UCL Logo

Comparison of network entropy between healthy and 6-OHDA lesioned rats

Bruno Averbeck

Institute of Neurology, UCL, UK

Recordings from deep brain stimulation electrodes in the basal ganglia of patients with Parkinson's disease (PD) have shown that there are increased beta frequency oscillations in these brain structures. These oscillations are stronger when patients are off dopamine replacement medication than when they are on medication. Furthermore, when on medication, the oscillations decrease even further when patients initiate movements. This suggests that the increased oscillations in the basal ganglia networks may be related to the movement deficits seen in the patients. To study this at the level of single neurons, our collaborators have developed an animal model in which 6-OHDA is injected unilaterally into the substantia nigra pars compacta. This unilateral injection depletes much of the dopamine on the affected side. Multiple single neuron recordings in the globus pallidus have shown that exaggerated beta frequency oscillations develop on the affected side several days after the injection.

We have been characterizing the effects of these oscillations on the network entropy, by comparing entropy between ensemble neural responses in lesioned and normal animals. Correlations decrease entropy and therefore, there are decreases in entropy in the lesioned animals. However, there are reductions in firing rates and increases in oscillations and cross correlations between lesioned and healthy animals. Therefore, we wanted to examine which factors most decreased the entropy. We have used a logistic regression model to examine the relative effects of firing rate, oscillations and synchrony on the entropy in the network. The reduction in firing rates of the lesioned animals appear to have the largest effects, the oscillations the second largest effects, and the synchrony the smallest effects, in pairs of neurons simultaneously recorded. However, effects of synchrony are likely to be minimal in pairs, and larger when larger populations of neurons are considered. We are currently developing a model of the network effects of synchrony, at the population level, so its effects can be more directly compared to firing rates and oscillations.