The Use of Euclidean Distance in 64-space and Behavioral Correlates to Optimize Filters for Gamma AM Pattern Classification

Abstract

Experiments with untrained animals showed that bulbar responses to novel odorants that were presented in sessions without reinforcement failed to develop novel AM patterns. These stimuli initially elicited an orienting response in company with replacement of gamma bursts by broad spectrum fluctuation resembling background activity. This suppression of bursts disappeared after 3 or 4 trials. If the novel odorant was followed by reinforcement, burst suppression gave way to renewed formation of gamma bursts also within 3 or 4 trials. My search for eventrelated content in bulbar gamma activity was guided by two hypotheses. First, when unconditioned stimuli were paired with one of two discriminative conditioned stimuli (the CS+) but not the other (CS-), resulting in a conditioned response (CR) to the CS+, and not to the CS-, randomly alternated in the same training session, odorant-related content must exist in the form of a pattern of neural activity in the olfactory bulb between the arrival of a CS at the nose and the performance of a correct CR. Second, this pattern elaborated by synaptic currents of the granule cells would be detected in field potentials in the bulbar AM patterns of gamma activity.