Abstract Body

Neural oscillations are thought to provide a temporal structure facilitating information processing and inter-area communication. These oscillations coexist in different frequency bands involving neural networks at different spatial scales and are related to different aspects of behavior.

We questioned the effect of learning on the dynamics of neural oscillations and the evolution of the associated neural networks. We tackled this question during an olfactory discrimination task in rats. Local field potentials were recorded in a large set of brain areas (including sensory, motor and limbic areas). We focused our analysis on the neural networks defined by rhythm synchronization in two frequency bands: (i) slow rhythms (1-10Hz), including animal respiration frequency and hippocampal theta rhythm, and (ii) beta band (15-30Hz), known to be modulated by odor learning.

For the slow rhythms, preliminary analyses during odor sampling showed that hippocampal and respiratory rhythms were in a close frequency range only during a short window allowing for phase synchronization (1 or 2 cycles). Functional connectivity showed that these slow rhythms were supported by two distinct but overlapping networks which evolved during initial learning but stabilized as soon as the task rules were acquired.

In the beta band, a large oscillation emerged at the end of odor sampling as the rat was learning the task. This oscillation was coherent across all recorded areas where it was expressed and when the context learning was acquired. In addition, rat behavior seemed no longer plastic when beta was strongly expressed.