Click for next page ( 304


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 303
SYMPOSIUM OVERVIEW: THREE PERSPECTIVES

OCR for page 303

OCR for page 303
PER SPECT IVE: SVELTE R MAKOUS - This volume contains a variety of interesting and excellent papers that represent a good sample of current research relating to night vision. It might not be totally out of line for me to congratulate the organizers of the meeting or which this volume is based on their excel- lent job and to thank the sponsors for their support. Here we focus on the pu rposes of the meeting on which this volume is based, and so I shall start with practical issues in the following cats gories: (1) current techniques, (2) personnel and equipment for a night vision laboratory, and (3 ) suggestions for improving night vision. CURRENT TECh~' IQU~ S I submit that the progress of science, visual science at least, is primarily limited by its techniques; i.e., periods of rapid advance depend on new techniques. The Visual Mechanism Advances in understanding the structure, physiology, and biochemis- try of the visual system have been especially rapid. Development of new histological stains and markers that entirely fill the cells that are in jected, development of biochemical tracers, ana advances in elec- tron microscopic techniques have produced a renaissance in neuroanatomy. Our view of the structure of the visual system has been revolutionized since the early 1960s; and the scale of the questions being asked has changed from that of the connections between gross structures of the brain, is passing through that of connections between cortical layers, and has begun to focus on connections between classes of cells within layers and between areas. Use of these new anatomical techniques in combination with current electrophysiological techniques and use of 2-deoxyglucose, which allows inentif ication of entire sets of st~mu- lated or otherwise active neurons, have 9 iven a new meaning to the concept of functional neuroanatomy. The proliferation of identified and putative neurotransir,itters (ana modulators, along with new methods for studying them, has created an 305

OCR for page 303
306 entirely new f ield concerned with identifying the transmitters used by different classes of cells and using this information to modify the sys- tem selectively. These workers promise soon to replace present methods of testing function through irreversible destruction of spatially defined regions of the brain with a spectrum of chemicals that will reversibly inactivate specific physiological functions or mechanisms. Development of the microelectrode since the 1950s has, of course, t rans formed neurophysiology, and to stain and identify impaled cells after recording is now routine. Recent introduction of suction and patch electrodes to investigations of the physiology of photoreceptors and other retinal cells has permitted, in a short tune, precise answers to questions that had defied even approximate answers for a century. Microspectrophotomet ry is also coming of age, although measurement of the membrane cur rents of outer segments may super sede these teab- nigues for many purposes. Reflection densitometry in the living eye is improving, but has yet to prove its validity and usefulness for many purposes. Funct tonal Propert ies: Psychophys Arcs The advances in psychophysics have not been as dramatic, but there have been some. Stimul i Perhaps the greatest advances for psychophysicists are in the gener- ation of stimuli. Improvement of the quality of cathode ray tube {CRT) devices and their control by computers have increased the complexity of stimul i in all the dimensions of time, space, and color. With suitable interf acing and computer control, it is now possible to produce temporal and spatial modulations along any vector in color three-dimensional space . Th is has produced significant advances in the neurophysiology (Derrington et al., 1984) and psychophysics (Krauskopf et al., 1984; D'Zmura and Lennie, in press) of color vision. Sine~ave modulations in time and space are now ubiquitous, as are the mathematical tools that such stimuli foster. The study of motion, facilitated by the new display capabilities, has seen a recent rebirth. Manipulat ion of the random propert ies of stimuli has been increasingly f ru itful, following its initial exploitation for the investigation of stereopsis by Juleez ( 1971) . Perfection of the laser interferometer by Williams (1985) promises to open previously inaccessible domains for exploration, for interfero- n~eters are powerf ul j ust where Cams are weak: while the luminances of CRTs are limited to about 100 milliLamberts (mL), there is practically no limit to the ret inal illuminances that lasers can achieve; inter- ferometers increase the spatial frequencies that can be thrown on the retina f row about 60 cycles/degree (cpd) 'co over 200 cpd; they raise the temporal frequencies possible from about 30 Hz, or at most 1 MHz, to 1 Liz; finally, they erect sine-wave gratings with no measurable d istort ion. Control of color and phase are under development.

