Modelling coupled oscillations in neural populations during cognitive processing

Oscillatory neural activity is a pervasive feature of both invertebrate and vertebrate nervous systems. There is a growing body of evidence demonstrating that oscillatory activity at various scales is correlated with behavior in a task-dependent manner. This has led to the hypothesis that oscillatory neural activity is produced and dynamically modulated by the nervous system in order to execute various functions. Oscillations provide the ability to filter information based on its frequency content and therefore can facilitate closed channels of communication and provide protection to a signal from intrinsic noise and distracting inputs. This filtering capacity is investigated in two computational models of decision making. Predictions of the models are investigated in local field and spike train data recorded from rodent prefrontal cortex and hippocampus during a spatial working memory based T-maze task. These two structures exhibit transient increases in theta frequency coupling at time points relevant to the decision, consistent with theta frequency oscillations being used to communicate task relevant information. An example of a prefrontal interneuron that fires rhythmically at theta frequency, as predicted by one of the models, is found. Hippocampus and prefrontal cortex have also been shown to demonstrate coupling of oscillatory activity occurring at two different frequencies, specifically theta (4-12 Hz) and gamma (30-100 Hz). A non linear firing rate model of neural populations, which produces gamma frequency oscillations with amplitude modulated by the phase of a theta frequency input, is presented. This system demonstrates both a Hopf and a Saddle-Node-on-Invariant-Circle (SNIC) bifurcation. To look for evidence of this type of coupling in the rodent data, three published algorithms for detecting such coupling are compared; the Envelope-to-signal Correlation (ESC), Modulation Index (MI) and Cross-frequency Coherence (CFC). Each measure shows superior performance on one particular type of simulated data. Analysis of data from the T-maze task shows evidence of coupling of hippocampal gamma to both hippocampal and prefrontal theta activity. This activity appears to differentiate between correct and error trials and between choice and forced trials on the task.EThOS - Electronic Theses Online ServiceGBUnited Kingdo