Reafference and attractors in the olfactory system during odor recognition

L M Kay,L R Lancaster,W J Freeman

DOI: https://doi.org/10.1142/s0129065796000476

Abstract:Olfactory bulb activity has been postulated to be chaotic, as measured in the EEG, and to be subject to an attractor with many "wings" enabling classification of different learned odor classes. Two parallel questions are thus addressed by the work presented here: (1) what is the evidence for attractors in the olfactory system, which can mediate learned odor classes? and (2) how does the olfactory system enter a specific attractor or attractor wing associated with the learned odor during the classification process? Both of these questions address the wider notion of endogenous activity preparing the system for an expected stimulus, which is at the basis of the reafference principle. By viewing the brain as a distributed complex dynamical system with global attractors, these questions can be answered together. Rats were implanted with bipolar macroelectrodes in the Olfactory Bulb (OB), Prepyriform Cortex (PPC), Entorhinal Cortex (EC), and Dentate Gyrus (DG), and then trained in an operant paradigm to press a bar for a reward in the presence of one odor and to receive no reward in the presence of another odor. Local Field Potentials (LFP) were recorded simultaneously from the structures during the operant task. We present evidence for three endogenous events: (1) preafference, which is manifested both by the EC entering an attractor and a mid-range signal (15-30 Hz) which appears to be passed from the EC to the OB just before the OB enters an attractor; (2) afference, where the OB enters an attractor during the odor recognition period of the experiment and the LFP recordings indicate that the OB drives the other structures in all frequency bands, especially the high gamma band (65-100 Hz) associated with the OB burst frequency; and (3) reafference or post-afference, which is accompanied by a lower frequency gamma band signal (40-60 Hz) originating in the PPC and passed to both the OB and the EC just before the onset of the motor response to the odor. We use a new method, NECTAR (Nonparametric Exact Contingency Table Association Routine), related to mutual information, to verify what is seen with coherence and phase estimates, the apparent driving of each structure at different times in the odor trials, and to display evidence for non-periodic attractors governing both individual physiological structures and the system of structures. This is the first evidence of an endogenous, limbic event associated with sensory perceptual tuning in a mammal. These results are also the first experimental confirmation that the attractors governing olfactory activity involve multiple sites in the olfactory/limbic system and implement the process of attention.