Offline Sorter

Offline Sorter (OFS) version 3 is a powerful, easy-to-use tool for viewing and classifying action potential waveforms (spikes) collected from single electrodes, stereotrodes, and tetrodes. Spikes can be displayed as points in either 2D or 3D feature space, where a variety of manual, semi-automated, or fully automated clustering techniques can be applied in order to classify (sort) the spikes. OFS can also perform spike extraction on continuously-recorded neural data using a variety of different thresholding methods. OFS allows verification of sorting through a variety of displays, and can calculate cluster separation statistics and sort quality metrics. The sorted spikes can be exported in several formats for subsequent analysis.

The parameter space for several fully automatic spike sorting algorithms can be systematically scanned through, with graphs displaying the sort quality metric as a function of the parameters varied. The results of the automated scanning through sorting parameter space can be saved in a .SCAN file. This, combined with new batch mode commands, allow the running of various sorting algorithms overnight, followed by rapid manual viewing of different sort outcomes and selection of the appropriate sort for each channel.

Offline Sorter version 3 also contains new time segmented views of the waveforms, with cluster vs. time and sort quality vs. time views that show how the waveforms change and evolve over time. Offline Sorter version 3 ships with two USB license keys. Plexon also offers additional USB license keys as well as upgrades from version 2 with either one or two USB license keys.

Multiple methods for spike detection in continuously digitized data files can be used, including:

• Voltage threshold
• Signal energy (~ voltage²)
• Nonlinear energy

Stereotrode and tetrode spike detection is supported. A low-cut filter may be applied to the continuous data prior to spike extratraction.

The typical analysis pathways and files related to OFS are shown in the diagram below.

Offline Sorter can be used to sort or re-sort data files containing either thresholded waveform segments or continuously digitized data. File types supported include:

• Plexon
• NeuroExplorer
• Cyberkinetics
• CED Spike-2
• DataWave
• Neuralynx, Inc.
• Multi Channel Systems
• Panasonic MED
• Neuroshare
• Generic binary continuous data

Powerful Waveform Viewing and Sorting

Manual cluster selection in 3D feature space using principal component projections, voltage slices in time, or more than 20 other waveform features such as peak, valley, full-width at half maximum, etc.  Box sorting (like in the Sort Client on-line MAP control software) Waveform selection in time-voltage space  Template matching algorithm  Tetrode and Stereotrode waveform extraction and sorting using any method  Unit cross-correlograms and ISI plots   Waveform density plots in 2D or 3D in any feature space Waveform alignment   Interval invalidation and artifact removal    Raster displays of spikes and continuous data   Printable sorting summary view with export to PowerPoint®  Sort summary statistics Waveform features and statistics that can be easily exported to MATLAB®, Excel® or text file Export of sorted data to NeuroexplorerTM or optimized binary format (for fast loading) Save and recall particular sorting instances for a selected channel (useful for comparisons)

Time Segmentation, Adaptive Sorting Methods, and Time-dependent Analysis

Semi-Automatic and Automatic Sorting

Enhanced automatic sorting methods with systematic scanning through parameter space and graphs displaying the sort quality metric as a function of the parameters that were varied; select the best sorting according to a selected sort quality metric Automatic sort scan results for each channel can be saved to and loaded from .SCAN files. This, together with the new batch mode scan commands, will allow the running of certain sorting scans overnight, followed by manual reviewing and selection of the most appropriate sort for each channel Automatic cluster selection in feature space using valley seeking or T-Distribution E-M Semi-automatic cluster selection in feature space using K-means or Standard E-M (Expectation Maximization)

Enhanced Batch Mode Processing for Running Sorting Algorithms Overnight

New batch mode commands for scanning through automatic sorting parameters New menu item to edit batch files with a user-selected editor   Memory of previous batch file locations  New Quick Batch Reference showing all batch commands  New menu item showing the log file resulting from the latest batch run

Enhanced Continuous Data Handling and Spike Detection

Up-sampling of continuous signals, with linear or spline interpolation FFT view of continuous data   Single and Dual Thresholds for spike detection Waveform detection and extraction now appropriately handles differences in voltage scales between spike waveform segments and continuous spike data

Improved User Interface, Displays, and Performance

For ordering information contact Plexon Sales at (214) 369-4957 or info@plexon.com. You may also download a fully functional demo version from the Plexon site.