New Techniques for Assessment of Mental Performance in the Field
Harris R. Lieberman1and Bryan P. Coffey
This chapter considers some of the difficulties associated with conducting behavioral research in the field, as well as more general issues regarding assessment of human performance. Although the focus is on assessing the effect of nutritional manipulations on performance in the field, the general concepts should apply to a variety of other independent variables, such as environmental factors, sleep loss, and toxic agents. After this background material has been presented, a new ambulatory monitoring technology under development at the U.S. Army Research Institute of Environmental Medicine (USARIEM) will be introduced. Finally, some recommendations regarding future research in this area will be presented.
ASSESSMENT OF HUMAN BEHAVIOR
For a number of years, many investigators have been interested in how it might be possible to measure human cognitive behavior in a more continuous and less invasive manner than is possible with conventional tests of cognitive function. Standard tests of cognitive performance usually involve filling out forms with paper and pencil, taking written tests on a computer, or using mechanical devices such as Peg-Boards or other manipulanda (items that can be manipulated). These tests measure functions like learning, memory, reaction time, vigilance, attention, manual dexterity, and sensory function. Using such tests, a very wide range of human behaviors can be assessed and, at least in theory, related to real-world performance.
Although psychologists and other scientists conducting research in this area generally are required to rely on such tests to assess performance, there are a number of problems associated with employing them, not only in the field, but also in the laboratory. One limitation is simply the difficulties associated with administering tests to people engaged in field exercises. It is generally necessary for the subject being tested to stop all ongoing activities to participate in the task. When the subjects in a research study are members of military units engaged in training exercises or other military operations, this is a particularly significant problem and often excludes behavioral assessment in field studies. Studies, nutritional or otherwise, conducted in the field generally need to be minimally intrusive with regard to use of soldiers' time. When military units agree to participate in research, they are not primarily in the field to be test subjects. Usually they are there to train. When scientists conduct research with military units, commanders generally impose significant limitations on the amount of time their soldiers will be available to the investigators. Furthermore, soldiers themselves, especially those from nonelite units, can easily become bored or frustrated with behavioral tests, and the quality of the data collected may deteriorate. Although studies cannot be conducted without prior written informed consent, soldiers often become disaffected when asked to take behavioral tests repeatedly that they regard as less than critical to their essential military duties.
Another key issue is the interpretation of data collected using formal laboratory behavioral tests. It can be difficult to relate data from performance tests to real-world performance. This is a generic problem, not just an issue of relevance to field behavioral research. There are many reasons why it is difficult to relate cognitive test performance to actual work performance. One is that a subject usually cannot engage in actual work performance and simultaneously take a behavioral test.
Another well-known problem with conducting behavioral testing is circadian variation in performance (Moore-Ede et al., 1982). Because of such variation, testing must be conducted at multiple times during the day to fully describe the daily pattern of any behavioral parameter. Alternatively, if testing is
to take place once a day, each test session must be conducted at the same time. It is not appropriate to compare data collected at one time of day with data collected at another time. In field studies with military units as opposed to laboratory studies, testing at exactly the same time each day can be extremely impractical and often significantly increases the extent of disruption produced by the investigators. Furthermore, circadian variation in a specific parameter may overwhelm any differences caused by nutritional treatments or environmental conditions being evaluated. Also, individual behavioral functions appear to have different patterns of circadian rhythmicity (Moore-Ede et al., 1982).
Another critical problem is that the underlying behavioral function a performance test actually assesses is often unclear. Almost all tests, even the simplest, require the subject to use a variety of sensory systems, information processing capabilities, and motor functions. For example, visual choice reaction time requires the visual system to process sensory information, the attentional functions to focus on the critical sensory parameters, the memory functions to choose the correct stimulus, the decision-making processes to determine whether to initiate a response, and of course, the motor system to make even the simplest response. It cannot be assumed that the limiting or critical factor that any test assesses can always be defined accurately. Also, under different environmental conditions or because of the presence of some extraneous influence on performance, the critical factor in any behavioral test may change. For example, in a cold environment where substantial shivering occurs, impaired ability of the subject to make the necessary motor responses may prevent assessment of the cognitive parameter the test is actually intended to monitor.
