Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
SUGARY: FINDINGS AND ~CO~NDATIONS em?
SUMMARY: FINDINGS A2iD RECOMMENDATIONS Night vision encompasses many different visual functions uncter a variety of ambient lighting conditions. Since night operations are a crucial part of around-the-clock combat readiness, the United States Air Force has been interested in evaluating visual performance at night. The Working Group on Night Vision addressed the topic of night vision with four specific objectives in mind: (1) a definition of the relevant parameters of night vision; (2 ~ an update of the literature pertaining to night vision, especially new f indings, test procedures, and concepts since 1950; (3 ~ the development of guidelines for establishing a compre- hens~ve night vision laboratory; and (4) recommendations for the devel- opment of night vision screening tests. The first two objectives were addressed by convening a symposium at Brooks Air Force Base in 1985 (see Appendix B. ~ The proceedings of the symposium form the basis of this report. The findings and recommenda- tions presented here are based on working group discussions following that meeting. Reference is made throughout this section to the papers in this volume that have some relevance to the recommendations under · ~ discussion. The recommendations address five broad topics: (1) the specifica- tion of ambient illuminance levels; (2) task analysis and characteri- zation of the work environment; (3) research areas of potential utility to the development of night vision tests; (4) the development of night vision screening tests; and (5) recommended equipment for a night vision laboratory. SPECIFICATION OF AMBIENT ILLU}lINANCE LEVELS In order to predict performance under low-illumination conditions, it is necessary to match information describing ambient illuminance levels, task requirements, and psychophysical data relating the capa- bilities and limitations of human performance to luminance levels. If the task in question is to be carried out indoors, specification of illuminance level is ~ .~1 at~vely simple procedure. Indeed, in most artificially illuminated environments, it i s possible to ad just the i Luminance levels to match the demands of the task. 3
4 However, for tasks performed under natural lighting, specification of illuminance is complicated by the continually varying levels of natural illuminance from the sun and moon and their attenuation by atmospheric conditions. Perhaps the most complex illumination condi- tions are those associated with mixtures of natural and artificial lighting, since there can be large variations in both the spatial and temporal characteristics of ambient illumination. Estimates of the ambient illumination during twilight and from the moon can be calculated on the basis of the date, time, latitude, and altitude. Less precise estimates can be made of the effects of atmos- pheric conditions. We suggest that consideration be given to the devel- opment of a hand calculator that would permit the rapid determination of ambient illuminance levels between sunset and sunrise. These data are particularly critical during twilight, when illum~nance is changing rapidly (by a factor of 2 every 4-5 min at 40 degrees north latitude), and during the darker portions of twilight and at night, when even the low illuminance from moonlight can have implications for performance. TASK ANALYST S AND CHARACTERI ZATION OF THE WORE: ENVIRONMENT The design of visual screening tests for night vision must incor- porate features that are based on the.nature of the tasks performed, the properties of the work environment, and the typical illumination and contrast levels that are present for night vision working opera- tions. In a similar fashion, the definition of operational light levels for various tasks requires an accurate assessment of the visual skills used to perform certain tasks and the physical characteristics of the work environment. Although this information is currently avail- able for some tasks performed under low illumination, specif ications are not yet available for the ma jor ity of night vision working opera- tions. In order to establish vision test procedures with h igh job rele- vance and validity, it is essential that detailed information be made available for task requirements and the work environment. We recom- mend that this information be obtained as the first step in the analy- sis of night vision. A quantitative description of the work environ- ment should include the measurement of ambient lighting conditions, contrast levels, the spatial and temporal frequency properties of the environment, and the range of stimulus conditions under which certain tasks are performed. Task analyses of various jobs should include a comprehensive survey of the most frequent tasks performed, the most critical task components, the types of visual skills necessary for properly conducting the task, an evaluation of the frequency and con- sequences of errors in task performance, and other related factors. It is important to define the illumination levels under which tasks are performed and whether they require visual detection, identification, localization, or other skills. The visual requirements for individual jobs performed at low Dominances (and the time, expense, and difficulty of administering night vision tests) can vary widely, depending on the specific characteristics of the task.
