Identification of Individuals Susceptible to Sleep Deprivations

This investigation validated the use of the Alertness and Memory Profiler and B-Alert algorithms in objective measurements of drowsiness and the capability to identify those most susceptible to acute sleep deprivation.

In this study 24 college-age healthy subjects were tested repeatedly during 44-hours of prolonged sleep deprivation to create changes in the level of alertness that permitted within-subject comparisons of alertness.

The graphs above show that percent correct and reaction time are inversely related measures of the impact of continuous sleep deprivation. The performance measures were very responsive to the 40-minute nap that was provided after 32 hours of sleep deprivation (at 1600 hours). Performance measures were independently used in a cluster analysis to categorize individuals into three groups based on susceptibility to sleep deprivation. The graph in the upper right shows the mean + SE of the canonical correlations between the B-Alert EEG drowsy classifications compared to performance measures (reaction time and percent correct) and technician-observed signs of drowsiness (eye closures, head nods) across the 10 time points. In the graph to the lower left, the percentage of epochs classified as drowsy are plotted with the associated reaction times (RT) for the three susceptibility groups. Note how well both measures relate across time for the three groups. The graph to the lower right shows that there was no difference in perceived sleepiness across the three susceptibility groups even though their performance was dramatically different.

These results indicate the B-Alert algorithms were capable of identifying fatigue resulting from a lack of sleep that leads to poor decision making. The combination of the B-Alert classifications and neuro-behavioral measures were demonstrated to be capable of identifying individuals whose performance is most susceptible to sleep deprivation. The fact that the subjective measures of sleepiness did not distinguish between the groups suggests that the subjects’ perception of their own level of fatigue did not correspond to the objective measures.

Berka, C., Levendowski, D. et al. (2005). EEG Quantification of Alertness: Methods for Early Identification of Individuals Most Susceptible to Sleep Deprivation. Proceedings of SPIE Defense and Security Symposium, Biomonitoring for Physiological and Cognitive Performance during Military Operations, Orlando, FL, SPIE: The International Society for Optical Engineering.

Objective Quantification of Daytime Drowsiness and Disease

The Alertness and Memory Profiler (AMP) provides assessment of physiological and neurocognitive factors using a non-invasive integrated and synchronized test battery (3-Choice Vigilance Test, Image Recognition (IR), IR with Interference, Verbal/Number-Image Paired Associate Learning, and Sternberg Verbal Memory Scan).

This approach facilitates rapid data collection from any size population, enabling large scale epidemiological studies, disease diagnosis or treatment outcome evaluations. The AMP is sufficiently rugged and robust for studies to be conducted in operational environments (e.g., in the desert at 29 Palms) and its metrics have been proven useful in predicting susceptibility to sleep deprivation or marksmanship skill acquisition in US Marines.

One of the AMP metrics called the ACES score can be derived with less than one-hour of testing. It uses EEG and performance measures to discriminate healthy controls from obstructive sleep apnea (OSA) patients prior to treatment, providing a sensitivity of 0.88 and a specificity of 0.96 (n=396). Eighty-three percent of OSA patients had an ACES score of 7 or greater and 95% of the healthy subjects had a score of 4 or less. The graph to the left shows the change in ACES in OSA patients pre- and post-CPAP or Mandibular Advancement Device (MAD), and repeated testing in healthy subjects.

A 2-hour AMP test battery allows neurocognitive factors scores (NCFS) to be calculated for visual-spatial processing speed (VSPS), sustained attention (SA), recognition memory accuracy (RMA) and recognition memory speed (RMS). The graph to the right compares NCFS across Rested and Sleep Deprived Healthy subjects, OSA patients from MAD and CPAP studies and a third group of patients who underwent full polysomnography followed by Maintenance of Wakefulness Test (MWT) at the New York University School of Medicine.

The AMP Drowsy Detector combination of EEG, EKG, and performance indices from a four-task, one-hour battery to provide a highly sensitive and specific classifier to accurately discriminated daytime drowsiness due to sleep deprivation or sleep disorders. The graph to the left compares the AMP to the MWT and Epworth Sleepiness Score. The AMP is 20% more specific than the MWT in identifying fully rested healthy subjects, 20% more sensitive in identifying sleep deprived healthy subjects and 51% more sensitive in identifying patients sleep deprived as a result of sleep disordered breathing. These data show the ESS is insensitive as a measure to identify sleep deprivation.

