Article Figures & Data
Figures
- Figure 1.
In LIP, the reduction in Fano factor becomes smaller with greater saccade target–RF distance. A, B, Timing (A) and design (B) of target-mapping task. Monkeys made delayed visually guided saccades to saccade targets randomly chosen from a 40° × 40° square grid of locations around the central fixation point (B). A distractor was flashed briefly at the RF center 500 ms into the 1050 ms delay period (see Materials and Methods). The saccade target overlapped the fixation point, and the monkey maintained fixation in control fixation trials. C, Population average PSTF of the Fano factor indicates a greater reduction of the Fano factor in the epoch between target and distractor appearance for smaller target–RF distances (colors indicated in boxed legend). D, Population average PSTHs of the corresponding mean spike rate for the PSTFs plotted in C suggests that the variation in Fano factors beyond 20° is not accompanied by changes in mean spike rate. The PSTF in C and the PSTH in D were calculated in 100 ms windows stepping every 25 ms. Vertical lines in C and D indicate the time of target, distractor, and go-cue onset. E, F, Pairwise comparisons confirm the decrease in Fano factor at smaller target–RF distances: scatter plots of each neuron's Fano factor (average of five non-overlapping 100 ms bins from −470 to 30 ms relative to distractor onset, gray bar in C) comparing saccades to the near surround (abscissa) with those to the far surround (ordinate in E; signed-rank test, p = 0.0047; n = 45) and to the RF (ordinate in F; signed-rank test, p = 0.0006; n = 72 neurons). One point (3.18, 6.01) was omitted from the display in E for better visibility. Green, black, and magenta circles indicate neurons from monkeys I, D, and Z, respectively. Part of B was reproduced with permission from Falkner et al. (2010).
- Figure 2.
Fano factors are reduced during the target-mapping task and show a steady decrease from the surround toward the saccade target. A–C, The values for the control fixation task were subtracted from that for the delayed saccade task with targets at different locations. A, C, In contrast, mean spike count shows a change from enhancement to suppression with increasing target–RF distance (A), and a similar pattern is visible for the variance of spike counts (C). D, The Fano factor increases with measurement window width but still shows the same pattern with target–RF distance: the plot for the 100 ms bin width is identical to that in B. A–C, Error bars are SEMs. Fewer than five neurons had data for target–RF distances beyond 50° (data not shown). Each point averages data from the bordering 5°. Δ represents change.
- Figure 3.
LIP neuronal Fano factor increases with mean spike rate in the surround. A, Population average PSTF of the Fano factor indicates a larger Fano factor in the epoch between target and distractor appearance for larger mean spike counts. Target locations >25° from the RF center were rank ordered by mean spike rate and then split into three equal classes for each neuron before averaging. The lowest one-third of spike rates is shown in blue, the middle one-third in red, and the highest one-third in magenta PSTF. B, Population average PSTHs of the corresponding mean spike rate for the PSTFs plotted in A confirm differences in spike rate, as expected from the sorting procedure. C, D, Pairwise comparisons confirm the trends shown in A: scatter plot comparing each neuron's Fano factor (average of five non-overlapping 100 ms bins, −470 to 30 ms relative to distractor onset, gray bar in A) for trials with the middle one-third of spike rates (abscissa) to those with the lowest one-third of spike rates (ordinate in C; signed-rank test, p < 0.0001; n = 72 neurons) and the highest one-third of spike rates (ordinate in D; signed-rank test, p < 0.0001; n = 72 neurons). Two points in D (x-axis: 1.63, 2.39; y-axis: 4.98, 4.09) were omitted from the display for better visibility. Figure format for PSTFs, PSTHs, and scatter plots are as in Figure 1.
- Figure 4.
Target–RF distance interacts with spike rate in a multiplicative manner to affect the Fano factor (FF) in the surround. A, B, Regressing Fano factor on mean spike rate in different target–RF distance windows gives a slope (A) that increases and an intercept (B) that decreases with the distance of the target from the RF center (for targets at least 10° away from the RF center). A, B, Slopes and intercepts as a function of target–RF distance calculated in partially overlapping 10° target–RF distance windows, centered at 5° intervals from 5° to 30° away from the RF center. The final window included only locations >35° away from the RF center. The dotted line is derived by averaging (across neurons) the predicted slope (A) and intercept (B) at each target–RF distance using a single linear multiple-regression model with a multiplicative interaction between spike rate and distance for all saccade target locations at least 10° from the RF center. Error bars represent SEM. Variations of both slope and intercept with distance (solid lines) are statistically significant (Friedman's test, p < 0.0001). C, Single-neuron example of relationship of Fano factor versus mean spike count with superimposed regression line for target locations between 5° and 15° (blue) and between 25° and 35° (black) from the RF center. D, Relationship of Fano factor versus mean spike count for different target–RF distances from 10° to 30° (in steps of 5°), based on the average parameters of the general linear model in the population (solid lines) and the individual regressions for each target–RF distance window (dotted lines). The range of mean spike count values for each line comes from the mean 10–90% range of mean spike count values in the population for target–RF distance windows extending 5° above and below the value for the line. E, F, Identical to A and B respectively, except that the Fano factor was calculated over the entire 500 ms predistractor epoch.
