A thalamic network model was developed based on recent data regarding heterogeneous thalamic reticular (RE) cell axonal arborizations that indicate at least two projection patterns, short-range cluster projections and long-range diffuse projections. The model was constrained based on expected convergence and the biophysical properties of RE and thalamocortical (TC) cells and their synapses. The model reproduced in vitro synchronous slow (3-Hz) oscillatory activity and the known effects of T-channel blockade and cholecystokinin (CCK) application on this activity. Whereas previous models used the speed at which approximately 3-Hz oscillations propagate in vitro to infer the spatial extent of intrathalamic projections, we found that, so long as the gamma-aminobutyric acid-B synaptic conductance was adjusted appropriately, a network with only short-range projections and another network with both short- and long-range projections could both produce physiologically realistic propagation speeds. Although the approximately 3-Hz oscillations propagated at similar speeds in both networks, phase differences between oscillatory activity at different locations in the network were much smaller in the network containing both short- and long-range projections. We measured phase differences in vitro and found that they were similar to those that arise in the network containing both short- and long-range projections but are inconsistent with the much larger phase differences that occur in the network containing only short-range projections. These results suggest that, although they extend much further than do short-range cluster projections, long-range diffuse projections do not spread activity over greater distances or increase the speed at which intrathalamic oscillations propagate. Instead, diffuse projections may function to synchronize activity and minimize phase shifts across thalamic networks. One prediction of this hypothesis is that, immediately after a collision between propagating oscillations, phase gradients should vary smoothly across the thalamic slice. The model also predicts that phase shifts between oscillatory activity at different points along a thalamic slice should be unaffected by T-channel blockers and decreased by suppression of synaptic transmission or application of CCK.