OCR for page 303
307 Responses Technical improvements have also enhanced measurement of responses. New instruments are available for inobtrusive observation of eye posi- tion, accommodation, and pupil size; and these-have been used to close the loop so that stimuli can be presented in stable positions on the retina, i.e., independent of normal eye movements, without direct ocu- lar attachments. Computer analysis of neuroelectric data has enhanced the usefulness of such measures of response as the electroretinogram and the electroen- cephalogram. One would be remiss not to mention the development of magnitude estimation. Although its use has not penetrated visual psychophysics as much as some other areas, it can sometimes be used validly to gather information more efficiently than by thresholds, and it provides a means of answering questions not accessible to threshold techniques. ; Other Areas The theory of signal detection has clarified the nature of thresh- olds and led to techniques that separate shifts of the criterion for responding from changes of sensitivity. Chief among these is the fcrcea-choice technique' introduced by Blackwell (1952). Laboratory computers have facilitated the use of these techniques and have per- mitted the efficient choice of the stimulus for each trial from the previously gathered information.- Combined tests of identification with detection have enhanced the power of inferences about underlying mech- anisms, and a recent issue of the Journal of the Optical Societv of America (2(9):1455-1610, 1985) has brought together a set of papers def ining the theory and illustrating the power of these techniques. Techniques in the field of human factors have changed as attention has shifted from things like anthropo~,etry and the design of dials and switches to things like perception and human information processing. Several important advances are encapsulated by the process of total system design (Bailey, 1982), which fosters the organization of human factors projects and promotes consideration of all the relevant factors at the appropriate stages of the project. Finally, there have been conceptual advances that are driven by technology. Among these are the increasing complexity and usefulness of models, as described in this volume by Andrew Watson.. Attempts to build machines that "see" have focused attention on more rigorous specification of the problems that the visual system must solve in the act of seeing and the constraints imposed by the information available in the environment. A N IGHT VI SI OF LAbO~TORY The goals of an Air Force night vision laboratory evidently would include (1) developing techniques to evaluate the night vision of

OCR for page 303
308 personnel and (2) answering more general questions about night vision. The first goal requires personnel that are trained and experienced in human factors, and the equipment would depend on the results of the initial steps of the program. The second goal requires personnel that are trained and experienced in visual psychophysics. Naturally , it is easiest and least expensive (disregarding the costs of delays) to choose the apparatuses after the experiments have been planned. There are, however, some pieces of apparatus that will certainly be neeaea, whatever the specific experiments. Essential Apparatus Photometer Whatever the experiment, the stimulus will have to be measured. A recent and useful innovation allows self-calibration of spectrally flat photodetectors that have extraordinarily high efficiency (Zalewski and Geist, 1980). Commercial implementation of this technology is now available (e.~., the UDT Model QED-200 absolute radiometric standards . In the long run, one will almost certainly want to measure the lumi- nances of small areas, as on a CRT; hence, a telescopic probe or "spot" meter will be required. Finally, most vision laborator ies eventually find it necessary to make spectral measurements and calibrations; so a spectro~adiometer (cf. Geist et al., 198C) will ultimately be needed, and it might as well be compatible with the telescopic attachment. Compu ter Any efficient laboratory will also require a laboratory computer, something like a PDP-ll or Microvax. Hard disks are required for adequate speed and storage, and dual floppy disks are required for transfer of data and software. Plentiful analog and digital lines in and out, and a real-time clock are required to monitor and control stimul i and to record responses. On-line computation requires f loafing-point hardware . The sof tware should provide automated control of the exper laments, providing for both forced choice and method of ad justment. I t should be possible to determine the stimulus parameters on each trial f roar, the accumulated data at the time (e.g., QUEST; Matson ana Pelli, 1983~. Access to a large, fast computer also is needed for modeling and possibly for data analysis. Stimuli Stimulus properties and their control depend on the specif ic exper iment, and perhaps the choice of the means of generating their is best left until the research is planned. However, in over half of the psychophysical papers in a recent, haphazardly chosen issue of Vision