 

Berka C., A. I., Burschtin O. et al. (2009). “High Throughput Brain-Behavior Assay: Quantification of EEG and Performance in Patients Referred for Assessment of Daytime Drowsiness.” Sleep 32: A383.

Kintz, N., Johnson, R., et al. (2009). “Mandibular Advancement Devices Significantly Improve Neurocognitive Function for Patients with Obstructive Sleep Apnea.” Sleep 32: A229

Continuous Monitoring of Workload over 24-Hour Sustained Operations

For this application the EEG headset was modified so the sensors could be worn under a Kevlar battle helmet. The patented B-Alert artifact identification and decontamination algorithms were applied (e.g., eye blinks, gross head movement, etc.) to the signals and the decontaminated signals and associated power spectra density values were transferred in real time to a third party for their use in a proprietary measure of cognitive effort.

Six members of a 32-member National Guard platoon wore the EEG headset intermittently during a 24-hour sustained training operation at a United States Army training and testing facility. The platoon leaders, platoon sergeants and squad leaders were selected because of the expected variability in cognitive effort imposed as a function of their tactical combat activity and interaction with electronic communication technology. During exercises of entering and clearing buildings in an urban environment with simulated enemy forces and simunitions, approximately 75% of the data were acceptable for analyses using the B-Alert Headset and algorithms.

Mathan S., W. S., Dorneich, M. et al. (2007). Neurophysiological estimation of Interruptibility: Demonstrating Feasibility in a Field Context. Augmented Cognition. D. Schmorrow, Nicholson D., Drexler J., Reeves L. Arlington, VA, Strategic Analysis, Inc: 51-58.

Impact of Closed-Loop Drowsiness Feedback on Driving Performance

In this application, real-time EEG classifications of engagement and drowsiness were used to trigger feedback in a closed-loop time/performance-locked driving simulator training session.

In a randomized cross-over design, partially sleep-deprived subjects performed four driving simulator scenarios over an eight-hour evening period. During two of the sessions, audio feedback was initiated when EEG metrics of extreme drowsiness were detected with the intensity and duration of the feedback modulated to match changes in the EEG. Six unique sounds were selected for the alarms; the feedback sounds became more urgent each time the alarm was triggered. Performance was monitored during the other two sessions without feedback.

When feedback was provided subjects showed statistically significant improvements in reaction times and correct responses to the driving simulator divided attention task. Although there was substantial variability in driving performance resulting from sleep deprivation, the number of drifts and veers and total accidents showed significant reductions when participants were most fatigued and feedback was provided. Most participants reported that the feedback alarms were beneficial in helping them maintain alertness.

Real-time closed-loop feedback based on EEG-indices holds the potential for a number of training or safety applications. Vibro-sensory motors are currently being explored as a means to provide tactile feedback, a novel alternative to audio or visual feedback that can be generated in a confidential, non-cognitive distracting manner.

Berka, C., Levendowski, D. et al. (2005). Implementation of a Closed-Loop Real-Time EEG-Based Drowsiness Detection System: Effects of Feedback Alarms on Performance in a Driving Simulator. 1st International Conference on Augmented Cognition, Las Vegas, NV.

Operational Profiling Sleep Deprivation and Stress

U.S. Marine Corps troops experience combined stressors including sleep deprivation, physical exertion and threat of enemy fire that can impair vigilance and decision-making with potentially dangerous consequences.

In this study USMC battalion/platoon leaders (n=17) were evaluated during a 28-day, continuous live-fire training exercise. Prior to the start of the exercise and then once a week wireless EEG and EKG were acquired in the field during a 20-minute, 3-Choice-Vigilance-Test (3C-VT). Physiological measures of engagement, distraction/drowsiness, heart rate variability, performance were assessed along with self-reported stress, fatigue and mood were assessed with Profile of Mood States, Stanford Sleepiness Scales, Brief Fatigue Inventory and Perceived Stress Scale.

