Over the past several decades, more than twenty different cortical areas have been recognized as processing loci of specific properties of visual scenes, such as forms, colours, motion, etc. At the same time, those investigations led to the conclusion that there are few, if any, places in the nervous system where the parcellated information of visual scenes is resynthesized to obtain a unified and coherent percept. Along this line, it has been suggested that synchronization of action potentials within a time-window of 1 to 5 ms between two or more neurons belonging to distant pools of cells may be an encoding process, which enables binding of various features of a single visual object. Binding, in general terms, is a process permitting the functional linkage of the distributed neuronal activity. Numerous previous studies have been carried out with multiunit recordings where several cells are recorded simultaneously. This method fails to disclose which cells participate in the synchronization process. Our investigations are aimed at further clarifying the processes leading to the formation of an encoding assembly by synchronization. Toward this end we investigated the modulation of synchronization between pairs of neurons sorted out from multiunit recordings. Our results show that synchronization may be observed in spontaneous activity. We also demonstrate that the amplitude of synchronization is modulated in relationship to the visual stimulus structure. Moreover, we found that, when considering one pair of cells, the spontaneous synchronization is high enough to mask stimulus-related-synchronization, whereas when other pairs of cells are added in the neuronal assembly, random synchronization is progressively outweighed, and the pattern of stimulus induced synchronization is progressively revealed.