OCR for page 303
309 Research a CRT of some sort was used to generate the stimuli. So a CRT system for generating stimuli is bound to be useful. At present, the closest to a general-pu rpose stimuli s generator that is commercially avai fable is the Picasso system. There are a number of devices, such as the ADAGE system, that provide the added dimensions of color and that are as flexible as the speed of the computer and the ingenuity of the programmer allow; but these are not commercially available as finished stimulus generators. Modeling The responsibilities of the laboratory are likely to require better access to the wealth of facts about vision, and their implications, than is now available f rom the literature and the minds of individual investigators. Hence, a system of the kind described in this volume by Andrew Watson is likely to be valuable in making the best guess, in the absence of a specif ic experiment, about the performance of the visual system in any g iven situation. Worthwhile Nonessentials The rest of the equ ipment needed will probably be dictated by the specif ic experiments that are planned. however, there are a few instru- ments that are also likely to be of value over a period of several years in an active vision laboratory concerned with practical problems. As the spatial properties of visual stimuli depend on the accommo- dati~re state of the observer, it is desirable in many situations, espe- c tally those outside the laboratory, to know the accommodative state of typical observers. Hence, an infrared optometer would be valuable. The locus on the retina at which stimuli fall is similarly impor- tant, and outside the laboratory the observers control it, not the experimenter. So in applied research it may often be desirable to know- where typical observers fixate. The best device for measur ing eye position depends on the specific application. However, for combined precision and convenience of use, it is hard to beat the the SRI eye- tracker. It measures the vertical and horizontal or ientation of the eye as well as accommodation, and allows use of these measurements for feedback control over the stimulus position. However, though this device requires no attachments to the observer, it does have limita- tions that complicate its use in many practical applications. In conclud ing any comments on the laboratory, I draw attention to the f act that answer s to many questions about vision depend on the visual ecology of the observer. In most practical situations, little is known about the visual environment. What are the luminances under which personnel work at night? What are the properties of visual tar- gets? What are their backgrounds? Typically' the answers to questions such as these are mere guesses. A valuable contribution to the under- standing of night vision would be to develop a means of sampling the visual ecology of those who must see at night. I suggest that the laboratory develop procedures and instruments for establishing the visual character istics of the tasks on which the laboratory is asked

OCR for page 303
310 to work. An example of an approach they might pursue is that used by Switkes et al. (1978) to measure the orientation and spectrum of spa- tial frequencies in different visual environments. IMPROVING N IGHT VI SI ON Here I list some possible ways of improving night vision, most of which arose dur ing the meeting on which this volume is based. Not all are practical, but it may be of some use to consider them anyway. Collect more light by: -- illuminating the ob jects of vision; -- using optical aids (e.g., telescopes primarily collect light); -- enlarg ing (more light) or constricting (better resolution) the pupil. Image intensif iers can improve performance 1OO times (van Meeteren, 1978) . Explore individual differences to determine whether selecting individuals for their night vision might be worthwhile. Crawford ( 1947) repo; ted a 60 percent standard deviation in recovery times (dark adaptation) among 26 nonclinical subjects. Flicker the f ield of view, as described in this volume by Donald MacLeod. o Ad just the magnif ication to put the important spat ial frequencies into the range of S-10 cpd, to which the eye is most sens i t ive . o Reduce intrinsic noise in the visual system -one source of such noise evidently arises from a dark-adapted eye (Makous et al., 1976; Auerbach and Peachey, 1984~; hence, there may be some benefit in light adapting one eye (to remove its contribution of noise) while using the other, dark-adaptea eye for seeing. Presenting and then removing ~ dim annulus also seems to leave the region surrounded by the annulus briefly more sensitive (Pulos and Makous, 1982~. Con- versely, avoid dark adapting one eye while trying to see with the other eye, as might happen if one eye is patched for some reason or while one eye is viewing through a monocular instrument or heads-up d isplay. I f an eye is to be patched for some reason, a translucent patch might be best. Conversely, Birch et al. (1980) reported that patching an eye for a per iod of a week lowers the absolute threshold. O Speed up afar k adaptation, or at least avoid slowing i t down: ~ onger 1 ight adaptation slows the process of recovery, and there is obscure evidence, which I have not been able to track down, that days or weeks of dark adaptation can increase the desensitizing ef feet of light adaptation in cats.