Repeated measures analysis of variance revealed significant interaction effects (p<.0001) across quartiles over time in the 3C-VT with increased Distraction/Drowsiness, decreased High-Engagement, decreased accuracy and increased reaction times across weeks of training. Heart rate variability, suggesting increased levels of stress, also increased significantly. Significant changes in self-report measures included only decreased POMS-Vigor.

This study confirmed the Marines also exhibit trait characteristics in susceptibility to sleep deprivation. The graph to the left shows a change in reaction times (histograms) and percentage of incorrect responses (lines) across weeks during the 28-day exercise. After normalizing the data to account for the fact that Marines as a group tend to be less susceptible to sleep deprivation than the general population, we found that speed and accuracy were highly affected in 11% of the Marines by the end of the training exercise. Conversely, performance in 35% of the group was minimally affected by sleep deprivation.

A closer investigation of the group with average susceptibility revealed a pattern that should be taken into consideration by those responsible for managing fatigue and situational awareness. The graph to the left shows that the speed of decision making (reaction time) increased during sustained operations which resulted in chronic sleep deprivation. The increase in decision making speed, however, contributed to a doubling in the number of incorrect decisions. Yet none of the questionnaire responses indicated that the Marines acknowledged the impact of the fatigue on their capabilities.

Berka, C., Davis, G. et al. (2007). “Psychophysiological Profiles of Sleep Deprivation and Stress during Marine Corps Training.” leep 30: A132

Effect of Nicotine and Withdrawal on Attention, Memory and Workload

Despite the widely publicized health risks of smoking, an estimated 25% of the U.S. population and between 25% - 37% of military personnel continues to smoke. The use of nicotine as a fatigue countermeasure may be problematic in operational environments because the stimulant properties may increase speed while sacrificing accuracy.

In this randomized cross over design, the Alertness and Memory Profiler’s Verbal Memory Scan (VMS) and, Image Learning and Recognition with Interference (ILR-I) tasks assessed neurophysiology and performance in cigarette smokers following 14 mg. transdermal nicotine administration vs. placebo nicotine patch. The results show that nicotine induces an excessive allocation of attentional engagement. After just 12 hours of nicotine withdrawal, fully-rested and otherwise healthy participants evidenced significant increases in EEG drowsiness and impaired performance on learning and memory tests.

Berka, C., D. Levendowski, et al. (2006). Nicotine Administration and Withdrawal Effects on EEG metrics of Attention, Memory and Workload: Implications for Cognitive Resource Allocation. Augmented Cognition: Past, Present and Future. D. Schmorrow, K. Stanney and L. Reeves. Arlington, VA, Strategic Analysis, Inc.: 174-183.

Mitigating Sleep Deprivation with Omega-3 Fatty Acids

The link between Omega-3 fatty acids and enhanced stress resiliency, memory, and performance was assessed in 30 subjects who completed two trials of 48-hour acute sleep deprivation.

One arm of the double blind cross-over design was on high dose Omega-3 and the other on placebo. The Alertness and Memory Profiler was used to objectively assess outcomes. The graphs to the left show that performance, measured by percent correct and reaction time during a 3-choice vigilance task, was substantially better when the subjects were on high dose Omega-3. High performance was maintained for approximately 39-hours after the start of the sleep deprivation. The amplitudes of the memory event related potentials (ERPs) across subjects and conditions in response to targets and non-targets were significantly greater while on high dose of Omega-3 as compared to placebo. ERP memory components were more pronounced after 24- and 48-hrs of sleep deprivation compare to both baseline and placebo in the frontal (executive function) and the posterior (spatial working memory) regions.

Subjects reported no difference in anxiety, depression, anger, or confusion resulting from sleep deprivation during their Omega-3 and Placebo sessions. While on placebo, subjects reported increased fatigue and decreased vigor. Additional studies are underway to assess the benefit of Omega-3 under chronic sleep deprivation conditions.

Johnson, R, Berka, C. et al. (2010) Mitigation of sleep deprivation through Omega-3 fatty acids: Neurocognitive inflammatory, EEG and EKG evidence, Neuroscience, San Diego, CA