Rats were administered equivalent doses of haloperidol for either 28 days or 8 months using one of two different drug regimens: intermittent (i.e., weekly injections) or continuously (via drinking water and osmotic mini-pumps). Oral movements were determined by human observers and by a computerized video analysis system, which determined number and amplitude of jaw openings and closings (computer-scored movelets "CSMs") as well as the slope (amplitude/duration) and frequency spectrum (fourier transform) of oral activity. The two drug groups developed distinctively different changes over time. Continuous administration resulted in late-onset oral activity changes at 1-3 Hz and withdrawal increases in CSMs, a pattern expected of tardive dyskinesia. Intermittent treatment produced a primed dystonia-like pattern: large amplitude CSMs which had steep onset slopes and a peak energy at 4-7 Hz. These results demonstrate the importance of drug regimen in determining the type of neuroleptic-induced dyskinesias which develop with prolonged neuroleptic treatment in rodents.