Figure 1.

A. Raster plots and spike density waveforms (σ = 5 ms) recorded from a representative visual transient (VT), visual sustained (VS), visuomotor transient (VmT), and visuomotor sustained (VmS) neuron to a delayed saccade task, which was used to facilitate neural classification. Data are aligned on target appearance (left column) and saccade onset (right column) in the delayed saccade task when the target appeared in the neuron's response field. B. The response of the same single neurons to the 7 stimuli in the standard repetition paradigm. The black bars across bottom of abscissa represent the stimulus timing. Spikes for individual trials are presented in raster format (only a subset of trials shown for display purposes) and overlaid with a mean spike density function (σ = 5 ms). C. Scatter plots and histograms of the metrics used to classify cells. The transient-sustained index is plotted against the visual-motor index for each cell (color indicates cell class), with smaller numbers indicating a more motor and more transient responses, respectively, as measured from responses in the visual delay task shown above. The histograms show the number of cells with each parameter value using a bin width of 0.025 units. The dashed lines show the cutoff values that were used to separate classes of neurons into the 4 categories. D. The mean depth for each cell class. The cell classes had significantly unequal variances (Bartlett's test, T(3) = 15.50, p = 0.0014), and consequently a Kruskal-Wallis test was conducted to evaluate differences in depth among the four cell classes. Cell classes significantly differed in depth (C2(3, 108) = 20.10, p = 0.0002), and pair-wise Wilcoxon rank-sum tests (Bonferroni corrected) indicated that visual transient cells significantly differed from both visual motor transient and visual motor sustained (z=3.78, p < 0.001 and z=3.79, p < 0.001, respectively.).