Effects of advanced candidate anticonvulsants under two rodent models of 'counting'.

The behavioral effects of a variety of advanced candidate anticonvulsants for organophosphate-induced seizures were evaluated under two rodent 'counting' models. Rats pressed the left of two levers a number of times (a 'run') before pressing the right lever. The targeted performance was a run of 12. The training contingency was a targeted percentile schedule, which provided food if the current run was closer to 12 than two-thirds of the most recent runs. Baseline performance was well controlled by the target, with mean run lengths slightly less than 12. Once this performance was acquired, half the subjects were switched to a procedure providing food following runs of different lengths with a probability yoked to previous percentile schedule performance. The two procedures generate comparable baseline performances, but behavioral disruptions generate reinforcement loss only under the yoked procedure. Atropine, scopolamine, azaprophen, aprophen, trihexyphenidyl, procyclidine, benactyzine, biperiden and diazepam were tested. All produced dose-related decreases in overall run length and response rate. Responding was disrupted more readily under the yoked procedure than under the percentile procedure. Only atropine affected responding at doses below those effective against soman-induced seizures. Of the present candidates, trihexyphenidyl, procyclidine, benactyzine and biperiden appear most promising for further development. Copyright 2001 John Wiley & Sons, Ltd.