One of the most difficult issues to address in any study of performance is determining which test or tests are optimal to answer the research question of interest. It is very difficult to determine what specific test or class of tests will be most useful to assess the effects of a particular experimental parameter. Selecting the test, particularly when only a few can be administered (as is almost always the case in field studies), presents a variety of problems to the investigator. Often, it is not known what test will be most sensitive to the effects of a nutrient, drug, or environmental condition on human performance. There is rarely agreement among scientists working in a particular area as to the best method to employ. In nutrition-behavior studies where, in most every case, the effects of nutritional manipulations are modest, this is a daunting problem. If the investigator selects the ''wrong" test, it will be insensitive to the treatment being evaluated. This could then result in a Type II error—failing to detect a treatment effect that actually is present. It may not even be possible to agree on the criteria that should be used to determine the optimal test to address a specific research question. For example, for many years, there was considerable controversy about whether moderate doses of caffeine, equivalent to those found in single servings of common foods, had effects on behavior (Dews, 1984). However,
when appropriate vigilance tests are employed, consistent effects of caffeine in this dose range are observed reliably (Fine et al., 1994; Lieberman, 1992).
A problem that often occurs in field studies is that subjects are inaccessible to the investigator for long periods of time. Military units frequently conduct operations in extreme and sometime dangerous environments. Often, it is not possible for investigators to accompany soldiers to administer tests. Frequently, it is these extreme environments that are of the greatest interest with regard to performance. One solution to this problem is to issue handheld computers to the soldiers participating in a study. The soldiers are then responsible for self-testing their own mental performance at designated times.
This technique, in combination with conventional tests of cognitive function, was employed in a month-long field study conducted by USARIEM and Natick Research, Development and Engineering Center (Askew et al., 1987). The study was conducted in northern Vermont in the autumn and was designed to evaluate an experimental, lightweight, 2,000 kcal/d ration (the RLW-30) in comparison with three Meals, Ready-to-Eat VI (MREs), which provide 4,000 kcal/d. Neither the conventional nor the self-administered tests of performance detected substantial changes in performance that were attributable to the RLW-30 ration, although there was some evidence of degradation of performance on some of the conventional tests among soldiers consuming the calorie-deficient ration (Askew et al., 1987; Lieberman et al., in press).
It is apparent that the problems associated with behavioral testing are substantial and magnified when testing must be conducted in the field. Several chapters in this volume provide partial solutions to some of the issues. Some require high technology and sophisticated equipment, such as simulators to collect data (see Watson and Papelis, Chapters 26 in this volume; Johnson, 1991) while others rely on simpler solutions such as paper and pencil tests (Mays, 1995; Shippee et al., 1994), or portable computers to assess performance (Banderet and Lieberman, 1989). In addition, there may be unique and innovative solutions that can be employed to provide data on soldiers in the field as they go about their daily activities without interrupting their training or other operations. Although advances in technology will provide partial solutions, no device or emerging technology will be the complete answer to these complex problems.
One way to gather useful behavioral data in the field is by using electronic activity monitors and other passive, electronic monitoring devices such as foot strike monitors (Askew et al., 1987; Hoyt et al., 1994; Lieberman et al., 1989; Redmond and Hegge, 1985). Although such devices currently do not assess mental performance, they can be employed effectively to provide information on soldiers as they go about their typical activities in the field. Activity monitors, which were developed in part with U.S. Army sponsorship, record
minute-by-minute patterns of the wearer's activity (Lieberman et al. 1989; Redmond and Hegge, 1985; Webster et al., 1982; for a comprehensive review see Tryon, 1991). The most advanced versions can record continuously for many days and are suitable for use in the field, even in extreme environmental conditions (Shippee et al., 1994). Attempts have been made to relate the levels of activity recorded on these monitors to energy expenditure. Although the relationship between activity and energy expenditure is complex, activity monitoring can improve estimates of energy expenditure provided by other methods (Hoyt et al., 1991; Patterson et al., 1993).
