The amnesic properties of benzodiazepines result from an impairment in expl
icit (conscious) acquisition of new material. Rationale: Explicit encoding
of new material has consistently resulted in an increase in regional cerebr
al blood flow (rCBF) in the left prefrontal cortex, as measured by positron
emission tomography (PET). Objective: PET was used to determine whether an
amnesic dose of midazolam (0.075 mg/kg) attenuated activation in this area
during explicit memory encoding. Methods: A second condition (condition A)
used a task to control for the automatic processing that occurs during exp
licit learning (condition E). Results: The subjects who received midazolam
(n=7) recognised significantly fewer words than those who received placebo
(n=8), but were not impaired with regard to automatic processing. rCBF was
significantly increased in the left prefrontal cortex during explicit encod
ing of word lists in all subjects and in the temporal lobe and parieto-occi
pital regions during automatic processing, rCBF was significantly decreased
in the prefrontal, superior temporal and parieto-occipital regions followi
ng midazolam. The midazolam-induced deactivation in the prefrontal cortex d
id not affect rCBF activations induced by the explicit memory condition (E-
A). Conclusions: These results suggest that a specific interaction with pre
frontal cortex activation does not underlie the amnesic effect of midazolam
. However, it remains possible that a threshold level of prefrontal rCBF is
necessary for encoding and that, after midazolam, this was not reached.