Journal of Neuro-Ophthalmology, March 2004, Volume 24, Issue 1

What's New in Perimetry

Despite important refinements that have improved quantitation and shortened test time, modern perimetry remains relatively insensitive and plagued by high test-retest variability. Some novel methods, though not yet fully vetted, offer the promise of improving sensitivity and reducing variability.

STATE OF THE ART What's New in Perimetry Michael Wall, MD Abstract: Despite important refinements that have im­proved quantitation and shortened test time, modern perim­etry remains relatively insensitive and plagued by high test-retest variability. Some novel methods, though not yet fully vetted, offer the promise of improving sensitivity and re­ducing variability. ( JNeuro- Ophthalmol 2004; 24: 46- 55) HISTORICAL ASPECTS Perimetry was introduced into clinical medicine in 1856 when von Graefe ( 1), using a chalk board and a piece of chalk, mapped scotomas and the physiologic blind spot, visual field constriction, and hemianopias. Using a fixation point, he described his experience with this type of kinetic flat surface perimetry ( campimetry), in which patients were asked to respond when they saw a light target emerge from a dimmer background. Campimetry could only map the central visual field. About a year later, Aubert and Forster ( 2) introduced the arc perimeter. The use of an arc- shaped arm enabled the full limits of the visual field to be investigated. This device dominated the field for the next 15 years. In 1889, Bjerrum ( 3) reintroduced campimetry. With his student, Ronne, he performed quantitative isopter pe­rimetry and mapped the prototypic scotoma of glaucoma, the arcuate defect that breaks out from the blind spot and bears his name. Ferree et al. ( 4,5), Sloan ( 6), and Aulhorn and Harms ( 7) all contributed significantly in the ensuing years. How­ever, perimetry took its present form in 1945 when Gold-mann ( 8,9) introduced the bowl perimeter. Using a single light source, a set of mirrors, and neutral density filters, this device allowed exquisite control of the background and stimulus luminance. This type of differential light sensitiv- University of Iowa, College of Medicine, Departments of Neurology and Ophthalmology, Veterans Administration Medical Center, Iowa City, Iowa. Address correspondence to Michael Wall, MD, University of Iowa, College of Medicine, Departments of Neurology and Ophthalmology, Veterans Administration Medical Center, Iowa City, IA 52246. E- mail: michael- wall@ uiowa. edu. ity became the standard for perimetry and allowed testing conditions to be uniform throughout the world. The use of the pantograph enabled elegant mapping of defect shape. Customized test point location and improvisation of strategies in real time still allows the most accurate map­ping of visual field defect shape. High resolution shape in­formation obtained this way can be invaluable for neuro-ophthalmologic diagnosis. However, there is a long training and testing time for perimetrists, and a relatively insensitiv-ity to shallow defects compared with static testing. Work by Lynn ( 10), Fankhauser et al. ( 11), Heijl ( 12,13), and Krakau ( 14) in the 1970s laid the foundation for current differential light sensitivity automated perim­etry. These investigators basically automated the Gold-mann bowl device. They standardized test location with a 6°- spaced grid using a size III differential light sensitivity stimulus. A 4 dB/ 2 dB staircase test bracketing procedure was used over a four log unit range. In 1980, the Octopus 201 perimeter ( Haag Streit AG, Berne, Switzerland) was launched with a price of about $ 100,000. Later in the de­cade, other perimeters were introduced until the Humphrey Visual Field Analyzer 600 series ( Humphrey Instruments, San Leandro, CA) became the industry standard. The Humphrey instrument brought some significant advantages. Although defect- shape rendering is coarse be­cause of the 6° spacing of stimuli, static perimetry proved to be more sensitive than kinetic testing. The test is very good for quantitation, and the testing conditions are reproducible. The long and complicated training required of perimetrists using the Goldmann device ( 15) was shortened with auto­mated perimetry. However, the objective of having a mini­mally trained individual perform the test has not been real­ized. Even though the procedure is automated, considerable human interaction is necessary to obtain good studies. Vari­ous reports in the last two decades have emphasized the importance of standardized test instructions ( 16), quality control ( 17), reliability indices ( 18,19), and eye monitoring. A series