Skip to content

OpenSorter User Guide

OpenSorter is stand-alone neurophysiological spike sorting software for TDT's OpenEx or Synapse software suite.

Before You Begin

Requirements

TDT Drivers and the OpenEx or Synapse Suite must be installed before installing OpenSorter.

The recommended operating systems for all TDT systems are Windows 7 and 10.

Organization of the Manual

This manual is organized in the following sections:

  • Overview - briefly describes the OpenSorter application and how it works with the OpenEx Software Suite.
  • OpenSorter Basics - describes the visual interface and how to select and view data.
  • Sorting - details the available sorting algorithms, manual sort techniques, and eliminating outliers.
  • Sorting Statistics - details the available statistics and how to view and export them.
  • Tank Filtering - details filtering events by time, epocs, or artifact rejection.
  • Processing Multiple Data Sets - details techniques for batch processing and concatenating data.
  • Menu and Toolbar Reference - a quick reference for the toolbars and menus.
  • Glossary

Overview

OpenSorter is designed to sort spike event data saved in the OpenEx DataTank format. Sorting is based on principal components feature space, waveform parameter feature space, or waveform shape. Semi-automated, K-Means, and manual sorting methods are supported. Sorting results can be manually edited and spikes re-assigned using editors in the feature space, timeline, or waveform space displays. Sorting analyses, such as Pseudo-F statistics, isolation distances, L-ratios and Silhouette indices, display sort quality and help to guide any manual resorting and editing that may be required.

OpenSorter features:

  • Fully manual to fully automated spike sorting modes
  • Sorting and editing in both the feature space and the waveform space
  • Supervised or unsupervised batch mode
  • Data set level and cluster level outlier elimination
  • Dynamic sort editing
  • Selectable sorting dimensions
  • Tank Filtering
  • Statistical analysis
  • Stepwise undo/redo for interactive data exploration, editing, and analysis

Sorting operations are carried out on the active data set displayed in the main display and editing window. Data sets are defined at the channel level and are selected in the Tanks panel using an expandable tree structure similar to that used in many Windows applications.

Two general sorting methods are available: graphical or manual selection and assignment of spikes using OpenSorter's mouse-based selection tools, and algorithm-based sorting methods that automatically assign sort codes to each spike based on user-defined algorithm parameter settings. Sort codes can then be saved into a SortID.

SortID

A SortID is a new level of the data tank format that contains the sort codes assigned to each spike in a data set. Since OpenEx and Synapse offer real-time spike sorting capabilities, spikes read into OpenSorter may already be labeled with pre-existing sort codes. In the event that real-time sorting was not used during acquisition, each spike will still be labeled with a sort code (0) and will appear as unsorted in the OpenSorter display and editing window. For all datasets, sort codes assigned during acquisition (online) appear as a read-only SortID labeled TankSort. The TankSort SortID and thus the sort codes stored during acquisition cannot be overwritten. This SortID is marked with a padlock icon in the Tank Browser window. Sort codes assigned offline using OpenSorter are instead saved to new SortIDs with names assigned by the end-user. Data sets can be sorted or edited any number of times, with the resulting sort codes stored to multiple SortIDs for each dataset.

The new SortIDs can be used with OpenEx Client applications such as OpenBrowser and OpenExplorer or via user-developed programs using OpenDeveloper calls. Events with associated SortIDs appear in the Events lists of these programs as an event followed by the name of their SortID (for example, SNIP_MySort). When you select these events, the named event (in this case SNIP), is selected along with the sort codes associated with the specified SortID (MySort).