Fig. 3.

Population responses of spatially selective GF neurons. (A) Population responses of 109 GF neurons from both monkeys for the target eliciting the strongest response (rank 1) in the spatial cue period (both 200 ms and 500 ms), the second-strongest (rank 2), and the third- (rank 3) and the fourth-strongest response (rank 4). The population discharge associated with the two most-preferred targets exhibited an increase in discharge rate and the ones associated with the least-preferred targets a suppression. (B) Population responses of GF neurons for the most-preferred and the least-preferred targets in the GF task compared with the population responses to the same targets cued by identity in the IM task. (C) For the population of spatially selective GF neurons, the median SSI values dropped significantly (Mann–Whitney U test, ***P < 0.001) from 0.48 for correct trials to 0.33 for error trials. SSI values for error trials could even become negative, indicating a reversal of preference, that is, a target that was the most preferred one in the GF task became less effective than the least preferred one when studied in the IM task. (D) SVM decoding of the discrimination between the most-preferred and least-preferred target based on the activity of all spatially selective GF, shown separately for correct and error trials and separately for the two durations of the spatial cueing periods (200 vs. 500 ms). SEs were obtained by deploying a bootstrapping procedure (n = 1,000). Shaded areas represent SEs in all figures panels.