Clustering

Cluster window

As an alternative to template matching, spikes can be sorted by cluster cutting. Values calculated by PCA or user-defined measurements are plotted to a 3D clustering window.

• View can be rotated in 3 dimensions using mouse or sliders
• K-means and normal mixtures algorithms for automatic cluster separation
• Manual cluster separation by placing ellipses
• View dots from a percentage of the entire time range and step through time to track moving clusters
• Generate interval histograms based on displayed clusters
• Show/hide particular clusters
• Visual identification of clustered spikes occurring within a specified refractory period
• Density plot option to aid cluster identification
• Copy values as text (underlying data values and display co-ordinates)
• Select individual spikes with a mouse click for manual classification

Principal Component Analysis

• Select principal components to plot on X, Y and Z axes
• Calculation of components for multiple trace (stereotrode or tetrode) channels based on entire waveform, peak amplitude or ratio of peak to mean amplitudes

User-defined measurements

• Select measurements for X, Y and Z axes to cluster the data
• Measure times, amplitudes, slopes and areas
• Save up to 10 configurations containing different sets of measurements
• For multiple trace data, measure individual trace or mean of trace measurements