The Neural Basis of the Psychomotor Vigilance Task

RAPID PUBLICATION 100 95 75 The Neural Basis of the Psychomotor Vigilance Task Sean P. A. Drummond, PhD1,2; Amanda Bischoff-Grethe, PhD1,2; David F. Dinges, PhD3; Liat Ayalon, PhD1,2; Sara C. Mednick, PhD4; M. J. Meloy, PhD1,2 University of California San Diego, Department of Psychiatry, San Diego, CA; 2VA San Diego Healthcare System, Research Service, San Diego, CA; University of Pennsylvania School of Medicine, Department of Psychiatry, Philadelphia, PA; 4The Salk Institute for Biological Research, La Jolla, CA 1 25 5 0 3 Study Objective: To identify brain regions underlying the fastest and slowest reaction times on the Psychomotor Vigilance task (PVT) under well-rested conditions, as well as brain regions related to particularly poor performance after sleep deprivation. Design: Subjects took the PVT twice while undergoing functional magnetic resonance imaging: once 12 hours after waking from a normal night of sleep and once after 36 hours of total sleep deprivation (TSD). Session order was counterbalanced. Setting: UCSD J. Christian Gillin Laboratory for Sleep and Chronobiology (the sleep core of the General Clinical Research Center) and UCSD Magnetic Resonance Institute. Patients or Participants: Twenty right-handed healthy adults (8 women; age = 27.4 ± 6.7 years; education = 15.6 ± 1.5 years.) Measurements and Results: After a normal night of sleep, optimal performance was related to greater cerebral responses within a cortical sustained attention network and the cortical and subcortical motor systems. INTRODUCTION THE PSYCHOMOTOR VIGILANCE TASK (PVT) WAS ORIGINALLY DEVELOPED IN 1985 AS A MEASURE OF SUSTAINED ATTENTION.1 SINCE THAT TIME, A LARGE number of studies have demonstrated its sensitivity to sleepiness in clinical, experiment, and operational contexts,2 making it one of the most widely used neurobehavioral tests in studies of sleep and circadian rhythm research.2 As examples, studies have shown the PVT to be sensitive to, among other things, alterations in the homeostatic3-6 and circadian systems,4,7 as well as work schedules,8 age,9 and sleepiness countermeasures such as naps,10 bright light,11 and caffeine.11 PVT results are generally interpreted as reflecting the arousal and attentional state of the individual. Comparisons are made both across conditions (eg, well rested vs sleep deprived) and within a given test administration. Across conditions, the PVT has proven to be a very sensitive measure of sleep loss.3,5,12 With- 100 95 75 25 5 0 Disclosure Statement This was not an industry supported study. Dr. Drummond has research support from Cephalon and Aventis. Dr. Dinges has received research support from Cephalon; has had a relationship including directorships, trusteeships, management relationship or paid speaking engagements with Cephalon, Pfizer, Merck, and Sanofi-Aventis; and has received substantial gifts, discounted or free use of material or equipment from Cephalon. Drs. Ayalon, Bischoff-Grethe, Mednick, and Meloy have indicated no financial conflicts of interest. Submitted for publication June 2005 Accepted for publication July 2005 Address correspondence to: Sean P. A. Drummond, Department of Psychiatry, 9151B UCSD / VASDHS, 3350 La Jolla Village Dr., San Diego, California 92161; Tel: 858 642 1274; Fax: 858 458 4201; E-mail: drummond@ucsd. edu SLEEP, Vol. 28, No. 9, 2005 1059 Slow reaction times, particularly after TSD, were associated with greater activity in the “default mode network” consisting of frontal and posterior midline regions. Conclusions: Optimal performance on the PVT appears to rely on activation both within the sustained attention system and within the motor system. Poor performance following TSD may result from a disengagement from the task and related inattention, and brain regions responsible for this localize within midline structures shown to be involved in the brain’s “default mode.” Finally, particularly poor performance after TSD may elicit a subsequent attentional recovery that manifests as greater activation within the same regions normally responsible for fast reaction times. Keywords: Sleep deprivation, cognitive function, functional magnetic resonance imaging, PVT, sustained attention, default mode network Citation: Drummond SP; Grethe AB; Dinges DF et al. The neural basis of the psychomotor vigilance task. SLEEP 2005;28(9): 1059-1068. in a given test bout, changes in performance (ie, reaction time [RT]) may reflect what has been termed “state instability,” which refers to PVT performance variability produced by the influence of homeostatically controlled sleep-initiating mechanisms on the endogenous capacity to maintain alertness and utilize executive attention.2 State instability is evident in the waxing and waning of attention and arousal over time (ie, milliseconds to minutes), especially during periods of sleep deprivation.12 As a result, when an individual’s arousal and attention are relatively normal, so too is PVT performance. However, when attention and arousal wane due to elevated homeostatic sleep drive, PVT performance changes such that RTs reflect commingling of errors of omission (lapses) and errors of commission (false responses). Depending on the degree of sleep deprivation, the fastest RTs on the PVT do not change or change only modestly relative to the well-rested state, which reflects the fact that individuals experience instances of relatively normal attention and arousal levels even when sleep deprived. On the other hand, the slowest RTs can lengthen dramatically after sleep deprivation, reflecting instances when individuals experience markedly reduced levels of attention and arousal.2,12 Despite the plethora of behavioral data collected with the PVT, there have been no published functional neuroimaging studies examining the cerebral correlates of PVT performance. Thus, little is known about the cerebral substrates underlying fast and slow performance on the PVT. Reports of electroencephalographic (EEG) measures of brain function related to PVT performance have either not collected the EEG and PVT performance data simultaneously6 or have not reported correlations between EEG and PVT performance measures.8 Some possible insight into brain systems subserving PVT performance can be gained, though, by examining related literature on attention. For example, neuroimaging studies during well-rested states have revealed a network related to sustained attention that includes the Neural Basis of Psychomotor Vigilance Task—Drummond et al right middle frontal gyrus and inferior parietal lobe, as well as, possibly, the left inferior parietal lobe and bilateral thalamus.1316 Neuroimaging studies of (nonsustained) attention tasks during sleep deprivation have reported greater responses in either the bilateral ventral lateral thalamic nuclei (for short-term attention)17 or right prefrontal and left inferior parietal regions (for divided attention)18 after total sleep deprivation (TSD). One (...truncated)