5 We recommend that these evaluations be carried out by a team of in- vestigators whose expertise includes visual psychophysics for measure- ment of light and contrast levels, analysis of spatial and temporal fre- quency properties of the work environment, and determination of visual functions critical to task performance; industrial/organizational psy- chology for conducting task analyses; human factors engineering for ergonomic considerations of the work environment; and knowledge of the operational concerns of the armed forces. The results of this analysis will have a significant impact on defining operational light levels for night vision tasks, defining the types of vision screening tests that should be developed, and determining the priorities for a night vision research laboratory. RESEARCH AREAS OF POTENTIAL UTILITY TO NIGHT VISION As indicated in the historical review (Appendix A), the majority of night vision tests evaluated during and following World War II were based on either scotopic detection thresholds or acuity-form perception tasks performed at mesopic or scotopic luminance levels. Since that time, the visual sciences community has produced many research findings and methodologies that are relevant to night vision. We recommend that the areas of research described below be considered for evaluation by a comprehensive night vision laboratory facility. .., Validation Studies Regardless of the specific visual functions and test procedures that are examined by a night vision laboratory, we believe that it is critical to correlate vision test results with task performance measures f ram studies with simulators, "field" studies, performance ratings of instructors, or other information sources. This information will be useful in determining which test procedures are the best pre- dictors of task performance and will thereby establish a basis for designing appropriate night vision screening tests. One of the goals of a military night vision laboratory should be to develop a test or battery of tests that are able to predict individual performance for various night vision tasks. As described In the his- torical review (Appendix A), previous attempts to predict night vision task performance on the basis of detection and acuity measures were not very successful. Recent investigations of the contrast sensitivity function, peripheral vision, accommodation and convergence, and other visual functions suggest that they may have predictive value for task performance. As the papers in this volume illustrate: (1) the accu- racy of accommodation and convergence responses is reported to be related to the visibility of objects during night driving (Owens); (2) the status of the peripheral visual field is reported to be corre- lated with driving performance (Johnson); (3) recognition of aircraft appears to be related to an individuals contrast sensitivity function (Harvey; Haber); and (4) the contrast sensitivity function and the
6 peripheral visual field are reported to have predictive value for orientation and mobility skills of patients with low vision (Bailey). None of these visual functions has been correlated with task perfor- mance under low-luminance conditions. In addition to evaluating the basic visual functions described in subsequent sections of this report, it is also important to correlate these test results with performance of visual tasks under low-luminance conditions. Emphasis should be placed on those tasks that are conducted frequently and those that are critical (i.e., nonoptimal performance can have serious consequences). There are a variety of procedures that may be used for deriving evaluations of task performance, including (1) the use of simulators of field studies, in which a task can be conducted under controlled conditions that permit measures of performance accu- racy and/or efficiency; (2) subjective performance ratings by instruc- tors or supervisors; and (3) critical incident evaluations, in which performance errors are reviewed. Performance evaluations can then be compared with psychophysical measurements to identify tests that may be good predictors of specific night vision tasks. Longitudinal Studies of Night Vision Performance We also believe it is important to establish a data base to track individuals over a period of time and to establish a population sample of reasonable size. Not only will this be helpful for present studies but it will also be invaluable for examining issues that may arise in the future. Many of today's questions might be answered simply if such a data base were available from previous research. We recommend that short-term (1-2 years) and long-term (7-20 years) studies be performed for military personnel engaged in night vision tasks. Both psychophysical tests under low-luminance conditions (e.g., contrast sensitivity, glare disability, peripheral visual function, dynamic visual acuity, oculomotor function) and task performance mea- sures should be obtained at periodic intervals. For both the short-term and long-term studies, individual differences and their significance for job-related night vision tasks should be evaluated. Short-term studies (with sampling intervals of 1-4 weeks) will help identify the variabil- ity of psychophysical and performance measures (both within and between subjects) and will also provide an opport'un~ty to examine the effects of practice and training. Long-term studies (with 1-year intervals of time) will help to identify trends in night vision performance varia- tions that may be related to aging, job experience, or environmental influences. Nearly all studies of long-term visual changes have been cross-sectional because of the logistic problems associated with lon- g~tudinal follow-up of individuals in a mobile society. The military environment is unique in that it is possible to conduct follow-up testing of the same individuals over extended per iods of time. By mon- itoring psychophysical and performance measures over many years, the effects of aging, job experience, "nd environmental influences can be readily examined. For example, changes in tonic accommodation and con- vergence (or, alternatively, the resting or dark levels of accommodation