One sophisticated, commercially available activity monitor is manufactured by Precision Control Devices, Fort Walton Beach, Florida. The device, the Motionlogger Actigraph, model AMA-32, has been employed effectively to assess patterns of rest and activity and estimate duration and fragmentation of sleep. It has been used successfully as a supplemental measure of energy expenditure in conjunction with other techniques in military field studies (Hoyt et al., 1991). In addition, it can provide invaluable information on levels of activity and sleep patterns of volunteers. The devices are 4 × 3.1 × 1 cm, weigh 57 g, and typically are worn on the wrist of the nonpreferred hand using a standard wristwatch band. Each device contains a microcomputer, 32 kilobytes of memory, an analog-to-digital converter, and a piezoelectric motion sensor and is powered by a standard wristwatch battery. Data collected by the AMA-32 can be downloaded to a laptop or other IBM-compatible computer for further analysis using a specially developed computer program (ACTION 3, Ambulatory Monitoring, Inc., Ardsley, N.Y.) or other software.
The ability of actigraphs to predict sleeping versus waking state of humans has been demonstrated by several investigators. Algorithms to classify activity patterns of individuals wearing activity monitors as representing a sleeping or waking state have been developed and validated (Cole and Kripke, 1988; Webster et al., 1982; Sadeh et al., 1989). Actigraphs are now widely employed to supplement the more accurate and detailed information provided on the extent and structure of sleep by polysomnography. Activity monitors also can be employed to assess circadian rhythms of rest and activity in normal individuals and those with disturbed rhythms because of psychiatric or other medical conditions or transmeridianal travel (Comperatore et al., 1996; Lieberman et al., 1989; Satlin et al., 1991; Teicher et al., 1986; 1988). In a study conducted at USARIEM with the Ambulatory Monitoring Inc.'s AMA-32 actigraphs, the effects on sleep quantity and quality of sleeping in a chemical protective mask were assessed. The significant degradation in sleep that the monitors documented during sleep in the mask was consistent with an observation of impaired performance the next day when soldiers were tested on computer-based performance tasks (Lieberman et al., 1994, 1996).
Figure 25-1 presents data collected with an AMA-32 activity monitor in a USARIEM nutrition field study conducted in a hot desert environment (Hotson et al., 1995). The study was designed to assess a new operational ration (the
Unified Group Ration), as well as the effects of a supplemental carbohydrate beverage on a Marine heavy artillery unit that was conducting a live-fire training exercise. As shown in Figure 25-1, which was generated by the ACTION III computer program, daily patterns of rest and activity are clearly documented over the 9 days of the study. In spite of temperatures as high as 130°F (54.5°C), subjects maintained a high level of activity on most days. Below the plots of the rest-activity pattern of each day are corresponding plots indicating the derived sleeping versus waking state of the subject at any given moment in time. The estimates of sleeping versus waking state are generated automatically by the Action III program using a validated algorithm (Figure 25-1).
One clear advantage of activity monitors over traditional cognitive tests is that they provide continuous assessment of a form of behavior, i.e., physical
activity. Although activity monitors cannot assess performance per se, the data they collect can be related to both physical and mental state.
THE USARIEM VIGILANCE MONITOR
Based in part on the demonstrated utility of activity monitors in the laboratory, clinic, and field, USARIEM has attempted to develop new ambulatory monitoring devices that will provide additional capabilities to researchers interested in assessing behavior and performance in the field. The devices may be particularly suitable for studying soldiers as they conduct field exercises and even actual operations. The monitors combine the characteristics and capabilities of actigraphs such as the AMA-32 with the addition of performance assessment and intervention capabilities. They may be especially useful for addressing how certain nutritional and environmental variables influence mental performance and evaluating the relationship between work performance and mental state. One of the chief capabilities of these devices is the ability to assess vigilance.
