Each recording period consisted of 10 min of spontaneous activity, followed by 20 min of tactile stimulation, and then another 10 min of spontaneous activity. The tactile stimulation consisted of 600 repetitions of 1 s stimulation at 20 Hz followed GS-7340 chemical structure by 1 s

without stimulation. The tactile stimulator consisted of a plastic rod attached at one end to a membrane of a speaker controlled by a computer. The other end of the rod was placed in contact with left hind limb. For auditory stimulation in anesthetized animals, the time course of experimental protocol was similar to that for tactile experiments in S1, and it is illustrated in Figures 5A–5D. After 10 min of recording spontaneous activity, tones were presented for 0.5 s interspersed with 1 s of silence. This timing allowed for more off-to-on transitions of tones, which evoked the greatest response than would be possible with the same period using tones of 1 s duration. Thus, 800 repetitions of tone stimuli were presented in the 20 min

stimulation period. For each experimental condition, we used a different tone frequency during stimulation (urethane only: 1 kHz; tail pinch or carbachol: 1.5 kHz; amphetamine: 2.2 kHz; MK801: 3.2 kHz). For experiments with awake, head-restrained rats, auditory stimulation was presented for over 40 min in each animal. The pattern of stimulation consisted of repetitions of tones for 1 s followed by 1 s of silence. Activity occurring 200 ms after stimulus offset and before the next stimulus onset was regarded as spontaneous. Stimuli consisted of pure tones tapered at the beginning and the end with a 5 ms cosine window. In Apoptosis Compound Library clinical trial data sets from awake animals, we did not have extended spontaneous periods these preceding or following stimulation period. Experiments took place in a single-walled sound isolation chamber (IAC) with tones presented free-field (RP2/ES1, Tucker-Davis). In order to quantify temporal relations among neurons, we calculated the mean spike latency as described previously (Luczak et al., 2009). Briefly, for each neuron, latency

is defined as the center of mass of a cross-correlogram of that neuron with the summed activity of all other simultaneously recorded cells (multiunit activity [MUA]) within a time window of 100 ms (Figure 2A). Before calculating the center of mass, cross-correlograms were smoothed with a Gaussian kernel with SD = 5 ms and normalized between zero and one to discard effects of baseline activity. Thus, this measure estimates the time when the corresponding neuron is most likely to fire with respect to the population activity. In addition to analysis of cross-correlograms between single neurons and multiunit activity as described above, we also calculated latency from pair-wise cross-correlograms to look at temporal relations between neurons in more details (Figures 2E, white bars, and 6F–6O).