7 and convergence) brought about by aging, greater amounts of course work, or other influences may have significant long-term consequences for cr itica1 night vision tasks. The data from long-term evaluations will also be helpful in def ining norms for visual function and task performance under low illumination conditions. Oculomotor Effects and Spatial Orientation Recent research has demonstrated that accor~odation and convergence responses become inaccurate under degraded viewing conditions and dis- sociate to different tonic or ~resting" levels in darkness (Owens). Although the resting positions of accommodation and convergence are located at intermediate distances (0.5-2 m) for most persons, large individual differences are present. These individual differences in oculomotor adjustment under low illumination may be an important compo- nent of the large individual variability that is typically encountered for night vision tasks such as target detection, recognition, and local- ization. Distance perception can also be influenced by accommodation and convergence responses. Tonic shifts in the oculomotor adjustments, resulting from prolonged viewing of targets at fixed distances, can also influence accommodation and convergence responses as well as pro- duce shifts in distance perception (see, for example, Post and Beckman; Ebenholtz). Reflexive and voluntary eye movements and their relation- ship to se'f-motion sensation (vection), induced motion of objects in the environment, spatial orientation, and postural stability are also important topics for night vision research. Many of the night vision tasks performed by Air Force personnel are conducted under "reduced cue" situations, in which vection (illusory self-motion), induced motion, spatial or ientation, and postural stability can signif icantly Papa fir pert ormance . The Spatial Contrast Sensitivity Function The spatial contrast sensitivity function has received increasing attention in the past 15 years ~ see, for example, Harvey) . I t has been shown to be a sensitive indicator of early clinical visual abnormali- ties, often in the absence of any deficits noted with standard clinical tests of visual function. In addition, recent studies have demonstrated that the contrast sensitivity function is a very good predictor of per- formance for certain visual detection and identification tasks, visual similarity judgments, letter identification and confusion, shape recog- nition, and related visual tasks. Many studies indicate that it is an important determinant of indi- vidual differences in visual performance for populations with "normal" vision. Although mesopic and scotopic contrast sensitivity functions have been measured for various locations in the visual field, little is presently known about the relationship between contrast sensitivity and visual performance under low-luminance conditions. The relationships between contrast sensitivity and visual performance that have been
8 reported for photopic luminance levels indicate that this area of research should be a high priority for investigation by a night vision laboratory. Peripheral Vision Function The peripheral visual field has been shown to be an important factor in tasks pertaining to orientation and mobility, vehicle guidance, and other related functions (Johnson; Bedell; Bailey). At low luminances the periphery becomes even more important, since performance with res- pect to many visual functions (detection, contrast sensitivity, resolu- tion, etc.) is better in peripheral vision than for the fovea under these conditions. There are several aspects of peripheral vision research that should be evaluated by a night vision laboratory. First, optimum performance of visual functions under low-luminance conditions occurs at different visual field eccentricities, depending on the specific task (detection, identification, spatial localization, etc.~. Observers are able to select and maintain a specific nonfoveal locus for performing tasks under free-viewing conditions at low lumi- nances, although it is not clear how this process is accomplished, or what factors are involved. A second area of peripheral vision research should be directed to the effects of practice and training. It is generally believed that the peripheral visual field is capable of significant improvement in visual performance with practice and training. However, the duration of these improvements, their generalization to nonlaboratory environ- ments> and their significance for task performance are not well under- stood. Third, there are many differences between central and peripheral processing, including reduced velocity and/or disappearance of slowly moving objects; changes in apparent brightness, size, and distance of peripherally viewed targets; and the Troxler effect (fading of station- ary peripheral stimuli in the absence of eye movements). The bases for these differences and their relation to visual task performance are not well understood. Finally, there is a paucity of information available on flicker and motion sensitivity in the peripheral visual field at low luminances, despite the fact that the perhiphery is particularly sensi- tive to interrupted or moving stimuli. Visual Search/Vigilance Night vision often forces the observer to perform under conditions in which imperfect visual information is available, and mental set and search strategies would necessarily play an unusually signficant role. Most laboratory studies of visual function involve the use of a static display, maintenance of steady fixation by the observer, and a high degree of certainty about the stimulus type, location, and time of pre- sentation. However, visual performance tasks are typically conducted in dynamic situations, with free viewing and a reasonable amount of stimulus uncertainty.
