Reproducibility of Changes in Behaviour and fMRI Activation Associated with Sleep Deprivation in a Working Memory Task

Abstract

Although the stability of inter-individual differences in vulnerability to sleep deprivation has been shown behaviourally, the neural basis for these differences has yet to be uncovered. In this study, we assessed the reproducibility of fMRI activation and performance on a working memory task before and after 24 hours of sleep deprivation (SD). All volunteers underwent 2 sessions (pairs of fMRI scans) at rested wakefulness (RW) and after SD. 19 healthy, right-handed subjects (mean age = 21.37 +/- 1.54 years). Brain activation was highly correlated across sessions in a frontoparietal network previously implicated in working memory function. The magnitude of decline in this activation after SD was preserved in bilateral parietal regions. Among several behavioural metrics investigated, the most robust marker of vulnerability to SD was the change in the intra-individual variability of reaction times. This was shown to be both stable over time and correlated with the drop in left parietal activation from RW to SD in both experimental sessions. Because of its reproducibility, the modulation of parietal activation may provide a good physiological marker of vulnerability to SD.