Vigilance is a behavioral function that can be readily assessed and may be an important aspect of some critical areas of military performance. Obviously, there are a wide range of behavioral parameters that are critical for soldiers as they conduct their missions. Reaction time, motor skills, attention, memory, and higher-order cognitive processing are all important if soldiers are to accomplish their objectives effectively and safely. One particularly important cognitive function required for many key military duties is maintaining vigilance. It is critical that soldiers maintain vigilance under the worst circumstances, such as when they are sleep deprived; in the middle of the night or early morning when they are at the nadir of the circadian performance rhythm; and when they are physically and environmentally stressed. If a key sentry, operator of surveillance equipment like radar, truck driver, or helicopter pilot cannot maintain vigilance, the consequences can be catastrophic. Assessment of vigilance is of great practical significance since many key civilian as well as military occupations, such as operation of motor vehicles and industrial equipment, require sustained maintenance of vigilance for long periods of time (Mackie, 1987). A number of accidents, such as the Three Mile Island nuclear reactor failure and commercial aircraft crashes, have been attributed, at least in part, to the failure of human operators to detect critical stimuli (Mitler, 1988; Office of Technology Assessment, 1991). Furthermore, even at the optimal time of day, vigilance deteriorates in well-rested individuals if it must be sustained for long periods time (Koelega, 1989).
Vigilance has been reported to be a sensitive measure of the functional capability of an organism. It reflects the ability of individuals to process relevant information and respond in a timely fashion. The central mechanisms underlying vigilance performance have been investigated, and a wide variety of tasks have been used to assess the underlying process or processes (Fine et al.,
1994; Hirshkowitz et al., 1993; Koelega, 1989; Tiplady, 1992). It also has been suggested that vigilance tasks may be more relevant to the performance of everyday activities than shorter tests of cognitive performance often employed in test batteries (Koelega, 1989).
It is important to note that vigilance performance is sensitive to effects of many factors, including sleep loss, drugs, hormones, food constituents, and environmental variables (Clubley et al., 1979; Dollins et al., 1993; Koelega, 1989; Lieberman, 1992; Wilkinson, 1968). Furthermore, there is a large body of literature in this area, which makes it possible to begin to relate vigilance to real-world performance. A good example is the work on the influence of caffeine on simulated sentry duty, presented by Johnson (1991).
Given the operational importance of vigilance and its sensitivity to nutritional and other variables, a device was developed in this laboratory that is capable of continually assessing this parameter as soldiers or civilians go about their normal daily activities. The device also can measure patterns of rest and activity simultaneously and can measure several key environmental factors. As currently configured, the device contains the equivalent of a first-generation personal computer, including 128 kilobytes of memory and an 8-bit microprocessor. It also contains substantial signal processing capabilities and simple auditory and visual output capabilities. It will record continuously for more than 5 days.
The device provides a variety of information about the environment of the wearer, as well as information about the wearer's activity and performance. It can be worn on the wrist, although it is somewhat larger than state-of-the-art activity monitors, such as the AMA-32.
As discussed above, assessment of motor activity has proven to be a valuable technique for gathering information about sleep, circadian rhythms, and energy expenditure of the wearer. The vigilance monitor developed at USARIEM can simultaneously collect multiple channels of activity data. The channels can be programmed to differ in their recording characteristics. One channel might be selected to be sensitive to low-amplitude activity, while another might be sensitive to moderate-or high-amplitude accelerations. The channels also can vary with regard to their temporal characteristics so that differences in the duration of motion can be assessed.
Figure 25-2 displays data collected in the field using the USARIEM vigilance monitor (Unpublished observations, H. R. Lieberman, USARIEM, Natick, Mass., 1994). Data from three channels of activity in counts per minute are presented in the first three panels (Acc1–Acc3). Channels 1 and 2 (Acc1 and Acc2) vary with regard to amplitude of the acceleration they will sense, while Channel 3 (Acc3) has a longer time constant, so it will be more sensitive to longer-duration accelerations. It is believed that the capability to record a variety of different channels of activity simultaneously will, among other things,
permit the development of improved algorithms to predict sleeping versus waking state, as well as improve the ability of monitors to predict energy expenditure.