9 Target detection, identification, and localization are tasks that are highly dependent on the properties of underlying neural mechanisms, visual search behavior, vigilance, search strategies, mental set, and other factors (see, for example, Copenhagen and Reuter; MacLeod and Stockman; Makous; Leibowitz, Sheehy, and Gish; Sanders). Additional research on these problems is needed, particularly with respect to night vision tasks. Prolonged and/or Stressful Conditions Prolonged effort and/or stress, both of which are common in the military environment, can produce significant alterations in the oculo- motor system, functional visual fields, gaze stability mechanisms, and cognitive processing functions (Leibowitz, Sheehy, and Gish). Very few of the currently available tests of visual function are designed, how- ever, to evaluate performance under those conditions. The majority of current evaluation techniques and procedures were designed to reflect performance for short-term tasks in a relatively unstressful environ- ment. Many of the changes involved are critical for tasks performed under low illuminance when the demands of the task frequently encroach on human performance limitations. One of the effects, for example, is visual field narrowing. Under a wide variety of psychological and physiological stress conditions, the functional visual field may constrict. As a consequence, visual signals in the periphery may go unnoticed, reaction time may be signi- ficantly lengthened, and a significant increase in stimulus energy may be required for detection of signals. Since there is less overall energy available under low illuminance conditions, it follows that the consequences of visual field narrowing under stress may be more signi- ficant at twilight and at night than under high illuminance conditions. Evaluations of night vision performance should be performed for physio- logical and psychological conditions that are known to produce visual field narrowing. Dynamic Visual Acuity The majority of visual evaluation tests and procedures have been designed for static observers viewing a static target. In most real- life situations, however, the observer, the target, or both are moving. Adequate tests for assessing functional capabilities under such dynamic conditions are generally not available. At the present time, we have little information pertaining to the functional properties of night vision under static versus dynamic conditions, individual differences in dynamic visual function, and the relationship between dynamic visual properties and task performance capabilities. We recommend that dynamic visual acuity and related tests be considered for evaluation by a night vision laboratory. In addition to providing dynamic visual test cor.di- t~ons, these procedures require the observer to coordinate oculomotor, sensory, and cognitive functions, thereby creating a more demanding visual task environment.
10 Glare Sensitivity The presence of a glare source can have a significant influence on a variety of visual functions, especially under conditions of low ambi- ent illumination (see, for example, Blackwell and Blackwell; Owsley; Bailey). Most of the information that has been obtained for glare disability has been derived from static displays. The influence of glare for dynamic viewing conditions needs to be evaluated, especially for task performance over extended periods of time. Aging and Vision It is important to evaluate the effects of normal aging on perfor- mance under low illumination levels (Owsley). Some investigators report that age-related changes in the ocular media produce an average transmission loss of 50 percent (0.3 log unit) every 13 years, while others have reported age-related transmission losses that are about half as large (0.15 log units every 13 years). Under high illumination these transmission losses may not be significant. During twilight and at night, however, age-related transmission losses may be a critical component of visual performance. We recommend that age-related changes in night vision performance be investigated as part of a night vision laboratory program. Computer Modeling The use of computer modeling and computational theories are playing an increasingly important role in vision research (Watson). Empirical data from optical, electrophysiological, psychophysical, perceptual, and cognitive studies of the visual system can be used to design a model, which can then be used to generate predictions about visual response properties under a variety of situations. Comparison of certain predictions of the model with experimental results can then be evaluated to refine the model and improve its applicability to specific problems. Computer modeling could be a valuable e tool for the study of night vision. The design of such a diodes requires specif ic, detailed information about the functional properties, interrelationships, and constraints of all components of the model. Thus, missing information, unspecified interactions among components, unknown stimulus and/or response properties, and other key aspects of the problem become readily apparent. This can serve as a guide to directing research to critical areas of night vision that are not well defined. In addition, the predictive capabilities of a com- putational model can often provide insights for complex problems and interactive, dynamic situations that might otherwise go unnoticed. In view of the current rudimentary understanding of the critical factors underlying the performance of night vision tasks, we recommend that computer modeling be considered as one of the activities of a night vision laboratory. 1
11 DEVELOPMENT OF NIGHT III SION SCREENING TE STS There are several purposes associated with screening night vision capabilities of Air Force personnel: (1) to identify those individuals with night vision problems stemming from ocular pathology; (2) to deter- mine that an individual has sufficient night vision capabilities to per- form a specific task; (3) to classify individuals on the basis of night vision capabilities so that the most critical night vision tasks can be carried out by personnel with the best night vision performance on appropriate screening tests; and (4) to periodically monitor individuals to ensure that adequate levels of night vision are being maintained. The first objective is readily achieved with the use of existing clinical evaluation methods (see Fishman; Berson), including a thorough ophthalmologic exam and history, psychophysical measures (dark adapta- tion, visual fields) and electrophysiologic studies (electroretinogram [ERG], electro-oculogram [EOG]~. A more formidable challenge is to design simple, rapid screening tests for night vision that will predict task performance and that will permit the reliable classification of individuals on the basis of their night vision capabilities. Previous research indicates that it is not possible to accurately predict visual performance at one background luminance level on the basis of performance at other background luminances, especially If they are separated by a large interval. Thus, it is not likely that a pho- topic vision screening test will be able to provide information that is predictive of night vision capabilities. In this view, the results of a thorough task analysis of various night vision operations will play an important role in the design of night vision screening tests. Depending on whether specif ic tasks are typically performed at mesopic or scotopic luminance levels, the time required for a night vision screening procedure can vary considerably. Presently, there are no standardized night vision screening tests that have been shown to be related to task performance. The design of night vision screening tests should ideally come from thorough labora- tory studies that are correlated with actual performance measures of night vision tasks. It will require considerable time and effort to obtain this information. In the meantime, there is general agreement that a performance-related night vision screening test should probably incorporate elements of contrast sensitivity, target detection and identification, and visual search in a free-viewing situation. A rapid screening o~ oculomotor performance at low luminance s can also be per- formed in a simple, straightforward manner. RECOMMENDED EQUIPMENT FOR A NIGHT VI SI ON LABORATORY A critical item for a night vision laboratory is a high-quality photometer/radiometer to measure and calibrate luminance and contrast levels accurately. The instrument should be capable of determining luminance p illu.minance, color temperature, and other measurements for a wide range of lighting conditions, object sizes, and temporal stimula- t ion cond i t ion s .