OCR for page 303
311 Screen for individuals with visual pathologies, especially retinitis pigmentosa, as described in this volume by Eliot Benson and by Gerald Fishman. Training and practice--there are few things that do not improve with practice. Some individuals require practice to learn to use per- ipheral or off-axis vision ~ i.e., the astronomer' s trick) . ~ Large doses of vitamin A--this seems like a long shot, as the evidence indicates that it does not help. According to theory, it only helps individuals with a deficit of vitamin A. How many indi- viduals have such def icits is, perhaps, an open question, but it seems unlikely that there are many; Dowling and Wald (1960} found that simply depriving rats of vitamin A was not sufficient to produce noticeable effects on their vision--the animals had to be depr ivea f rom birth. REFERENCES Auerbach, E., and N.S. Peachey 1984 Interocular transfer and dark adaptation to long-wave test lights. Vision Research 24 :1043-1048. Bailey, R.~ . 1982 Human Performance End ineer ing: A Guide for System Designers. - Englewood Cliffs, N.J.: Prentice-Hall. Birch, D.G., E.E. Birch, and J.M. Enoch 19 8 0 Vi sue l sens i t iv ity, resolu t ion, and Rayle is h matche s follow ing monocu la r occ lu s ion f or one wee k . Jou r nal of the Opt ical Society of Amer ice A 70: 954-958. Blackwell, Y. . R. 1952 Studies of psychophysical methods for measur ing visual thresholds. Journal of the Optical Society of America 42: 606-616. Crawford, B .H. 1947 Visual adaptation in relation to brief conditioning stimuli. Proceedings of the Royal Society, London B134 :283-302. D' Zmura, M ., and P. Lennie In Shared pathways for rod and cone vision. vi Lion R-~;'r~h press Derrington, A.~., J. Krauskopf, and P. Lennie 1984 Chromatic mechanisms in lateral geniculate nucleus of macaque. Journal of Physiology 357 :241-26S. Dowling, J .E . and Wald, G. 1960 The biological function of vitamin A acid. Proceedings of the National Academy of Sciences USA 46: 587-608. Geist, J., E.~. 2alewski, and A.R. Schaefer 1980 Spectral response self-calibration and interpolation of silicon photod ioces. App: fed Optics 19: 3795-3199. Julesz, B. 971 Foundations of Cyclopean Perception. Chicago: The University of Chicago Press.

OCR for page 303
312 Krauslcopf, J., D.R. Williams, and D.W. Heeley 1982 In search of the cardinal directions in color space. Vision Research 22 :1123-1131. Makous, W., D. Teller, and R. Boothe 1976 Binocular interaction in the dark. Pulos, E., and W. Makous 1982 Changes of visual sensitivity caused by on- and off-transients. Vision Research 22 :879-887. . Switkes, E., M.J. Mayer, and J.A. Sloan 1978 Spatial f requency analysis of the visual envi ronment: Anisotropy and the carpentered environment hypothesis. Vision Research 18 :1393-1399. van Meeteren, A . 1978 On the detective quantum eff iciency of the human eye. Vision Research 18 :257-267. - Watson, A.B., and D.G. Pelli 1983 QUEST: A Bayesian adaptive psychometr ic method. and Psychophysics 33: 113-120. Willia~r.s, D.R. 1985 Aliasing in human foveal vision. Vision Research 25 :195-205. Zalewski, E .~., and J. Geist 1980 Applied Optics 19:1214 . Vi sion Research 16: 473-476. Perception .

