For both data selections we found an average β1 coefficient that was significantly larger than zero (p < 0.001, t test across monkeys), indicating a significant shift of the psychometric function toward more preferred choices for convex- and concave-selective

sites. Finally, there was no significant difference between the β3 coefficient (indicating the slope change due to microstimulation) of the convex- and the concave-selective sites (p = 0.14, t test). Hence, slope changes were similar among convex- and concave-selective sites. Both monkeys displayed a small but significant response bias toward concave choices equivalent to Z-VAD-FMK datasheet on average 5.5% stereo-coherence (p < 0.01; logistic regression analysis on 3D-structure-selective and -nonselective sites with no significant effect of microstimulation to avoid misestimating the response bias due to e.g., probability Raf inhibitor matching effects

[Salzman et al., 1992]. If microstimulation in IT elicited activity that was unrelated to the sign of the 3D structure (that is, concave versus convex 3D structure), the task would be expected to become more difficult and the monkey would most likely rely more heavily on his response bias to make a choice, i.e., to choose concave. One would therefore expect a higher proportion of stimulation-induced psychometric shifts toward more concave choices. Nevertheless, we observed stimulation-induced psychometric shifts toward convex choices in 96% of all convex-selective sites. Hence, considering the convex 3D-structure-selective sites, our results cannot be explained by an activation of the monkeys’ response bias, since this would have produced shifts in the opposite,

concave direction. Microstimulation significantly biased the monkey’s choice toward more preferred choices at each of the three positions-in-depth of the stimulus (p < 0.0001 for Far-, Fix-, and Near-position-in-depth; Suplatast tosilate Figures 4A and 4B, M1; Figures 4C and 4D, M2). In addition, the strength of the microstimulation effect tended to increase with the 3D-structure selectivity of a site. Figure 5 shows the shift of the psychometric function plotted against the 3D-structure selectivity of the MUA measured at each stimulation site. For this purpose, negative and positive psychometric shifts denote shifts toward more concave and convex choices, respectively. Signed d′-values measure the 3D-structure preference of the MUA-sites, with positive and negative values indicating convex and concave preferences, respectively (see Experimental Procedures). We observed a significant correlation between the signed d′ and the signed psychometric shift in each monkey (M1: 0.79, p < 0.001; M2: 0.62, p < 0.001). The previous analysis is based on all 68 sites in which we stimulated, including 34 sites not selective for 3D shape (see below).