Using a computer model to explore impairments of acquisition processes following ingestion of diazepam

Hypotheses about the information processes impaired in diazepam-induced amnesia were tested by fitting the output from a computer simulation of list learning to observed serial position curves and to overt rehearsal protocols. Twenty-four subjects received an average weight-relative dosage of 0.18 mg/kg oral diazepam; 24 subjects received placebo. Immediate free recall of 16-word lists was examined at 2- and 8-s presentation times. Subjects receiving diazepam recalled significantly fewer words than placebo subjects (diazepam = 6.77 +/- 2.39 words; placebo = 9.29 +/- 1.42 words); their memory impairment was greater at the 8-s than 2-s presentation time. Tests of nonlinear regression models based on computer simulations of list learning performance were consistent with the hypothesis that diazepam reduces rehearsal capacity and disrupts the formation or utilization of contextual and inter-item associations. Among these causes of diazepam-induced amnesia, the disruption of contextual associations appears most important. The results further suggest that quantitative modeling of memory data may complement traditional methods of inferring relationships between brain processes and cognitive dysfunction in amnesic states.