A New Approach to the Analysis of Multidimensional Neuronal Activity: Markov Random Fields

How can information hidden in a spatial configuration of neuronal activity be addressed? The Markov Random Field method for the analysis of the spatial component of a multidimensional neuronal process is introduced and after simulations is applied to experimental data on rat at olivocerebellar activity. Using this method it was determined, for the first time, that the activity demonstrates dynamic coupling and may have different fine spatial substructures. The results obtained support the view that the inferior olive serves as a movement organizing centre that controls motor activity by means of spatially as well as temporally organized patterns of coherent activity. Copyright 1997 Elsevier Science Ltd.