Role of ER Export Signals in Controlling Surface Potassium Channel Numbers

and 265 from the other. Sample interval activity was summed over 800 ms, beginning 100 ms after stimulus onset. The delay interval activity was summed from 300 ms after sample offset to 100 ms after the end of the delay. Baseline activity was from the 500 ms of fixation before sample onset. 21. T test versus baseline activity, P , 0.01. Parametric statistics such as t tests assume normal distributions. Because neuronal activity is sometimes not normally distributed, we also computed nonparametric statistics for all main effects. They yielded a virtually identical pattern of results. 22. T tests on activity from the sample and delay intervals, both P , 0.001. 23. P . 0.6. 24. Paired t tests between activity of all stimulus-selective neurons in response to the two categories computed in successive 100-ms time bins. A significant difference (P , 0.01) began 100 to 200 ms after sample onset, when the earliest PF neurons began responding. The immediate appearance of category information was also evident in average histograms across the neuron population. 25. One-way analysis of variance (ANOVA) on the 54 sample stimuli. Sample interval, 62 neurons; delay interval, 33 neurons; P , 0.01. 26. T test that BCD/WCD ratios were significantly different from 1. Sample interval, BCD/WCD mean 5 1.30; delay interval, BCD/WCD mean 5 1.49; both P , 0.001. Category information was significantly stronger during the delay; one-tailed t test, P 5 0.04. An index of (BCD 2 WCD)/(BCD 1 WCD) yielded similar results. 27. Excluding neurons with firing rates below 2 Hz (which produce spurious values when normalized) yielded 55 and 29 neurons with selectivity in the sample and delay intervals, respectively. We normalized each neuron’s activity as a proportion of its activity in response to the most effective morph level. To ensure that analyses were not biased toward a category effect, we used only the single stimulus that evoked the maximum response to determine preferred and nonpreferred category. 28. Two-way ANOVA of category membership and level of category (60%, 80%, 100%), test of the category factor. P , 0.01 for both intervals. 29. Two-way ANOVA, test of the level factor, P . 0.6 for both intervals.