OCR for page 303
PERSPECTIVE: DONALD G. PITTS In this overview, I will highlight four major points mentioned in the papers in this volume. This is not meant to be an overview of all the topics that were presented. SPECIF ICATI ON OF LUMINANCE LEVELS ~ was very intrigued by Walter Makous' work in which he states that if you put light in one eye, you will lower the threshold in the other eye. My problem is that he did not give the levels and he did not give the dimensions of the targets. If this work is to be useful, we will need those spec if icat ion s. Alfred Owens, describing his dark focus research, also did not spec- if y what light levels th is dark focus involves and whether it changes with levels of light. For an operational problem, these types of data are very important. HUMAN PERFORMANCE ISSUES Lewis Harvey described his work with modulation sensitivity, which he calls "contrast sensitivity. n He mentioned a decrease in the peak. Under dark or less light conditions, this is not all that happens to the system. There is not only a decrease in the height of the peak but also a shift in the high frequencies. If you look at the curves you will see a shift. In addition, you have a shift in the peak f requency. It not only occurs with decrements of light, it also occurs with age. Stan Smirch talked about the Blackwells' data. I was disappointed that there were no graphs comparing those formulas with anything. Those formulas do not mean anything unless 9 raphs can be seen that relate them to human performance, preferably giving the formula line on the graph and hut an performance for that formula as a comparison against it. Dr. Smith said that the cornea is the source of the stray light . I would like to challenge that statement because we do not f ind great changes in transmittance in the cornea with age. me lens must be the source of stray light. I would cite data by Peale of England, 313

OCR for page 303
314 our laboratory, and Wolf's laboratory. They are not new data, rather they are old data. I was very intr igued by the possibility of utilizing part of Donald MacLeod's data to increase target recognition and perception. But we need to have a reg imen or methodology for doing this. In other words, we need to know how to take advantage of this increased interaction. Harold Bedell is in our laboratory down the hallway. We talk occasion- ally, but I did not know that he was working on spatial localization. The ideas of enhancement, which he brought out at this meeting, and of localization of the area of the retina both need to be expanded further. In fact, I think these are ideas that the military should look at very closely. NIGHT VISI ON TESTING I was rather intrigued that Ian Bailey came here and provided the general rules for mobility and levels of light as well as general rules for reading. These rules are very practical. Then he discussed two tests: the contrast board test and the threshold test. These are the types of tests that the military could use as screening tests. They are simple to administer and interpret. They provide a method to determine: You can go or not go. You have sufficient light vision to begin or you do not have it. " We are talking about whether we can determine who can serve in the military and what level of night vision is required for various jobs. Of course, we can take a great deal of time to select the F16 pilot who flys a 830-million aircraf t. For other personnel, a test like the one Dr. Bailey described would be very helpful. LITERATURE REVIEWS I would also like to make a few comments about A. Sanders and his work: he is the only European representative we have here, and he brought a refreshing view. He gave us a good review of the literature in this area and some good ideas for future directions. His talk was really r ich: take a good look at the references he gave alone. Inci- dentally, some of the psychologists here may not remember or nearer knew that the or iginal research done in this area by the Air Force was con- ducted by Bob Boynton, before he got into color vision. He did the or ig inal ~ search and published three or four thick reports back in the 1950s. So, he is another classic investigator at whom we should take a look. He was the f irst one to study the effects of clutter. CONCLUSION These are just some ideas that I picked up as we went through the meeting. I know there are many more ideas that I have ~r.issec,. We lack knowledge of the job requirements that the military is con- cerned about. Do not be su rpr isea that we do not get the information

OCR for page 303
315 we need for the simple reason that some of these jobs are ~classified.. We have heard reference to performance tests, but we have not defined which performers or what tests should be given. T f we are going to refer to performance tests, I think we need to identify them at least. Another area that is missing is computers. In order to do the job that the military needs, we need computers for storage and analysis of the data. We may need some software people to do the job. We had a lot of good information presented here that the U.S. Air Force School of Aerospace Medicine can use to help solve operational problems. The f inal utilization of that information depends upon the people here helping us pull this report together and get it to the people who asked us to do the job for their..

