According to many theories of brain function, the computational power of the brain depends upon the number of independent synapses it contains. A synapse will not be independent if its receptors are activated or modified by neurotransmitter released at neighbouring synapses. Recently, there have been several reports suggesting the occurrence of 'crosstalk' or 'spillover', and a large number of results consistent with crosstalk. However, the quantitative importance of this phenomenon remains uncertain. We estimate the significance of crosstalk using a simple model which predicts that, during concentrated synaptic activity, crosstalk between distinct synapses is likely to activate high-affinity receptors and may also desensitize certain receptors. Comparison of these predictions with the experimental data highlights the information that is required for a more detailed model of crosstalk.