of artifacts that interfere with interpretation have been recognized ( 20- 22). Before the advent of the Swedish Interactive Thresh­old Algorithm ( SITA), the static automated visual field examination took 15 minutes or more ( 23). SITA has halved the test time but the test remains a boring and diffi­cult task. 46 J Neuro- Ophthalmol, Vol. 24, No. 1, 2004 What's New in Perimetry JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 Test- Retest Variability: A Major Shortcoming Among the greatest shortcomings of conventional automated perimetry is its high test- retest variability. For defects of 1 log unit ( 10 dB) or greater, there is an exponen­tial rise in test- retest variability. The enhancements of con­ventional perimetry, including elegant statistical indices, SIT A, and Short- Wavelength Sensitive Automated Perim­etry, have not eliminated this high test- retest variability. Why does this variability occur? Automated perim­etry was developed on normal subjects who have a prob-ability- of- seeing curve with an extremely steep slope ( Fig. 1), indicating high reproducibility. The developers of the technique did not anticipate that visual field damage would flatten the slope of the frequency- of- seeing curve ( Fig. 2). For example, the subject in Figure 2, whose visual field re­sults are typical of 10 tested glaucoma subjects ( 24), has a 10- 20 dB loss in a glaucomatous visual field defect. Note the shallow slope of the frequency- of- seeing curve and the broad potential range of values that may occur within the visual field defect. The rise in variability has been shown to be exponential. This leaves conventional automated perim­etry with a critical flaw: with increasing visual field dam­age, variability increases and impedes the ability to detect visual field change over time. So, in the very patients in whom one needs to detect change, test- retest variability in­terferes with that detection. In glaucoma patients with nor­mal sensitivity determined by Statpac software, the fre­quency- of- seeing curve is also often abnormal ( 24). Many factors contribute to test- retest variability, in­cluding fatigue, temperature, time of the day, learning ef­fect, threshold, attention ( 24a), and testing strategy. Among 100% T 90% 4- 80% a 70% - $ 60% C 50% | 40% £ 30% - 20% -- 10% 0% 10 20 30 Light Intensity ( dB) 40 FIGURE 1 . Frequency- of- seeing curve in a normal subject. The curve is created by intensively testing each location repeatedly with a range of stimuli. The frequency at which a stimulus is seen ( in percent) is graphed against the stimu­lus intensity. The stimulus location used to generate the frequency- of- seeing curve is indicated by the upper arrow in the inset. The steep slope signifies low variability. 10 0 30 Light I n t e n s i t y ( dB) 40 u 100% 90% 80% " 70% - 60% 50% - 40% 30% - 20% 10% 0% 0 10 B Light Intensity ( dB) FIGURE 2. Frequency- of- seeing curves in a patient with 10- 20 dB loss from glaucoma. Stimulus location is indi­cated by the left arrow in the inset. A. Stimulus size III. Note the shallow slope of the curve, signifying high variability. B. Stimulus size V. Note the steep slope of the curve, sig­nifying low variability. optic neuritis patients whose visual fields were tested 5 times in 1 day, and once a week for 5 weeks, 3 variability patterns emerged ( 25) ( Fig. 3). About one- third of subjects had reproducible visual field defects (" consistent" defects; Figure 3), about one- third had moderate variability, and about one- third had high variability defects (" variable" de­fects; Figure 3). All subjects had excellent reliability indi­ces and were tested by trained perimetrists. Fortunately, a strategy exists to reduce a substantial portion of the variability. Figure 2b shows a frequency- of-seeing curve at the same test location in the same subject as in Figure 2a, except that a size V stimulus was used instead of the standard size III. This pattern of lower variability with frequency- of- seeing curves using size V as compared with size III stimuli held across a group of 10 subjects ( 24). The larger stimulus is thought to raise the signal- to- noise ratio and result in lower variability. Small stimuli may first 47 JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 Wall CONSISTENT SAME DAY DIFFERENT DAYS m m iOtAn UL• a * i i gj • I i 1 • * If • a H i l l m i l KIII VARIABLE SAME DAY DIFFERENT DAYS « 1 1 1 * if I l f l l ' I I * *[ • * •? » a! I t l l l 1 1 1 ! n • i : » ; a m \ I f i l f • Suva I I I • a g m s: @ i > i X : l l i i * « ft i i ft a s i l l l i ' # * * » i: ,||", j|-^--" T,::" ™ ~"^ g I I I « • t t f ! ff * » » • * m m • a a • i l l i l l