OCR for page 303
PERSPECTIVE: W] LLIAll A . MONACO - I think that the authors of the two previous summaries have made it easier for me to summarize, because they covered much of what I was going to discuss. Let me back up, though, and relate what has been presented in this volume to a meeting that I attended in October 1985. mat meeting addressed visual standards, that is, trying to relate visual performance to tasks and personnel selection. It is a very difficult problem. As a matter of fact, it was difficult enough that one of the participants, after we had been there for a couple of days, got up and said, HI think we're all suffering from practical apathy.. By that he meant that the problem is already too hard. We can only compound that problem by making more tests and by expanding them. Because I am in the military and because I have been the Navy representative to the National Research Council' s Committee on Vision, I have tried to respond to the presentations at this meeting with a d if ferent k ind of hat that i s, the hat of the user rather than that of the sc dentist or clinician. I have tried to put together some ideas, based on what has been said, that are directly applicable to the kinds of questions that General Doppelt and Colonel Tredici have asked: bow do we see at night? Chow do we OCR for page 303
317 issues using available information. he actually went to the fleet and stressed the importance of knowing what they did' trying to quantity the environment in a productive way. And I can state from practical experience that they are very cooperative and very much interested in ways to make their job easier. So, based on that experience, I would like to key in on some of the comments that have been made in this vol- ume and maybe stimulate thinking in a slightly different direction-- that is, toward more applied, operational, or clinical applications that may address this difficult problem. LEVELS OF ILLUMINATION We started with Herschel Leibowitz quantifying light levels. Granted, it is not specific, but it is quantification of the light levels. Were are data available through the Air Force that indicate that we do have some data in actual f light scenar ios for quantifying that environment. PHOTORECEPTOR PROPERTIES - Regarding David Copenhagen's comment about the stability of rhodop- sin permitting us to see at low light levels, is it possible that we can control this? That is a basic science kind of question. Gnat can we do with existing clinical tools? When I asked that question at the initial session, Eliot Berson volunteered that you could selectively screen some of the more obvious problems from a family history. The flicker issue might be one that is diagnostic. Dietary manipulation to extend our sensitivity is another consideration. Gerald Fishman described the modification of the adaptometer. I was very impressed with his approach to an existing clinical tool ana his modification of it in such a way that it asked a different clinical question. I think that that is a manageable way of dealing with some of these problems. OCULOMOTOR AND SPATIAL ORIENTATION FACTORS Alfred Owens discussed clinical screening for dark focus, but is it necessary to screen for it? hill retinoscopy do an effective job of screening? Could we just do visual acuities at mesopic levels of illu- mination? Robert Post mentioned the role of the environment--like a cockpit destroying the input for relative motion. Sheldon Ebenholtz addressed vestibular-ocular response and the distinction to be drawn between the central and the peripheral retina, which I think have training implica- tions.

OCR for page 303
318 SPATIAL AND TEMPORAL FACTORS Harold Bedell stressed the distinction between Vernier acuity and visual acuity. Is there a way we can modify that or incorporate it into our existing test battery? Variability in fixation is something I think that we have to be very concerned about, particularly when we are asking questions like where does a person look, what is his scanning performance, and what is the importance of knowing where to look in particular in target detection. Can we optimize it with screening and training? Lewis Harvey described the importance of contrast sensitivity and/or edge detection. Contrast sensitivity may have applicability in a more manageable format. He pointed out that the visual system is most acti- vated by something with a high visual demand--is that applicable to demand? If it is, are there design criteria that we can put into air- craft, tanks, or anything else that would draw their attention to that particular piece of equipment at the appropriate time? Ralph Haber talked about defining the task. Again, in the Navy we do have a way of quantifying that. It is not perfect, but we try. Donald MacLeod mentioned rod sensitivity and inoculating the background, in effect, increasing the sensitivity of the rods. Again, that might be a good indication for an engineering specification. We have what are called Heads-up displays. in al rcraf t that supposedly keep a per- son's eyes out of the cockpit and increase the support of seeing things that are threats. Can we optimize the design of those heads-up displays in such a way that it optimizes the sensitivity of vision? Ian Bailey's comments about low vision offer an excellent model and one that we in the Navy have tried to pursue through the Pennsylvania College of Optometry's Feinbloom Center. The fact that people with restrictions of visual f ield have trouble locomoting is no clitterent than a person who is moving at very high velocities with the same kinds of restrictions, so those are good models. HIJR~N PERFORCE I SS ~ S A. Sanders addressed the importance of visual search. Th is cannot be overemphasized. This is crucial to our being successful in many mission scenarios. It also lends itself well to training. Andrew Watson's comments about modeling offered another excellent way to approach the development of our understanding of night vision functions.