There are, of course, a multitude of environmental factors regulating human behavior. Many cannot readily be assessed, but several can be acquired automatically with minimal difficulty because of the availability of miniature low-power sensors. Therefore, several environmental sensors have been included in the device developed at USARIEM. Specifically, the device can continuously assess and record ambient illumination, sound, and temperature levels. Although the importance of the presence or absence of light in the control of human behavior is readily apparent, it was not until a few years ago
that the critical nature of bright light in the regulation of human circadian rhythms became known (Lewy et al., 1980). A light sensor has been included in the device so that information regarding the duration and amplitude of individual exposure to light can be continuously acquired (Channel 5 [Light], Figure 25-2).
Another factor influencing human rest and activity is the pattern and level of sound in the environment. Using a miniature microphone, this information also is collected continuously by the monitor (Channel 7 [Sound], Figure 25-2).
Also included is a thermistor that measures ambient temperature on the surface of the monitor (Channel 6 [Temp], Figure 25-2). The importance of ambient temperature in regulating behavior cannot be understated. When field studies are conducted, information on ambient temperature conditions to which soldiers are exposed can be difficult to collect. Furthermore, even when it is available, it does not describe the conditions an individual soldier experiences but rather the conditions at a stationary weather station (Santee and Hoyt, 1994). By individually monitoring temperature on each subject participating in a study, more accurate tracking of environmental exposure may be possible.
As configured for the study presented here (Unpublished observations, H. R. Lieberman, USARIEM, Natick, Mass., 1994), the monitor also assessed the vigilance of the wearer (Channel 4 [Rtime], Figure 25-2). On average once every 15 minutes, the monitor's speaker presented a short-duration tone. The subject was asked to respond by pushing a small switch on the monitor as soon as he or she heard the tone. If the subject did not respond to the initial tone in a few seconds, a second louder tone was presented. Finally, a third tone was presented if the subject still failed to respond. In addition to determining whether or not the subject responded to the tones, the monitor also recorded the time required for the subject to respond—his or her reaction time. Therefore, the monitor can continually assess alertness in the field as soldiers or civilians perform their daily duties. Whenever the individual wishes to turn off the behavioral task or tasks, for example to sleep, he or she can do so.
It also should be noted that because the monitor contains a user-programmable microcomputer, it can be used not only to assess vigilance but to actively intervene, if desired, to prevent degradation in performance of the wearer.
Like most other activity monitors currently available, the monitor developed at USARIEM, once programmed, will record data for many days with no investigator input. The computer program executed by the monitor is written on a standard IBM-compatible computer and is downloaded to the monitor for execution. When the experiment is complete or the memory is full, the collected data are uploaded to a standard IBM-compatible personal computer for analysis.
Some of the capabilities of the USARIEM vigilance monitor are:
The wearer can continue with most daily activities while wearing the monitor. Of course, it will interfere somewhat with ongoing duties, but
compared with stopping to take a behavioral test, responding to the monitor is a minor inconvenience.
The monitor is suitable for field use and provides information on several environmental factors: light, sound, and ambient temperature.
With sufficient monitors, it is possible to monitor a large number of subjects for many days in the absence of an investigator.
The monitor is fully programmable so that it can be employed in a variety of configurations to address different experimental questions.
The monitor may be used for the active prevention of reductions in alertness, if validated for this use.
There is a distinct need for innovative technologies that can be employed to assess human performance and other aspects of behavior in the field. Such technologies should assess behaviors of clear relevance to the critical duties of soldiers and other warfighters.
Although there are currently some mature technologies, such as activity monitors, for assessment of limited aspects of behavior in the field, the development of devices with much greater functionality would significantly advance military nutrition field research.
Personal monitoring devices are relatively inexpensive technologies to utilize in the field once the appropriate hardware and software capabilities are developed and implemented.
There is considerable need for the development of analytic tools for processing and integrating the information collected by ambulatory monitors.
It is recommended that laboratory or other controlled studies be conducted to relate data provided by new technologies for field assessment of performance to the information provided by standard techniques.
Askew, E.W., I. Munro, M.A. Sharp, S. Siegel, R. Popper, M.S. Rose, R.W. Hoyt, K. Reynolds, H.R. Lieberman, D. Engell, and C.P. Shaw 1987 Nutritional status and physical and mental performance of soldiers consuming the Ration, Lightweight or the Meal, Ready-to-Eat military field ration during a 30-day field training exercise (RLW-30). Technical Report T7-87. Natick, Mass.: U.S. Army Research Institute of Environmental Medicine.
