Many IB cells do not receive direct VPM inputs whereas most RS cells do (Agmon and Connors,

1992 and Baranyi et al., 1993), which is consistent with the greatest changes Screening Library in the initial component of the wPSP occurring in RS cells. VPM inputs (Wimmer et al., 2010) may also underlie the LVb depression observed in vivo 3 days after trimming, when intracortical circuits haven’t yet depressed. Inhibition can also be activated early in the PSP by thalamic input (Gil and Amitai, 1996) and so could also conceivably be involved in modulating the amplitude of the direct thalamic drive to the cells. Inhibition could have a differential impact on RS and IB cells since RS cells are thought to receive more inhibitory control than IB cells (Schubert, Staiger et al., 2001). Testing which of these pathways are involved will require specific further studies. For practical reasons we were obliged to use different methods to classify RS and IB cells in the in vivo and ex vivo aspects of the study. Classically, injection of current at threshold elicits a complex of action potentials in IB cells and a single action potential in RS cells, which is sufficient to classify them (McCormick et al.,

1985) and this method was used in the ex vivo studies. In the in vivo studies, we used this method in addition to identifying the action potential find more (AP) complexes that are characteristic of IB cells where the APs decrease in amplitude and ride upon a slow depolarization envelope (Baranyi et al., 1993, Connors et al., 1982, Ergoloid Dégenètais et al., 2002, McCormick et al., 1985 and Nuñez et al., 1993) but see Nowak et al. (2003) for a quantitative analysis. Both methods are judged to be equivalent and indeed,

in support of this view we described in the Experimental Procedures section different post hoc tests confirming that RS and IB cells have comparable properties in both our in vivo and ex vivo studies. However, the following qualifications need to be kept in mind: Schwindt and colleagues have reported high threshold bursting neurons that do not fire bursts at threshold current injections for producing APs (Schwindt et al., 1997 and Steriade et al., 1993); such neurons would likely be included in the bursting neuron population in vivo but not ex vivo. Conversely, while slow depolarizations in vivo can be generated intrinsically they can also result from the activity of synchronized inputs, as occurs during the generation of up-states. Moreover, synaptic activity, release of neuromodulators in vivo or differences in physiological temperature can preclude or obscure the occurrence of bursts in IB cells (Wang and McCormick, 1993, Waters, 2011, Steriade, 2001 and Steriade et al., 1993).