I I a UC Davis • Br --•.. I S ™ l i s S! *' Tiwi" ir:"": : i 1 1 I I i - « * I S i l l ' * i # m * f • i s si* 1 I I a ml I f i l « l UC Pawls FIGURE 3. Repeated visual field testing in two patients with optic neuritis. The left half of the figure shows the results of testing a single patient who displayed consistency. The right half of the figure shows the results of testing a single patient who displayed variability. Same day columns display the results of testing every two hours during the same day. Different days columns display the results of testing on different days matched for time of day. be seen and later missed because of microfixation shifts that result in the stimulus first falling on an intact portion of the visual field and later onto a defect. A larger stimulus would fall onto a larger number of receptive fields so that small fixation shifts would have less of an effect on frequency-of- seeing. But using a larger stimulus reduces sensitivity for small defects. An appropriate compromise is to use a perimetric method that finds threshold by changing stimu­lus size. The ring test ( high- pass resolution perimetry), which uses this method, resulted in low variability ( 26) and earlier detection of glaucoma progression in a multi-year study ( 27). Newer Perimetric Methods The visual field methods introduced in the past 10 years include short wavelength sensitive perimetry ( 28), pattern discrimination perimetry ( 29), color contrast perimetry ( 30), and contrast sensitivity perimetry ( 31), noise field campimetry ( 32), pupil ( 33), and flicker perim­etry ( 34). None of these methods has a substantial incre­mental increase in sensitivity, but some have lower variabil­ity than conventional perimetry. SITA An advance in reducing the test time of perimetry has come from SITA. The SITA method uses two maximum likelihood visual field models: normal and glaucoma. The likelihood of a set of data is the probability of obtaining that particular set of data, given the chosen probability dis­tribution model ( in this case either a normal or glaucoma model). These models are based on the fact that the slope of the frequency- of- seeing curves ( variability) increases with threshold. Figure 4 shows the two models. Log likelihood is on the y- axis and threshold ( intensity) on the x- axis. At the 48 © 2004 Lippincott Williams & Wilkins What's New in Perimetry JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 O O T - j bJO O 30 40 Threshold value ( dB) 5-. f question # 46 seen 10 20 30 40 T - I " 1 I I I i I I ' 0 10 20 30 40 i i i i i- rTT ™ n 0 10 20 30 40 FIGURE 4. Visual field models for normal and glaucoma results and their change throughout the pe­rimetry test using SITA ( 23). As the test continues, the maximum like­lihood model is updated based on the results at the tested location and surrounding correlated loca­tions. In this example, the glau­coma model gradually becomes the most likely model by the end of the test. 10 20 30 40 10 20 30 40 Intensity ( dB) beginning of the test, there is a likelihood that the test loca­tion results are either normal or abnormal. As the test progresses, the model is updated until the threshold is estimated with the given level of confidence. In an effort to reduce test time, the SITA Tast algorithms interrupt the threshold even earlier. However, this reduction in test time is at the expense of accuracy and variability. Introduction of the SITA algorithm resulted in a 50% reduction in test time. Thresholds are 1 to 2 dB higher than with the use of the full staircase ( full threshold) method. The shortening of test time is likely the result of a combi­nation of less patient fatigue and interruption of the stair­case procedure before threshold is reached. The algorithm has been so successful in cutting down test time that full threshold testing has largely been replaced. However, the efficacy of SITA for serial examinations to detect change remains untested. The Ring Test ( High- Pass Resolution Perimetry) The first commercially available computer graphics perimeter, the ring test, is designed to measure visual acuity throughout the visual field ( 35- 38). It uses high- pass fil­tered stimuli of various sizes ( Fig. 5), termed vanishing op­totypes. A key property of this stimulus type is that the space- average luminance of the stimulus is equivalent to the background. This stimulus has a special property: the threshold for detection is approximately the same as the threshold for resolution. Fifty test locations are used, and the test continues until the smallest ring- shaped stimulus at each test location is detected using a staircase procedure. There are several