Our previous work has shown that chronic haloperidol treatment decreases striatal symmetric synapses preferentially in rats which develop oral dyskinesias (vacuous chewing movements (VCMs)). The present experiment tests the hypothesis that olanzapine, which does not cause dyskinesia in humans or rats, would not cause the ultrastructural changes produced by haloperidol. After 6 months of treatment, VCM scores for the olanzapine group (5.1 +/- 4.5) were similar to those of controls (5.2 +/- 3.9), whereas rats in the haloperidol group were either nondyskinetic (4.3 +/- 2.2) or dyskinetic (16.9 +/- 6.7). The volume of the striatum (mm(3)), did not differ among the groups: control, 37.5 +/- 4.7; olanzapine, 36.4 +/- 4.3; haloperidol, nondyskinetic, 40.5 +/- 6.3; haloperidol, dyskinetic, 36.6 +/- 5.9. Synaptic density (per 1 microm(3)), obtained from the central region of the striatum, did not differ between the olanzapine (0.699 +/- 0.146) and control groups (0.652 +/- 0.108). The number of asymmetric synapses in the olanzapine group (0.624 +/- 0.136) was also similar to that of controls (0.550 +/- 0.090). The number of symmetric synapses in the olanzapine group (0.074 +/- 0.032) was not significantly different from that of controls (0.096 +/- 0.043). Thus, olanzapine, in contrast to haloperidol, did not produce dyskinesias or synapse loss. These results strengthen the correlation between the expression of VCMs and striatal synaptic changes and indicate that olanzapine has fewer behavioral and anatomical side effects than does haloperidol.