Banderet L.E., and H.R. Lieberman 1989 Treatment with tyrosine, a neurotransmitter precursor, reduces environmental stress in humans. Brain Res. Bull. 22:759–762.
Clubley, M., C.E. Bye, T.A. Henson, A.W. Peck, and C.J. Riddington 1979 Effects of caffeine and cyclizine alone and in combination on human performance, subjective effects and EEG activity. Br. J. Clin. Pharmacol. 7:157–163.
Cole, R.J., and D.F. Kripke 1988 Progress in automatic sleep/wake scoring by wrist actigraph. Sleep Res. 17:331.
Comperatore, C.A., H.R. Lieberman, A.W. Kirby, B. Adams, and J.S. Crowley 1996 Melatonin efficacy in aviation missions requiring rapid deployment and night operations. Aviat. Space Environ. Med. 67(6):520–524.
Dews, P.B., ed. 1984 Behavioral effects of caffeine. Pp. 86–100 in Caffeine. New York: Springer-Verlag.
Dollins, A.B., H.J. Lynch, M.H. Deng, K.U. Kischka, R.E. Gleason, R.J. Wurtman, and H.R. Lieberman 1993 Effect of pharmacological daytime doses of melatonin on human mood and performance. Psychopharmacology 112:490–496.
Fine, B.J., J.L. Kobrick, H.R. Lieberman, B. Marlowe, R.H. Riley, and W.J. Tharion 1994 Effects of caffeine or diphenhydramine on visual vigilance. Psychopharmacology 114:233–238.
Hirshkowitz, M., L. De La Cueva, and J.H. Herman 1993 The multiple vigilance test. Behavior Research Methods, Instruments, and Computers 25(2):272–275.
Hotson, N.G., A.D. Cline, R.L. Shippee, S. McGraw, M. Kramer, W.J. Tharion, J.P. DeLany, R. Hoyt, C.J. Baker-Fulco, and H.R. Lieberman 1995 Nutritional challenges for field feeding in a desert environment. Unpublished manuscript. Natick, Mass.: U.S. Army Research Institute of Environmental Medicine.
Hoyt, R.W., T.P. Jones, T.P. Stein, G.W. McAninch, H.R. Lieberman, E.W. Askew, and A. Cymerman 1991 Doubly labeled water measurement of human energy expenditure during strenuous exercise. J. Appl. Physiol. 71:16–22.
Hoyt, R.W., J.J. Knapik, J.F. Lanza, B.H. Jones, and J.S. Staab 1994 Ambulatory foot contact monitor to estimate the metabolic cost of human locomotion. J. Appl. Physiol. 76(4):1818–1822.
Johnson, R.F. 1991 Rifle firing simulation: Effects of MOPP, heat, and medications on marksmanship. Pp. 530–535 in Proceedings of the Military Testing Association. San Antonio, Tx.: Armstrong Laboratory Resource Directorate and U.S. Air Force Occupational Measurement Squadron.
Koelega, H.S. 1989 Benzodiazepines and vigilance performance: A review. Psychopharmacology 98:146–156.
Lewy, A.J., T.A. Wehr, F.K. Goodwin, D.A. Newsome, and S.P. Markey 1980 Light suppresses melatonin secretion in humans. Science 210:1267–1269.
Lieberman, H.R. 1989 Cognitive effects of various food constituents. Pp. 267–281 in Psychobiology of Human Eating and Nutritional Behavior, R. Shephard, ed. Chichester, England: Wiley.
1992 Caffeine. Pp. 49–72 in Factors Affecting Human Performance. Volume II: The Physical Environment, D. Jones and A. Smith, eds. London: Academic Press.
Lieberman, H.R., J.J. Wurtman, and M.H. Teicher 1989 Circadian rhythms of activity in healthy young and elderly humans. Neurobiol. Aging 10:259–265.
