The temporal relationship between sniffing and the limbic 19 rhythm was studied in rats during odor discrimination reversal learning. The I3 rhythm was monitored as rhythmic slow wave activity (RSA) in the dorsal hippocampal formation, and cyclic nasal airflow (sniffing) was monitored with a thermocouple in the nasal cavity. The training procedures required animals to perform a sequence of whole body locomotion toward one wall of an arena, followed by investigatory sniffing of stimuli through a port while otherwise standing still. Hippocampal RSA was present reliably during the periods of investigatory sniffing. Analyses based on the fast Fourier transform (FFT) demonstrated that this RSA tended to be lower in frequency and amplitude than RSA which occurred during locomotory approach. Other analyses based on the FFT were developed to characterize the nature and parameters of the temporal relationship between rhythmic sniffing and hippocampal RSA as a function of the dominant sniffing frequency during the periods of stimulus sampling. The phase difference between sniffing and RSA tended to vary linearly with frequency so as to maintain a preferred latency relationship between the onset of each sniff cycle and a particular phase of the hippocampal RSA. The phase of RSA to which sniffing was related differed across animals and was correlated with electrode position relative to the phase reversal layers within the hippocampal formation. These results therefore are consistent with the interpretation that, during the periods of stimulus sampling, the sniffs were being timed to maintain a preferred latency relationship with the pacemaker activity which drives the 8 rhythm and the recorded RSA.The consistency with which the animals exhibited the preferred latency relationship varied during the course of training. Across animals, this correlation between sniffing and 8 activity was consistently high during the trials which immediately preceded the achievement of criterion level performance, and the correlation was reduced during the criterion run and/or subsequent trials of overtraining. Thus, the tendency of the animals to exhibit this relationship was not associated specifically with correct performance. Rather, the correlation tended to be highest when the animals were most likely to be evaluating the behavioral relevance of stimuli and were in the process of modifying their responses to those stimuli. The timing of investigatory sniffs as a function of 6' cycle phase may be important for the neural processing of sensory and/or motor information of relevance for response modification.

No abstract