, 2007). Thus, layer 5 pyramids could have a smaller margin C646 in vitro to further increase responsiveness

upon M stimulation compared to unimodal stimulation. Finally, regardless of the layer 2/3-to-5 projection, the possibility cannot be excluded that infragranular neurons could display ME in response to more complex or ecologically relevant stimuli. ME was scarcer and less common in Pv-INs than in pyramids. This could be due to the fact that AP responses of Pv-INs are less sensitive to an increase in the strength of synaptic inputs compared to pyramids (Tateno et al., 2004). The weaker ME of Pv-INs could originate during the conversion of PSPs into APs. However, we did not record subthreshold FK228 cell line activity from Pv-INs, so MI might be already weaker at the subthreshold level. In fact, Pv-INs have briefer excitatory currents and shorter EPSPs that integrate less strongly over time compared to pyramids (Angulo et al., 1999 and Thomson, 1997). The results of our optogenetic manipulation of Pv-INs show that these inhibitory cells play a permissive role in the MI of pyramidal cells, because the lack of ME in Pv-INs enables the pyramidal cells to effectively integrate the inputs of different sensory modalities. Indeed,

when we optogenetically induced an “artificial” ME at least in a subgroup of Pv-INs, we selectively disrupted ME in neighboring pyramids.

Albeit what causes ME in pyramids is the concurrent arrival of two distinct sensory inputs in pyramidal cells that have intrinsic integration capabilities, our data indicate that the lack of integration in Pv-INs is a necessary condition enabling ME in pyramids. One mechanism that could explain why photostimulation only promoted ME in a subgroup of Pv-INs is that adding some extra (photo)excitation during both uni- and multisensory stimulation might selectively Thalidomide favor M responses in those inhibitory cells with slightly more hyperpolarized membrane potentials. In these cases multisensory stimuli might elicit PSP responses significantly closer to the AP threshold than unisensory ones, and therefore promote higher ME. These results suggest that the integrative properties of the major class of interneurons are important in enabling ME in excitatory cells. Other interneuron subtypes integrate inputs in different ways, and thus may differentially modulate MI in the excitatory cortical network. For example, recent work showed that somatostatin-positive inhibitory cells are effective integrators of sensory-driven synaptic inputs in V1 (Adesnik et al., 2012), contrasting with our results for Pv-INs in RL. It is also interesting to compare our data on Pv-INs with the results obtained with optogenetic manipulation of Pv-INs in primary sensory cortices.