No such bursting was observed in nonmitral cells recorded in close proximity to mitral cells. excitability at ~25 msec after spikes, as reflected by a peak in the interspike interval distribution and in the corresponding hazard function. About half also showed a peak at about 6 msec, reflecting the common occurrence of doublet spikes. Nonmitral cells showed no such doublet spikes. Bursts typically increased in intensity over the first 20C30 sec of a burst, during which time doublets were rare or absent. After 20C30 sec (in cells that exhibited doublets), doublets occurred frequently for as long as the burst persisted, in trains of up to 10 doublets. The last doublet was followed by an extended relative refractory period the duration of which was independent of train length. In cells that were excited by application of a particular odor, responsiveness was apparently greater during silent periods between bursts than during bursts. Conversely in cells that were inhibited by a particular odor, responsiveness was only apparent when cells were active. Extensive raw (event timing) data from the cells, together with details of those analyses, are provided as supplementary material, freely available for secondary use by others. t t by 50% and spikes no (R)-MG-132 longer follow the stimuli faithfully, apparently because of increased lateral inhibition. (C) Average peristimulus time histogram for 29 mitral cells showing consistent suppression of activity for ~150 LW-1 antibody msec after LOT stimulation (at 1 Hz for 5 min). The stimulus artifact and antidromic spike events are in the period covered by the red bar. The bar on the right of the histogram shows the mean SE spikes/bin for the 29 mitral cells in the period 200C250 msec after stimulation. (D) Orthodromic excitation of interneurons in the olfactory bulb following LOT stimulation. Variable latency spikes (arrowed) following LOT stimulation (red bar). (E) Poststimulus time histogram of cell shown in D; orthodromic action potentials occur at a latency of between 6.5 and 6.9 msec. = 29 cells tested), indicating that LOT\evoked inhibition is synaptically mediated. This involves lateral inhibition rather than recurrent inhibition alone because recurrent inhibition should be present (R)-MG-132 after spontaneous spikes as well as after antidromic spikes. We therefore looked at the responses of nonmitral cells recorded from the region of the mitral cell layer, to see if the timing of their activation corresponded to the timing of inferred inhibition of mitral cells. These presumptive interneurons responded variably to LOT stimulation, but of 23 cells tested for their responses to 1 1 Hz stimulation, nine were (R)-MG-132 strongly excited at a nearly constant short latency of 4C8 msec (intercell range; Fig. ?Fig.1D1D and E). The other cells were either unresponsive or had late excitatory or inhibitory responses. Thus, in the region of the mitral cell layer, some interneurons displayed a nearly constant latency to LOT stimulation at latencies only slightly longer than the range for antidromically identified cells, at a timing that could account for their mediating the inhibitory effects of LOT stimulation upon mitral cells. Phasic bursting of mitral cells Forty\seven of the 89 mitral cells in female rats fired spontaneously with long bursts separated by long silent periods; these cells had an intraburst firing rate of 14.3 1.1 (7.1C27.7) spikes/sec and an activity quotient (proportion of time active) of 50 3 (21C77)%; thus typically they spent as much time silent as they did active. The mean burst length was 122 10 (50C303) sec and the interburst time was 129 11 (39C251) sec. The other 42 mitral cells also fired in long bursts, but were not silent between the bursts; these had an intraburst firing rate of 13.0 1.5 (3.4C31.3).