- Figure 5.
Increasing motivation via expected reward has a differing (push–pull) effect on mean spike rate for target and distractor responses but reduces Fano factor globally. A, Timing (left) and schematic (right) of visually guided saccade task with variable reward. B–E, The PSTHs (B, C) and PSTFs (D, E) show data for large rewards in red and for small rewards in blue. Data in B, D, and F (left column) from target in RF, distractor in surround task condition, and data in C, E, and G from target in surround and distractor in RF task condition. B, C, Population average PSTHs of the spike rate (15 ms non-overlapping bins) show that increasing expected reward increases the mean target-response and decreases the mean distractor response. The left and right graphs in each panel show data aligned to distractor onset and saccade onset, respectively; the left graphs replicate a plot in Falkner et al. (2010) and are reproduced with permission. D, E, Corresponding PSTFs for the Fano factor, however, show a reduction in Fano factor when a larger reward is expected, independent of target location. Gray bar at bottom denotes time window for scatter-plot computation (F, G). PSTFs and PSTHs are as in Figure 1C. F, G, Scatter plot of data from each neuron (calculated as an average of non-overlapping 100 ms windows from 400 ms before to 500 ms after distractor onset, marked in gray in B–E) shows a significant decrease in Fano factor with increased motivation for both conditions (F: p = 0.0008, n = 38; G: p = 0.0030, n = 46; signed-rank test). One point in F (x-axis, 3.10; y-axis, 4.51) and G (x-axis, 4.91; y-axis, 3.36) was omitted from the display for better visibility. Right panels in F and G shows normalized frequency polygons of spike counts for two-example neurons with data from the small (blue) and large (red) reward conditions for each neuron, using the entire analysis window. With larger expected reward, for the neurons in the upper and lower plots, the Fano factor was reduced from 5.29 to 2.37 and 7.12 to 5.41 (target in RF) and from 2.78 to 1.99 and 5.42 to 4.26 (distractor in RF), respectively.
- Figure 6.
A negative-binomial model fit captures the effects of expected reward on the spike count distribution. A–D, Blue and red indicate small and large rewards, respectively. Solid lines are frequency polygons of the distribution of mean spike count across all recorded neurons (n = 35 neurons with target in RF in A and C; n = 40 neurons with distractor in RF in B and D; only neurons with variance greater than or equal to the mean and at least 10 trials for both large and small reward conditions are included): one randomly chosen trial from each neuron was used to compute a mean spike count, and the distribution of the mean spike count was estimated by repeating the process 25,000 times. Dotted lines (appearing almost superimposed on the solid ones) are similar, except that the sum was calculated by randomly drawing one count from the best-fitting negative-binomial model to each neuron's distribution rather than from the data itself; vertical lines in A and B indicate the theoretical means of these distributions. C and D show the data from A and B after subtracting the mean to center the distributions. Spikes were counted in the epoch from 400 ms before to 500 ms after distractor onset for the target in RF condition (A, C), and from 470 ms before to 30 ms after distractor onset for the distractor in RF condition (B, D).
- Figure 7.
Larger neuronal Fano factor is associated with longer-latency memory-guided saccades to the surround. The PSTFs and PSTHs show data for saccades with latencies longer than the median in red, and for saccades with latencies shorter than the median in blue. Graph format as in Figure 5. A, B, Timing (A) and schematic diagram of task (B). C–E, Population PSTHs for the mean spike rate (D) and PSTFs for the Fano factor (E), as well as the scatter plot for Fano factor associated with short-latency (abscissa) and long-latency saccades (ordinate) in C (evaluated 200 ms before the go-cue; gray window in E). The data in C and E indicate a larger Fano factor for long-latency saccades; this is statistically significant (signed-rank test, p = 0.0247; n = 39). Two points in C (x-axis, 1.60, 8.27; y-axis, 2.02, 6.34) were omitted from the display for better visibility.
