Objective: To study the post-operative cognitive and psychomotor recovery from midazolam conscious sedation, after reversal with the benzodiazepine antagonist flumazenil over a prolonged recovery period. Design: A prospective, double-blind, randomised, crossover trial. Setting: Out-patient Sedation Department, Newcastle Dental Hospital and School Method: Eighteen patients, ASA I or II, received midazolam on two separate occasions to undergo equivalent dental treatment Following treatment patients were reversed with intravenous flumazenil or saline [placebo] at alternate appointments. Assessment of mood and cognitive function was undertaken using a highly sensitive and specific computerised battery of cognitive tests administered by telephone. Cognitive and psychomotor tests were administered prior to sedation and every hour for 6 hours post reversal. Results: Results indicated no significant effect of flumazenil on simple reaction time and choice reaction time but did show a trend of reversing the effects of midazolam on numeric working memory and word recognition. Conclusion: The cognitive and psychomotor effects of the sedation were not fully reversed by flumazenil. Cognitive impairments were still present up to 6 hours post-reversal, despite patients appearing clinically more alert. This has important implications for treatment protocols and discharge instructions.