical potential (in , black curves), on a particular electrode in the  array. After baseline subtraction and high-pass filtering, a spatial whitening filter was applied (see Methods). Red curves show the result of our fitting algorithm, in this case a single template waveform representing an individual neural unit. (B) Detail of a more complex event and its fit, in which a single unit fires a burst of 9 spikes of varying amplitudes (upper left channel), while a different unit fires 5 other spikes (upper right channel). Simultaneous data from four neighboring electrodes are shown. (C) Example of an overlap event and its fit, which now is a linear superposition of 7 templates. (D) Detail of (C), showing signals on four of the electrodes. This time individual fit spikes are displayed. The red and green traces show fit templates that, although similar, differ significantly in their overall strength, and in the relative strengths of their features. The black trace shows a fit to a low-amplitude template that was later classified as unusable, and hence was discarded, by the procedure in Methods, Step 4.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3140468&req=5

pone-0019884-g004: Template fitting to complex events.(A) Example of a single-spike event. Each subpanel shows the time course of electrical potential (in , black curves), on a particular electrode in the array. After baseline subtraction and high-pass filtering, a spatial whitening filter was applied (see Methods). Red curves show the result of our fitting algorithm, in this case a single template waveform representing an individual neural unit. (B) Detail of a more complex event and its fit, in which a single unit fires a burst of 9 spikes of varying amplitudes (upper left channel), while a different unit fires 5 other spikes (upper right channel). Simultaneous data from four neighboring electrodes are shown. (C) Example of an overlap event and its fit, which now is a linear superposition of 7 templates. (D) Detail of (C), showing signals on four of the electrodes. This time individual fit spikes are displayed. The red and green traces show fit templates that, although similar, differ significantly in their overall strength, and in the relative strengths of their features. The black trace shows a fit to a low-amplitude template that was later classified as unusable, and hence was discarded, by the procedure in Methods, Step 4.

Mentions: A major challenge for a spike sorting algorithm is dealing with variability in spikes produced by individual neural units. An even greater challenge arises from spatio-temporal overlaps between spikes from different neural units. Our low error rate in analysis of synthetic data containing both of these complexities (Fig. 3) provides evidence that our algorithm is effective at resolving overlaps and identifying variable spikes from given units. To test this further, we manually examined many events in the real data which a human observer could identify as representing overlaps or neural variability; and our algorithm typically did an excellent job of dealing with variable-amplitude bursts (Fig. 4B), as well as events that overlap in space and time (Fig. 4C).


Fast, scalable, Bayesian spike identification for multi-electrode arrays.

Prentice JS, Homann J, Simmons KD, Tkačik G, Balasubramanian V, Nelson PC - PLoS ONE (2011)

Template fitting to complex events.(A) Example of a single-spike event. Each subpanel shows the time course of electrical potential (in , black curves), on a particular electrode in the  array. After baseline subtraction and high-pass filtering, a spatial whitening filter was applied (see Methods). Red curves show the result of our fitting algorithm, in this case a single template waveform representing an individual neural unit. (B) Detail of a more complex event and its fit, in which a single unit fires a burst of 9 spikes of varying amplitudes (upper left channel), while a different unit fires 5 other spikes (upper right channel). Simultaneous data from four neighboring electrodes are shown. (C) Example of an overlap event and its fit, which now is a linear superposition of 7 templates. (D) Detail of (C), showing signals on four of the electrodes. This time individual fit spikes are displayed. The red and green traces show fit templates that, although similar, differ significantly in their overall strength, and in the relative strengths of their features. The black trace shows a fit to a low-amplitude template that was later classified as unusable, and hence was discarded, by the procedure in Methods, Step 4.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3140468&req=5

pone-0019884-g004: Template fitting to complex events.(A) Example of a single-spike event. Each subpanel shows the time course of electrical potential (in , black curves), on a particular electrode in the array. After baseline subtraction and high-pass filtering, a spatial whitening filter was applied (see Methods). Red curves show the result of our fitting algorithm, in this case a single template waveform representing an individual neural unit. (B) Detail of a more complex event and its fit, in which a single unit fires a burst of 9 spikes of varying amplitudes (upper left channel), while a different unit fires 5 other spikes (upper right channel). Simultaneous data from four neighboring electrodes are shown. (C) Example of an overlap event and its fit, which now is a linear superposition of 7 templates. (D) Detail of (C), showing signals on four of the electrodes. This time individual fit spikes are displayed. The red and green traces show fit templates that, although similar, differ significantly in their overall strength, and in the relative strengths of their features. The black trace shows a fit to a low-amplitude template that was later classified as unusable, and hence was discarded, by the procedure in Methods, Step 4.
Mentions: A major challenge for a spike sorting algorithm is dealing with variability in spikes produced by individual neural units. An even greater challenge arises from spatio-temporal overlaps between spikes from different neural units. Our low error rate in analysis of synthetic data containing both of these complexities (Fig. 3) provides evidence that our algorithm is effective at resolving overlaps and identifying variable spikes from given units. To test this further, we manually examined many events in the real data which a human observer could identify as representing overlaps or neural variability; and our algorithm typically did an excellent job of dealing with variable-amplitude bursts (Fig. 4B), as well as events that overlap in space and time (Fig. 4C).

Bottom Line: Our method can distinguish large numbers of distinct neural units, even when spikes overlap, and accounts for intrinsic variability of spikes from each unit.Human interaction plays a key role in our method; but effort is minimized and streamlined via a graphical interface.We illustrate our method on data from guinea pig retinal ganglion cells and document its performance on simulated data consisting of spikes added to experimentally measured background noise.

View Article: PubMed Central - PubMed

Affiliation: Department of Physics and Astronomy, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America. jprentic@sas.upenn.edu

ABSTRACT
We present an algorithm to identify individual neural spikes observed on high-density multi-electrode arrays (MEAs). Our method can distinguish large numbers of distinct neural units, even when spikes overlap, and accounts for intrinsic variability of spikes from each unit. As MEAs grow larger, it is important to find spike-identification methods that are scalable, that is, the computational cost of spike fitting should scale well with the number of units observed. Our algorithm accomplishes this goal, and is fast, because it exploits the spatial locality of each unit and the basic biophysics of extracellular signal propagation. Human interaction plays a key role in our method; but effort is minimized and streamlined via a graphical interface. We illustrate our method on data from guinea pig retinal ganglion cells and document its performance on simulated data consisting of spikes added to experimentally measured background noise. We present several tests demonstrating that the algorithm is highly accurate: it exhibits low error rates on fits to synthetic data, low refractory violation rates, good receptive field coverage, and consistency across users.

Show MeSH
Related in: MedlinePlus