Evidence for synchronised responses in the piriform cortex by using Gibbs potential analysis

The piriform cortex is a large paleocortical area which receives direct projections from the olfactory bulb. In order to study the spatiotemporal distribution of the piriform cortex activity, we chose optical recording of the responses evoked by olfactory bulb electrical stimulation. Such a stimulation elicited a large signal corresponding to cortical reactivation (disynaptic activity) via intrinsic association fibres. As the disynaptic activity was observed over the entire piriform area, we wondered whether or not this redistribution contributes to a synchronisation of the activity in the piriform cortex. In order to answer this question, we developed a statistical approach which allows us to take the temporal dimension into account. The analysis was performed by using the Gibbs potential analysis. The neural response of the diode is represented by a stochastic point process (occurrence of latency peak), and the response of the diode array is given as successive realisations of a binary random field defined on a finite set. The Gibbs measure associated with this field is then estimated through the interaction potentials of the field's configurations, which provide a quantitative evaluation of the interaction and the synchronisation between the neural sites. The analysis was performed on the latency of the peak of disynaptic activity, which was determined from signals from 60 different acquisitions realised with the same stimulus parameters. From these 60 files of latency values, we estimated the Gibbs interaction potential of singletons and pairs. The former gave an image of the spatiotemporal distribution of the disynaptic activity, which appears to propagate from the anterior to the posterior part of the area recorded. The estimation of the interaction potential of pairs allows us to characterise the degree of synchronisation between two neighbouring recording sites. It appeared that, in the anterior half of the area recorded, the disynaptic activity was statistically desynchronised whereas, in the posterior part the disynaptic activity appeared strongly synchronised. The functional implications of such a spatiotemporal distribution of the activity are discussed.