Figure 4B shows the distribution of slopes between firing rates and the number of saccades for all FEF sites. The median slope was −5.28 spikes/s/saccade. Thus, smaller responses led to greater numbers of saccades to find the target. This result is consistent with the idea that the response enhancement to the target stimulus in the FEF helps learn more guide the eyes to the target location. It was not possible to do the same analysis in V4 because the response to the target on a given trial was too highly dependent on the stimulus preferences of the individual cells. We investigated this relationship between response enhancement and saccades

in another way: by calculating the response to the target in the RF in those fixation epochs when the target stimulus would be selected for a saccade two saccades later, compared to fixation epochs when the target stimulus would be selected for a saccade more than two saccades later (Figures 5A–5C). If greater response enhancement to the target leads to fewer saccades to find the target, then the response to the target in the RF should have been greater when it subsequently took two saccades to find target (Type I target, Figure 5) than when

it took more than two saccades (Type II target, Figure 5). We only considered fixations when the two subsequent saccades were all away from the RF to avoid the influence of saccades into the RF. The predicted result was indeed found, as shown in Figures 5D–5F for early search and Figures 5G–5I for late search. Response Veliparib in vitro enhancements were significantly larger to the target when it was found after two saccades than when it was found after more than two saccades (Wilcoxon signed rank test, p < 0.05). This enhanced response to the target continued for approximately 100 ms after the initiation of the first saccade but ended before the second saccade began (see Figures 5E–5I), during which many period only distracters sharing neither

the color nor the shape with the target appeared in the cell’s RF. For comparison, Figure 5 also shows the responses to the no-share stimuli that were matched in properties to the target stimuli in the above comparisons. For these no-share stimuli, the responses were smaller than to the target stimuli in all conditions (Wilcoxon signed rank test, p < 0.05). The effects of feature attention were larger when the animal took only two saccades to find the target, but they remained significant even when the animal took more than two saccades (Wilcoxon signed rank test, p < 0.05). This specificity of the enhanced responses to the target versus no-share stimuli is consistent with a feature attention effect and is inconsistent with increases in general arousal, etc., on trials with fewer saccades to find the target. As show in Figure 6, a similar pattern of results was found in V4.