Dynamic sculpting of directional tuning in the primate motor cortex during three-dimensional reaching

Abstract

In the present study, we investigated how directional tuning of putative pyramidal cells is sharpened by inhibition from neighboring interneurons. First, different functional and electrophysiological criteria were used to identify putative pyramidal and interneuronal subtypes in a large database of motor cortical cells recorded during performance of the three-dimensional center-out task. Then we analyzed the relationship between the magnitude of inhibition and the tuning width, and a significant decrease of the latter as a function of the former was found in a population of putative pyramidal cells. In fact, the coupling of inhibition with narrow tuning was observed before and during movement execution on a cell-by-cell basis, indicating an important dynamic role of inhibition during movement control. Overall, these results suggest that local inhibition is involved in sculpting the directional specificity of a group of putative pyramidal neurons in the motor cortex.

Figure 1.

Cell type properties. A, Typical examples of the action potential shape (left), spontaneous activity (center), and normalized average discharge rate (NRA; right) for tuned neurons for the PI, PP1, and PP2 cell types. The red crosses in the panels on the left represent the times of beginning and end of the action potential. The action potential width and the discharge rate of the neurons during the control period are given at the top. In the NRA, the activity was sorted in descending order using the discharge rate during TET for each movement direction and then plotted in the order specified on the center right panel (d1, d2…, d8); the activity was aligned to movement onset (vertical line; M), and the black horizontal line represents the SD of the appearance of the target centered on the mean value. B, Examples of monosynaptic inhibition (top) and excitation (bottom) using STA analysis. STA (left) showing the number of spikes of the triggered cell as a function of the lag with respect to the triggering cell (PI, top; PP1, bottom) is shown. The mean ± SD of the action potential shape for the triggering and trigged cells are show in the center and right panels, respectively. The numbers on the top right of each panel are the widths of the action potentials.

Figure 2.

Theoretical Gaussian tuning curves of cells that show inhibition (left) or no inhibition (right) in their anti-PD. The discharge rate in the anti-PD of the neuron on the left is lower than its spontaneous activity (SA), depicted as a black line and a gray arrow. The black dashed line illustrates the half-width dispersion.

Figure 3.

A, Distribution of tuning dispersion for each cell type. B, Cumulative distribution of tuning dispersion. C, Cumulative distribution of the R² value of the Gaussian function fitting for each cell type. Dotted line, PI; dashed line, PP1; black line, PP2.

Figure 4.

Preferred directions of neurons corresponding to the three cell types.

Figure 5.

Mean onset response latency for neurons of the three types sorted in ascending order using the magnitude of discharge rate (Max Dir). Dotted line, PI; dashed line, PP1; black line, PP2.

Figure 6.

Scatter plots of the spontaneous (control period) discharge rate against the activity in the anti-PD direction, with the tuning dispersion indicated as the size of the point depicted. Cells with dispersion tuning <65° were plotted as filled circles, whereas cells with dispersions >65° were plotted as open circles. The left and right parts of the figure are for PP1 and PP2 cells, respectively.

Figure 7.

Inhibition index plotted as a function of tuning dispersion. The inhibition index fall with dispersion for PP1 cells but remained similar for the PP2 cells. Dotted line indicates an inhibition index = 0. The left and right parts of the figure are for PP1 and PP2 cells, respectively.

Figure 8.

A, The tuning dispersion in degrees (open circles, dashed line) and II (filled circles, continuous line) plotted as a function of the time window. The right and left ordinate scales correspond to dispersion and II, respectively. The analysis was centered on the movement onset. The transparent gray box depicts the II values between −1 and 1 as a reference. B, Tuning functions of the three time windows marked with asterisks in A. The dotted line ellipse corresponds to the half-width dispersion, the abscissa corresponds to the angle θ, the ordinate to the angle φ (Fig. 1A) (Naselaris et al., 2006a), and the firing rate is represented in a gray scale. The decrease in tuning dispersion in the intermediate tuning function is associated with an increase in the II shown in A.

Figure 9.

Tuning dispersion and II as a function of time window for six PP1 cells. The same conventions as in Figure 8A are used. Mov, Movement.

Figure 10.

Dynamic coupling between the decrease in tuning dispersion (top) and the increase in the inhibition index (middle) for the PP1 cells (n = 124) with an overall positive inhibition index, and the increase in discharge rate on the preferred direction (bottom) of the PI cells (n = 187) with significant directional tuning. Box plots for each sliding window where the white bar corresponds to the median of the distribution and the top and bottom of the box correspond to the 75th and 25th percentiles, respectively. The analysis was centered on the movement onset, and the two vertical lines signal the approximate beginning and end of the effects on the three measures.

Figure 11.

Maps of recording sites of the different cell types. A, Recording sites of PP1 neurons with a tuning dispersion <60° (small gray dots) and of PI (black dots). B, Recording sites of PP1 neurons with a tuning dispersion >60° (large gray dots) and of PP2 neurons (black dots).