AbstractZonisamide (Zonegran®) is a broad‐spectrum antiepileptic agent that shares some pharmacological properties with other anticonvulsants, including phenytoin, carbamazepine, and valproic acid, but is differentiated from these agents by the ability to significantly block T‐type calcium channels. Zonisamide interacts with the γ‐amino‐butyric acid (GABA) receptor in an allosteric manner, and thus does not modulate GABA receptor effects. However, given the potential of drugs within the latter class for drug abuse in humans, an evaluation of zonisamide for abuse potential is an important component of its potential side‐effect profile. In the present study, zonisamide was tested in animal models of the subjective and reinforcing effects of central nervous system (CNS) depressant drugs, e.g., diazepam discrimination in rats and intravenous self‐administration in rhesus monkeys, respectively. In addition, zonisamide was evaluated for physical dependence liability in a chronic infusion model using rats. Zonisamide did not substitute for diazepam in rats trained to discriminate 2.5‐mg/kg diazepam from vehicle nor was it self‐administered by rhesus monkeys experienced in methohexital‐reinforced responding. Continuous infusion of zonisamide (400 or 600 mg/kg/day) did not prevent the loss of body weight associated with discontinued pentobarbital infusion. These doses of zonisamide did produce some incomplete attenuation of observable signs of pentobarbital withdrawal, likely due to direct sedative or depressant effects of these high doses. These results suggest that zonisamide would not produce diazepam‐like intoxication in humans nor would it likely be subject to abuse when made more widely available. Further, when administered chronically, zonisamide would not be expected to produce physical dependence of the CNS depressant type. Taken together, these results support the prediction that zonisamide would have low abuse liability. Drug Dev. Res. 54:66–74, 2001. © 2001 Wiley‐Liss, Inc.