Waveform analysis in NeuroExplorer

NeuroExplorer version 5.005 released on November 15, 2014, introduces a new analysis – Waveform Comparison. This analysis was added at a request of one of the NeuroExplorer customers who needed to compare average waveforms from one interval to another. This is a standard requirement when one does optogenetical identification of neurons, i.e. spikes are triggered by light-induced activation. To make sure this manipulation does not dramatically change the waveform of the spike, one needs to compare shape of light-triggered spikes to spontaneously occurring spikes. When you run Waveform Comparison analysis, you can display average waveforms and waveform standard deviations (shown as a gray background around averages):

Waveforms1

 

You can also display waveforms in Principal Component space:

WaveformsPC

To compare  light-triggered spikes to spontaneously occurring spikes, you need to specify two interval filters – one for stimulation periods and one for the periods with no stimulation. You can compare waveforms side-by-side:

WaveformCompValues

In the graph above the left column shows waveforms that were recorded during stimulation, the right column shows waveform averages that were recorded in periods with no stimulation.

You can also overlay averages for each neuron:

WaveformsComp

 

as well as overlay PC projections:

WfPCOverlaid

and run MANOVA to evaluate differences in the waveform projections in principal component space:

WfManova

 

Exporting Data to Text Files

NeuroExplorer version 5.002 was released on September 8, 2014. Here is a summary of what’s new in this version:

  • Coherence, Spectrograms and PeriEvent Spectrograms analyses now have all the spectral calculation options available in other spectral analyses (selection of 3 types of data preprocessing, 6 types of windowing functions and multi-taper options).
  • Burst Analysis now saves properties of each burst in Numerical Results:

Burst Properties

  • When exporting data or numerical results to text files, a user can select what character is used to separate items in a line (comma, space or tab; previously only tab was available). The new comma-separated output (.CSV file format) is especially convenient when working with Excel since this file format is natively supported by Excel.
  • When exporting data to text files, a user can now export waveform variables and marker variables (previously, NeuroExplorer could only export neurons, events, intervals and continuous variables).
  • You probably know that you can open a .nex file in NeuroExplorer by double-clicking at the file in File Explorer. For some reason, Windows sends several open file commands to NeuroExplorer when you double click at a .nex file. You could see some files opened twice, and some files generating a “sharing violation” error. This problem is fixed in NeuroExplorer 5.002 (and while we are talking about opening files, note that you can also open a .nex file by dragging the file into NeuroExplorer).

Numerical Results for PeriEvent Rasters Are Available in Version 5

Many NeuroExplorer users asked me why perievent rasters analysis lists only perievent histograms in the numerical results table.

The reason was that listing all the timestamps for each line of the perievent raster can use a lot of RAM and can slow down the program. For example, for a spike train with 10,000 spikes, we may have 10,000 raster lines containing up to 10,000 spikes in each raster line. This adds to 100 million timestamps which would use 800 MB of RAM. Besides, the user can analyze dozens of neurons, so using so much RAM may not work, especially for a 32-bit application that cannot use more than 2 GB of RAM. The situation is somewhat better for a 64-bit program, but the program would still be too slow.

In NeuroExplorer Version 5, I came up with a solution — the program stores some of the results of the perievent raster analysis in memory and calculates all the timestamps for perievent rasters on-the-fly:


PeriEvent Rasters Numerical Results

What I found interesting was that the graphical representation of the perievent raster:


PeriEvent Raster

creates an impression that each line of the raster contains many data points while in reality each line contains only a few points as we can see from numerical results above.

Complete numerical results for PeriEvent Raster analysis and for all the other analyses that did not provide all the results in version 4 (Cumulative Activity Graphs, Instant Frequency, ISI versus Time, Poincare Maps and Synchrony versus Time) are available in NeuroExplorer 5.001 and later.

If you purchased NeuroExplorer Version 5 license, download the latest Version 5 installer.  Otherwise, download NeuroExplorer Version 5 Demo today!

 

New NeuroExplorer Website

I am sure you noticed a new design of the NeuroExplorer website. The main goal of the redesign was to make the website more dynamic — it is much easier to add new content.

However, I will be adding blog posts on a regular basis and will also be adding more web pages with the  information about NeuroExplorer. Here is a list of topics I would like to cover:

  • New and upcoming features of NeuroExplorer Version 5
  • Best practices of using NeuroExplorer user interface  (for example, a video demonstrating how to use all the bells and whistles of 1D View)
  • Interesting support cases (for example, help with scripting)
  • A gallery of screenshots of the results of various analyses in NeuroExplorer

Do you have any specific questions or requests? E-mail me at [email protected]

What’s New in NeuroExplorer Version 5

Version 5 is a truly significant expansion of functionality over the previous version. Version 5 adds several new analyses (Hilbert transform, head direction analysis and others), multi-taper spectral calculations, statistical tests (t-tests, ANOVA) and immediate data preview. The new version is also faster and can handle multi-gigabyte data files.

Here is the list of the main improvements implemented in version 5 of NeuroExplorer.

When NeuroExplorer version 5 loads a data file, it shows a preview of all data channels:

v5DataPreview

NeuroExplorer version 5 provides vastly improved spectral analyses of spike trains and continuous channels:

  • Single-taper and multi-taper calculations of spectra
  • 6 windowing functions
  • 3 preprocessing options
  • Jackknife confidence
  • Ability to replicate Matlab spectral calculations
  • Up to 10 times faster calculation of spectral results

Several new analyses of continuous data (LFP) are available:

  • Perievent rasters for continuous

perirastercont1

  • Power spectra for continuous
  • Coherence for continuous
  • Single trial spectrum analysis
  • Single trial phase analysis using Hilbert transform
  • There is a new ‘Analyses of Continuous Data’ analysis group that lists all the analyses that can be
    used with continuous channels:

ContVarAnalyses

 

Digital filtering options for continuous data are expanded and now include both IIR and FIR filters.
FIR filters can be used when a band-pass filter with a narrow frequency range is required.

  • 3 new IIR filters
  • 3 new FIR filters
  • Power spectra for continuous
  • Coherence for continuous
  • Single trial spectrum analysis
  • Single trial phase analysis using Hilbert transform

There are also improvements in spike train analyses:

  • Multi-taper options in spike train spectral analyses
  • Firing rate versus head direction analysis

New statistical tests to compare results across conditions can be run directly in NeuroExplorer:

  • Implemented via integration with freely available R-project
  • ANOVA test across multiple conditions
  • T-test test across two conditions
  • Wilcoxon test test across two conditions

ttest3

New version is available as a 64-bit Windows executable. This removes previous memory limitations
(all the available RAM can now be utilized by the program). The new version is also faster: computationally demanding analyses are run in parallel using all CPU cores.

NeuroExplorer Version 5 supports new .nex5 data file format (currently in development) that is more flexible than .nex file format:

  • Allows to save files of more than 2 GB in size
  • Allows to save unlimited metadata for the whole file and for every file variable in JSON format

The new version has the ability to save and restore results:

  • Saves all numerical and graphical results in a series of linked files
  • A user can open results files without recalculation and access both graphical and numerical results
  • A user can recalculate analysis results with a single mouse click

National language support is also implemented in the new version:

  • NeuroExplorer Version 5 correctly displays folder and file names in any language supported by Windows
  • A user can use any language to name files and variables in files
  • A user can also use any language to enter comments in data files and scripts

languages

See also: Waveform analysis in NeuroExplorer, Working with Results Files, Dealing with Noise and Artifacts in Data Viewer, New Phase Analysis, Python Scripting in NeuroExplorer