pioneering features of this method. Using computer graphics, the subject obtains feedback by way of messages and graphics. Following a ring presenta­tion, each legitimate response is acknowledged by a black square displayed on the screen at the tested location. This 49 JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 Wall FIGURE 5. Ring ( high- pass resolution) perimetry. With this type of perimetry, the smallest size stimulus that the subject can see is determined for each test location. The intratarget contrast of ring test stimuli is equal to the background, resulting in a vanishing optotype. The stimulus is seen ( de­tection) when it can be identified as a ring ( resolution). gives the subject immediate feedback as to whether the response is correct. In addition, the ring test was designed to be ergonomically comfortable. By using a size threshold to measure peripheral visual acuity, the results reflect reti­nal ganglion receptive field density. Thus, the results pro­vide a direct relationship to the underlying visual sy­stem anatomy. This test has been shown to be both sensitive and spe­cific in glaucoma and non- glaucomatous optic neuropathies ( 26,27,39). However, it is not more sensitive than conven­tional perimetry ( 40). The printout of results is unfamiliar to eye care professionals accustomed to looking at scoto­mas or displaced isopters, and there are few instruments in use beyond the borders of Sweden, where it was invented by Frisen ( 36). It has advantages of short test time ( 40,41) and especially low variability, allowing it to detect visual field change significantly earlier than conventional auto­mated perimetry. Frequency doubling technology perimetry ( FDT) Developed by Johnson et al.( 42,43), FDT uses the frequency doubling illusion. A 10° square contrast target with vertical bars is presented at a spatial frequency of 0.25 cycles/ degree ( Fig. 6). When the stimulus is temporally modulated at 25 Hz, the number of bars appears to double. Seventeen test locations are used, covering the central 20°. The rationale behind FDT is based on the activity of the My cell, a subtype of retinal M ganglion cells respon­sible for detection of motion and flicker. The My cells are a subgroup of M cells that have nonlinear response properties and very large receptive fields. They enhance the response to transient stimuli and are thought to mediate the frequency doubling illusion. This occurs when a series of low spatial frequency bars are alternated at rates above 15 Hz and the number of bars appears to double. Subjects respond as to whether they see movement, single bars, partially doubled bars, or completely doubled bars. The sensitivity and specificity of this test is the same as that of conventional perimetry ( 42,44). By using a tem­porally modulated stimulus, the test is resistant to blur up to about 5 diopters. It has been validated as an excellent screener for glaucoma ( 43,45) and for detection of other optic neuropathies ( 44). However, it often fails in the detec­tion of hemianopic defects ( 44). This failure stems from the fact that the large stimulus lined up along the vertical me­ridian may be perceived by the intact hemifield because of scatter and microfixational shifts ( Fig. 7). The test may also fail to detect hemianopias because of the steep slopes to some of the defects. If the patient sees any part of the large 10° stimulus, a small defect can be missed. FDT 2, also termed the Humphrey Matrix, is an alter­native version of FDT. A 5° stimulus substitutes for the standard 10° stimulus, and is presented in a 24- 2 pattern test grid ( 6°- spaced grid of the central 24°). A 30- 2 pro­gram ( 6°- spaced grid of the central 30°) and a 10- 2 program ( 2°- spaced grid of the central 10°) are also available. The test time is about 5 minutes per eye, and because the stimuli are smaller with the Humphrey Matrix, hemianopias are better detected. FDT is an outstanding glaucoma screener and detects optic neuropathy as well. It is portable and robust against Spatial Frequency of 0.25 cycle/ degree Temporal Frequency of 25 Hz 40 Degrees FIGURE 6. Frequency doubling technology perimetry ( FDT). A 25 Hz alternating sinusoidal contrast- grating stimulus with a spatial frequency of 0.25 cycles per degree is used. The stimulus for FDT version 1 is a 10° diameter test patch. In version 2, it has a 5° diameter. The 17 test loca­tions of FDT version 1 are shown in the figure. The nasal half ( n) and temporal half ( t) are each composed of 8 test loca­tions. The central location ( c) completes the test series. 50 © 2004 Lippincott Williams & Wilkins What's New in Perimetry JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 CAP Grayscale CAP Grayscale mm. Bh .