Scribd
Upload
2 views

Study - Multitrodes

The document discusses methods for sorting overlapping spike waveforms from electrode and tetrode recordings, highlighting the challenges posed by simultaneous recordings from multiple neurons. It presents an automated clustering approach that improves waveform resolution without requiring fixed similarity thresholds, and compares the performance of heptodes and tetrodes in various noise conditions. The findings indicate that heptodes outperform tetrodes in high noise scenarios with many active neurons, while both perform similarly under less demanding conditions.

Uploaded by

Richard Ardelean
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
2 views

Study - Multitrodes

The document discusses methods for sorting overlapping spike waveforms from electrode and tetrode recordings, highlighting the challenges posed by simultaneous recordings from multiple neurons. It presents an automated clustering approach that improves waveform resolution without requiring fixed similarity thresholds, and compares the performance of heptodes and tetrodes in various noise conditions. The findings indicate that heptodes outperform tetrodes in high noise scenarios with many active neurons, while both perform similarly under less demanding conditions.

Uploaded by

Richard Ardelean
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

TETRODE/MULTITRODE

Sorting Overlapping Spike Waveforms from Electrode and Tetrode Recordings

When recording extra-cellularly from multiple neurons, it is common to have action potential

waveforms from one neuron altered by the action potentials of other neurons. This often
results

in waveform shapes and amplitudes that are significantly different from single-unit
waveforms,

thereby posing problems for spike sorting algorithms, especially when those algorithms are

automated.

The most common approaches used to resolve overlapping waveforms in extracellular


recordings create combinations of previously identified single-unit waveforms, usually
referred to as templates, and then measure the similarity between the overlapping
waveforms and the template combinations. These methods usually use the amplitude values
of the waveforms to find the best match, although other features such as Fourier coefficients
of the waveforms have also been used. Template-matching methods usually tend to be slow
as well, because comparisons of all potential overlapping waveforms with all template
combinations are needed.

Other methods that do not involve template matching concentrated on extracting more
robust features other than amplitude to cluster overlapping waveforms with single-unit
waveforms. Wavelets+KMeans clustering and superparamagnetic clustering

Our automated method for resolving overlapping waveforms produced good results when
sorting the simulated datasets we studied. The main difference between our method and
methods previously reported is that we used a clustering method to find the best match for
each overlapping waveform among possible synthetic waveforms. This improved the
performance, as we were able to use more than one feature to compare the overlapping
waveforms with available templates. Also, unlike nearly all the methods we discussed in the
Introduction, our method did not require setting a fixed threshold on similarity measures
(e.g., correlation coefficients, root-mean-square error, etc.) to identify the synthetic
waveforms that were similar to the overlapping waveforms, which makes it easy to apply this
method to different datasets. Instead, we chose to let theclustering algorithm decide if
waveforms belong together. While this implies that appropriate features have to be used
during the clustering at this stage, we found in our experience that the same features used to
identify clusters of single units can be used successfully to resolve overlapping waveforms.

Another advantage of our method was that we generated superpositions at all possible
phase shifts, which gave us the ability to resolve overlapping waveforms that were
generated by synchronous firing of two single units.
Our method, however, cannot deal effectively with differences in spike waveforms from a
single cell caused by electrode movement or bursting cells

Spike sorting: the overlapping spikes challenge

When recording action potentials (spikes) from many neurons simultaneously via
multichannel microelectrodes the overlapping of spikes from different neurons is a
demanding problem for detection and classification of spikes (spike sorting). Since
multichannel electrodes provide better possibilities to separate the superimposed waveforms

The amplitude of a spike on a channel corresponds to the distance between neuron and
electrode tip. This effect leads to a better spatial resolution

Results indicate that the performance of the algorithm increases with higher numbers of
used signal channels. With low noise amplitudes the algorithm achieves reliable results for
all numbers of signal channels. With rising noise level, the results of the heptode and the
tetrode only decline slowly, while the results of the monotrode fall off even with little rise of
noise. The improvement even the stereotrode achieves in comparison to the monotrode is
clearly visible

Increased number of neurons: The results of the monotrode decline rapidly when more
neurons are enclosed in the simulation, while the results of the heptode and the tetrode stay
reasonably constant. Once again, already the use of a stereotrode leads to a significant
improvement compared to the monotrode.

Our results indicate that the performance increases with rising number of signal channels
especially under conditions with high noise amplitudes and a high number of neurons. Due
to the fact that neurons produce spikes with stereotypic shapes the waveforms can be quite
similar. The use of multichannel electrodes leads to a greater dissimilarity due to the
characteristic amplitude distribution and consequently provides a better separability. The
runtime and the computational eort increase with rising number of channels. However, the
optimal number of channels is yet unknown.

Are Heptodes Better than Tetrodes for Spike Sorting?

For less noise, both heptodes and tetrodes perform similarly. Since in these signals are 14
neurons present it gets clear that heptodes perform better than tetrodes when many neurons
are active and when the noise level is high

The dependence on the number of principal components is shown in Fig. 6. It can be seen
that heptode and tetrode based spike sorting results are similarly depending on the number
of principal components used for analysis. Both yield maximum performance for five
principal components.

Our results indicate that heptode outperform tetrode recordings when many neurons are
present in a signal with a poor signal-to-noise ratio. In signal conditions that are not that
demanding, both, heptodes and tetrodes, perform equally for spike sorting.

The automatic selection of spike detection thresholds is not a trivial task, especially for low
signal-to-noise ratios. Here, a better threshold estimation method for simulation experiments
as well as for analysis of real signals has to be found.

When the spikes of many highly active neurons are present in a recording, the problem of
overlapping spikes gets relevant for the spike sorting performance.

Past, present and future of spike sorting techniques

LICENTE

A Generative Adversarial Approach for the Detection of Typical and Drowned Action
Potentials

Weighted Principal Component Analysis based on statistical properties of features for Spike
Sorting

Phd - Vlad

The superposition of distant contributions within a radius of about 140-300 µm [Henze et al.,
2000] appears on the electrode in the form of low frequency ripples (up to 300 Hz) called
local field potential (LFP). LFPs are produced by slow waves that include synaptic activity,
somatodendritic spikes, afterpotentials, and voltage gated subthreshold membrane
oscillations. LFPs reflect cooperative events both in the input and in the local processing of
an area [Logothetis, 2003] comprising excitatory as well as inhibitory neurons [Rasch et al.,
2008]. Consequently, an intimate relation exists between spiking activity and LFP [Rasch et
al., 2008]. Recorded activity of LFPs and APS is separated by filters: a low pass filter, with a
cutoff frequency of 300 Hz, isolates the LFP; a high pass filter isolates the spiking activity.
Usually, a notch filter is additionally embedded in the recording setup to reject the power line
frequency (50 or 60 Hz).

You might also like