Benztropine mesylate (intravenous [IV] and oral) challenge was compared with brief neuroleptic withdrawal on dyskinesia ratings and symptom measures. Thirty-six neuroleptic-treated patients underwent a placebo-controlled acute IV challenge with 2 mg benztropine and a placebo-controlled two-week trial of oral benztropine mesylate (2 mg three times a day), followed by a double-blind placebo-controlled neuroleptic withdrawal involving four weeks of dose tapering and six weeks of placebo treatment. Benztropine given IV had no significant effect. Orally administered benztropine, however, led to statistically significant increases in dyskinesia and dysphoric mood. The brief neuroleptic withdrawal significantly increased dyskinesia scores and dysphoria and resulted in early termination of therapy in 12 of 36 patients (33%) due to symptom exacerbation. There was a striking absence of correlation between dyskinesia change measures brought about by benztropine and changes following neuroleptic withdrawal. Therefore anticholinergic challenge does not appear to be a fruitful procedure for identifying patients with covert dyskinesia.