12 There are several commercial devices for clinical evaluation of night vision that would also be appropriate for a night vision labora- tory. A sophisticated dark adaptometer, a clinical electrophysiological test system for measuring EGG and ERG responses, and an automated pro- jection perimeter would be important equipment items for a night vision laboratory. A number of clinical devices for the measurement of glare disability have also become available. Many of these devices have not undergone formal evaluations to determine their accuracy, reliability, and validity. In addition, there are some glare devices that use a high background luminance level that would not be appropriate for investiga- tions of night vision. Although a device to measure glare disability is highly recommended for a night vision laboratory, several issues must be carefully considered prior to purchasing a specific glare tester. In particular, the device should be applicable to the measurement of glare disability under low luminance conditions, its performance characteris- tics (sensitivity, accuracy, reliability) should be documented in pub- lished studies, and normative values should be available for the general population. A key equipment item for a night vision laboratory is a general pur- pose laboratory computer system. The computer system should be capable of real-time control of stimulus displays and other laboratory devices, as well as acquisition of psychophysical and electrophysiological res- ponses. A full complement of analog and digital I/O interface boards (and accompanying control software), peripheral devices (printer, plot- ter, graphic terminal, etc.), and disk storage should be included in the computer system. In addition, a high-quality spreadsheet/data base management software package should be included to permit correlations of laboratory findings with performance scores or ratings from night vision tasks, longitudinal evaluations of individuals, determinations of the efficacy of various training regimens, and other related issues. The establishment of a comprehensive data base is an important aspect of long-range planning for a night vision laboratory. In order to accommodate the needs of computer modeling, there should be suf f icient memory and mass storage available on the computer system to allow computational models of night vision characteristics to be evaluated. In addition to a computer system, other general psychophysical laboratory equipment (high-resolution display oscilloscopes, function generators, optical bench equipment, filters, etc.) should also be available. For studies of per ipheral vision, a manual per imeter than can measure a variety of v, sual functions (detection, resolution, flicker sensitivity, dark adaptation, etc.) over a large range of back- ground luminance s would be extremely useful. A large screen monitor or CRT display would also be valuable for assessment of perhipheral visual function. Evaluations of oculomotor function can be carried out with several devices. A laser or vernier optometer can provide steady-state measure- ments of accommodation, and steady-state convergence measures can be determined by the use of a device incorporating the nonius line tech- nique. For measurements of the dynamic properties of accommodation, a continuously recording infrared optometer is recommended. The dynamics of conjugate and disconjugate eye movements should be evaluated with an
13 infrared eye movement monitoring system. An infrared scleral reflec- tion-type system mounted on a trial lens frame is a relatively inex- pensive device that can be used for this purpose. For more comprehen- sive evaluations of eye movements, an infrared eye-tracker that moni- tors the relative positions of the Purkinje images would be a most useful tool in a night vision laboratory. Ideally, one would like to have night vision laboratory personnel with expertise in the following areas: (1) visual psychophysics, (2) human electrophysiology, (3) human factors, (4) clinical ophthalmic sciences, (5) industrial psychology, and (6) illuminating engineering. This multidisciplinary team would provide a comprehensive background for addressing the variety of complex problems associated with night · ~ vlslon. ,.(
