Table 3. Change in selectivity with local vs. distant return, using alternate selectivity measure.

Selectivity was defined as the largest difference, across current levels, in the activation probabilities of target and non-target cells. Each entry in the table shows the mean and 90% confidence interval (obtained by resampling), across target/non-target cell pairs, of the difference in selectivity in the local and distant return configurations. Therefore, positive values indicate increases in selectivity with local return stimulation. The cell pairs are segregated into near vs. intermediate and near vs. far, based on distance from the stimulating electrode (see Results).

LR vs. DR selectivity measured by activation probability mean (90% confidence interval)
	near vs. intermediate	near vs. far
retina 1	0.538 (0.480, 0.606)	0.195 (0.144, 0.248)
retina 2	0.344 (0.189, 0.489)	0.265 (0.0515, 0.472)
retina 3	5.37E-4 (−3.70E-3, 4.65E-3)	1.85E-4 (−3.00E-4, 6.92E-4)
retina 4	0.0639 (0.0511, 0.0786)	0.0264 (0.0139, 0.0393)
pooled	0.237 (−1.83E-4, 0.553)	0.122 (4.22E-05, 0.304)