4 versus 10.3 Hz, Wilcoxon rank-sum test, p = 0.35). Conversely, M responses in Pv-INs became larger than the preferred unisensory responses upon photostimulation (37.8 versus 12.0 Hz, Wilcoxon rank-sum test, p < 0.05). PCI32765 Accordingly, the ME index for Pv-INs was increased by laser activation (Figure 8C, left; medians: 0.01 versus 0.13, p < 0.05, paired Wilcoxon rank-sum test). The distribution of the ME indexes of Pv-INs upon photostimulation suggested that photostimulation increased the percentage of Pv-INs that displayed ME by ∼50% with respect to no-photostimulation control condition (see single trial analysis for

single cells of Tables S3 and S4). The analysis of the AP response of putative pyramids Galunisertib order confirmed that Pv-INs photostimulation selectively disrupts ME in these cells (Figures 8B, 8C, and S6; Table S4). Note the opposite changes of the ME indexes for pyramids and Pv-INs upon photostimulation (decrease and increase, respectively). The fact that optogenetically promoting integration in Pv-INs selectively disrupts ME in pyramids indicates that the lack of integration in Pv-INs enables the positive ME we observed in RL pyramidal neurons. To identify anatomical sources of multimodal inputs to RL, we performed

IOI-targeted, retrograde tracers injections in RL (Figures 9A and 9B; see Experimental Procedures). Retrogradely labeled cells were found

in V1 and S1 (Figures 9C and 9D) and, at subcortical level, in the associative PO thalamic nucleus (Figure 9E), but not in visual thalamus (dLGN and lateral posterior nuclei). To characterize the role of corticocortical connections in shaping RL sensory responses, we had to overcome the problem that pharmacological blockade of the entire V1 or S1 would have invariably caused diffusion of the silencing agent into RL. We therefore exploited the retinotopic organization of the V1-to-RL projections. We performed IOI-targeted injections of the GABA-A agonist muscimol in caudal V1, which represents the upper visual field and projects to rostral RL (Wang and Burkhalter, 2007). We then recorded responses to upper and lower visual field stimulation in rostral RL, that preferentially responds to the upper field (Figure 9F), mafosfamide before and after fluorescent muscimol injection. The selective pharmacological blockade of caudal V1 was verified by IOI, and the diffusion of fluorescent muscimol was monitored under epifluorescence (Figure S7A). As expected by the topography of the V1-to-RL projection, selectively silencing caudal V1 reduced upper field responses in rostral RL (Figure 9G; medians: 3.8 versus 2.4 Hz; Wilcoxon rank-sum test; p < 0.001), whereas lower field responses were not affected (3.1 versus 3.1 Hz; Wilcoxon rank sum test; p = 0.82).