Decreased Cortical Response to Verbal Working Memory Following Sleep Deprivation

STUDY OBJECTIVE: To investigate the cerebral hemodynamic response to verbal working memory following sleep deprivation. DESIGN: Subjects were scheduled for 3 functional magnetic resonance imaging scanning visits: an initial screening day (screening state), after a normal night of sleep (rested state), and after 30 hours of sleep deprivation (sleep-deprivation state). Subjects performed the Sternberg working memory task alternated with a control task during an approximate 13-minute functional magnetic resonance imaging scan. SETTING: Inpatient General Clinical Research Center and outpatient functional magnetic resonance imaging center. PATIENTS OR PARTICIPANTS: Results from 33 men (mean age, 28.6 +/- 6.6 years) were included in the final analyses. INTERVENTIONS: None. MEASUREMENTS AND RESULTS: Subjects performed the same Sternberg working memory task at the 3 states within the magnetic resonance imaging scanner. Neuroimaging data revealed that, in the screening and rested states, the brain regions activated by the Sternberg working memory task were found in the left dorsolateral prefrontal cortex, Broca's area, supplementary motor area, right ventrolateral prefrontal cortex, and the bilateral posterior parietal cortexes. After 30 hours of sleep deprivation, the activations in these brain regions significantly decreased, especially in the bilateral posterior parietal cortices. Task performance also decreased. A repeated-measures analysis of variance revealed that subjects at the screening and rested states had similar activation patterns, with each having significantly more activation than during the sleep-deprivation state. CONCLUSIONS: These results suggest that human sleep-deprivation deficits are not caused solely or even predominantly by prefrontal cortex dysfunction and that the paretal cortex, in particular, and other brain regions involved in verbal working memory exhibit significant sleep-deprivation vulnerability.