Based on our neurophysiological findings, we developed a neural network model that performs feature linking via modulatory interactions. In the linked state, the activities of the model neurons are highly synchronized. The model's performance is demonstrated in an example of contrast-dependent region linking and in examples of region definition in the Craik-O"Brien-Cornsweet illusion.We believe that stimulus-dependent synchronizations as we found them in the cat's primary visual cortex are very likely to be a universal principle for associative linking in the central nervous system at stages ranging from pre-attentive linking up to higher levels of visual integrations and even up to associative linkings among other sensory modalities and memory.