Lieberman H.R., M.Z. Mays, B. Shukitt-Hale, B.E. Marlowe, D.I. Welch, W.J. Tharion, P.D. Gardner, A. Pappas, and K.S.K. Chinn 1994 Sleep in the M40 mask: Sleep quality, mask fit factor and next day performance. Technical Report T94-8. Natick, Mass.: U.S. Army Research Institute of Environmental Medicine.
Lieberman H.R., M.Z. Mays, B. Shukitt-Hale, K.S.K. Chinn, and W. J. Tharion 1996 Effects of sleeping in a chemical protective mask on sleep quality and performance. Aviat. Space Environ. Med. 67(9):841–848.
Lieberman, H.R., E.W. Askew, and M.A. Sharp 1997 Effects of 30 days of undernutrition on plasma neurotransmitters and other amino acids. J. Nutr. Biochem. 8(3):119–126.
Mackie, R.R. 1987 Vigilance research—are we ready for countermeasures? Hum. Factors 29(60):707–723.
Mays, M.Z. 1995 Impact of underconsumption on cognitive performance. Pp. 285–302 in Not Eating Enough, Overcoming Underconsumption of Military Operational Rations, B.M. Marriott, ed. A report of the Committee on Military Nutrition Research, Food and Nutrition Board, Institute of Medicine. Washington, D.C.: National Academy Press.
Mitler, M.M. 1988 Catastrophes, sleep and public policy: Consensus report. Sleep 11(1):100–109.
Moore-Ede, M.C., F.M. Sulzman, and C.A. Fuller 1982 The Clocks That Time Us. Physiology of the Circadian Timing System. Cambridge, Mass.: Harvard University Press.
Office of Technology Assessment, U.S. Congress 1991 Biological rhythms: Implications for the worker. OTA-BA-463. Washington, D.C.:U.S. Government Printing Office.
Patterson, S.M., D.S. Krantz, L.C. Montgomery, P.A. Deuster, S. M. Hedges, and L.E. Nebel 1993 Automated physical activity monitoring: Validation and comparison with physiological and self-reported measures. Psychophysiology 30(3):296–305.
Redmond, D.P., and F.W. Hegge 1985 Observations on the design and specification of a wrist-worn human activity monitoring system. Behavior Research Methods, Instruments, and Computers 17(6):659–669.
Sadeh, A., J. Alster, D. Urbach, and P. Lavie 1989 Actigraphically based automatic bedtime sleep-wake scoring: Validity and clinical applications. J. Ambul. Monitor 2:209–216.
Santee, W.R., and R.W. Hoyt 1994 Recommendations for meteorological data collection during physiological field studies. Technical Report T94-9. Natick, Mass.: U.S. Army Research Institute of Environmental Medicine.
Satlin, A., M.H. Teicher, H.R. Lieberman, R.J. Baldessarini, L. Volicer, and Y. Rheaume 1991 Circadian locomotor activity rhythms in Alzheimer's disease. Neuropsychopharmacology 5:115–126.
Shippee, R., K. Friedl, T. Kramer, M. Mays, K. Popp, E.W. Askew, B. Fairbrother, R. Hoyt, J. Vogel, L. Marchitelli, P. Frykman, L. Martinez-Lopez, E. Bernton, M. Kramer, R. Tulley, J. Rood, J. Delaney, D. Jezior, J. Arsenault, B. Nindl, R. Galloway, and D. Hoover 1994 Nutritional and immunological assessment of Ranger students with increased caloric intake. Technical report T95-5. Natick, Mass: U.S. Army Research Institute of Environmental Medicine.
Teicher, M.H., J.M. Lawrence, N.I. Barber, S. Finkelstein, H.R. Lieberman, and R.J. Baldessarini 1986 Altered locomotor activity in neuropsychiatric patients. Prog. Neuropsychopharmacol. Biol. Psychiatry 10:755–761.
Teicher, M.H., J.M. Lawrence, N.I. Barber, S.P. Finklestein, H.R. Lieberman, and R.J. Baldessarini 1988 Altered circadian activity rhythms in geriatric patients with major depression. Arch. Gen. Psychiatry 45:913–917.
Tiplady, B. 1992 Continuous attention: Rationale and discriminant variation of a test designed for use in psychopharmacology. Behavior Research Methods, Instruments, and Computers 24(1):16–21.
Tryon, W.W. 1991 Activity Measurement in Psychology and Medicine. New York: Plenum Press.
Webster, J.B., D.F. Kripke, S. Messin, D.J. Mullaney, and G. Wyborney 1982 An activity-based sleep monitor system for ambulatory use. Sleep 5:389–399.
Wilkinson, R.T. 1968 Sleep deprivation: Performance tests for partial and selective sleep deprivation. Pp. 28–43 in Progress in Clinical Psychology, B.F. Reiss and L.A. Abt, eds. New York: Grune and Stratton.
JOHANNA DWYER: Knowing your wife, who is a pediatrician, I told Bill [Beisel] that you made a computerized baby that was always crying. My question is, how do you maintain the motivation to turn this thing [the vigilance monitor] off?
HARRIS LIEBERMAN: In the preliminary study that we conducted, we saw relatively stable levels of responding in the volunteers. When we test hundreds of subjects, how consistent they would be in responding to frequent stimuli is something we will just have to answer by doing the research. It is a good question.
ROBIN KANAREK: Along similar lines, it seems to me that it would depend a lot on the environment that the soldier was in at that particular time, for example if he is being shelled. Have you looked at how that would interact?
HARRIS LIEBERMAN: That is one of the nice things [about the device]. We can actually do minute-by-minute correlations, look at ambient sound levels and ambient light levels, and see if performance is affected by those particular parameters and also by some variations in temperature.
The configuration that I am showing you is fairly flexible. It would be possible to add other environmental sensors if you had a particular issue that you were concerned about, for example, exposure to a toxic gas. So you also
could have that capability fairly easily, although the device would not be as field hardened with special modifications on it.
Those are experimental questions that I hope we will be able to start to address with this kind of technology. It is just the beginning.
DAVID DINGES: What you have presented is a fascinating device, and, I think a logical step to move toward, Harris. What is your best guess right now for where this will be most useful? In other words, in the back of your mind, is it that this is going to be a fall-asleep detection device? Are you going to be able to identify that the soldiers were not paying attention at the right time? What is this device going to be for?
HARRIS LIEBERMAN: In the first instance, it will be used for us to gather data for experiments in which we are testing the effects of environmental and nutritional parameters. That is really what it was built for in the first place. The capability for intervening and actually keeping people alert is a little bit farfetched, perhaps. The only way to find out whether it is going to be useful is to get out there and try it out. I think finding the right parameters will be extremely difficult.
BERNADETTE MARRIOTT: What about another scenario where people may be concentrating on the task at hand to the point where they say, ''I don't want to be bothered with this" because it is an interruption rather than a measure of lapses?
HARRIS LIEBERMAN: The device can be turned off; that is, the auditory portion of the device can simply be turned off by the subject for any length of time simply by punching it in. Yes, that is a very good question. If somebody is trying to sleep, we do not want to interrupt their sleep, and if someone is trying to do a task that requires great concentration, we do not want to be in a situation of interfering with that.
JOHN VANDERVEEN: Do you have any telemetry [monitoring] capability with this?
HARRIS LIEBERMAN: No, there is not any telemetry capability. We have sufficient memory to record all these channels of data for more than 5 days, actually. The battery, not the amount of information we can store, is the limiting factor.
For the applications that we are interested in out in the field, it is really better, I think, to collect the data this way than to do telemetry. You would have to go to a whole new device, I think, if you wanted to do telemetry with this.
PATRICK DUNNE: I was wondering if you could correlate when the subjects were eating, with the activity monitor because a real issue when you are out in the field is why the people do not take in a full caloric load if they are just not given any down time to eat it.
HARRIS LIEBERMAN: I do not think it would really tell us.
PATRICK DUNNE: There seemed to be a few low spots within a given day, and I was just wondering whether the soldiers might have gone under a shade and if that may have been when they were eating.
HARRIS LIEBERMAN: We could set up a task and simply ask them to respond by pushing a button whenever they ate. That would be one way to do it. But just looking at the data we have now, we could not tell.