••<?• • I • • ! • • • • • i • • i • • Tot Dcv Pat Dev • • • • I • • • • I I I a u 1 • • • ' i • • • • • • i • • I • • • • • • • • • • • • i Tot Dev Pat Dev I • • • • • • • I • • • « • • • • • I • « • • I I I • • • • I 1 1 • • s • I • I • • • 1 • • • • a • • I a • IK • I • • I I I • • • * m FDT Tot Dev Pat Dev T<> 1 m K> FDT Tot Dev Pat Dev 3< S 7^ 1- "= Khl l= g FIGURE 7. Comparing Conven­tional Automated perimetry ( CAP) and Frequency Doubling Technol­ogy perimetry ( FDT) in two pa­tients with hemianopias. In both patients, FDT failed to accurately map the hemianopias. Tot dev = total deviation; pat dev = pat­tern deviation. blur. It has also been found to have a low variability. How­ever, it is not quite as good at detecting hemianopias as standard perimetry, and its sensitivity is similar to that of conventional perimetry. In addition, the large stimulus size may make it less sensitive for detecting small defects; with its lower variability, it may be better for detecting visual field change over time. Rarebit Perimetry ( The Rabbit Test) This is another new perimetric method introduced by Frisen ( 46). It uses high contrast, briefly- exposed micro­dots. With less than a 0.5 minimum angle of resolution, the microdots are presented 1 or 2 at a time. Test probes are initiated within a 5° circular area centered on 30 test loca­tions. Normal subjects respond to at least 96% of probes. With visual field damage there is a lower hit rate. In a pilot study, this test showed the promise of being a more sensi­tive test than conventional perimetry ( Figs. 8 and 9) ( 46). Multifocal Visual Evoked Potential A promising new method of objective perimetry is the multifocal visual evoked potential ( mfVEP). Two sys­tems are currently available: the Veris system ( 47,48) and the Opera or Accumap system ( 49). Both systems use FIGURE 8. Rarebit perimetry. From left to right, the same circu­lar test area is probed five times, using twin probes ( small circles) in different locations. The white circle indicates stimulus seen; the black circle indicates stimulus not seen. The result map ( bottom row) re­flects the cumulative percentage of probes seen in the testing area ( 46). One hundred percent correct is expected in normal subjects. Pass Number Probe Positions Sum Shown Sum Seen Hit Rate % Result Map 1 ( ^ W 2 1 50 1 r » A 2 )\{( oT S\ 4 3 75 ! i( T) 3 kC^^ 6 3 50 1 r » " 4 iyr O 8 5 63 ) ( • ) 5 rL 1 0 6 60 ! r. 51 JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 Wall o o o o o o o * ® m o ® ® m m i po o o O O ( ft (•) < § )" (•) xllv v^/ o o o B O o O o o O o O o o FIGURE 9. Rarebit ( A) and Ring test ( B) perimetry results from the OD of a patient with a temporal hemianopia. Note that the defect is more extensive with rarebit pe­rimetry ( filled- in black circles) than with the ring test ( larger rings) ( 46). patches of cortically scaled, checkerboard stimulus grids ( Fig. 10) that alternate in a pseudo- random sequence. Sixty test locations are used, and either peak- to- trough ampli­tudes or signal- to- noise ratios are calculated. These systems use a binary sequence and cross- correlation technique to extract visual evoked potential signals from discrete loca­tions. Most of the work to date has been done in glaucoma, where mfVEP defects correlate well with those of conven­tional perimetry ( 49). We have found that defects inpatients with anterior ischemic optic neuropathy and idiopathic in­tracranial hypertension also correlate well with conven­tional perimetry ( Fig. 11). However, in optic neuritis pa­tients there is usually a substantially greater defect with mfVEP than with conventional perimetry. Unfortunately, many hemianopias are missed by this technique, which is primarily designed for testing glaucoma patients. FIGURE 10. Multifocal visual evoked potential ( mfVEP) stimulus grid ( Veris method) ( 55). Notice that the stimulus size increases with eccentricity from fixation. Motion Perimetry Whereas the central few degrees of the visual field are important for perception of fine detail and color, the main task of the peripheral visual field is motion perception. Loss of motion perception should therefore better predict the ability to navigate one's environment than loss of differen­tial light sensitivity. The test is aimed at isolating a small set of retinal ganglion cells, the M cells. These cells comprise about 10% of all retinal ganglion cells and are important for perception of flickering or moving stimuli. Since there are relatively few M cells, there is less redundancy in the system ( 50). Therefore, it has been proposed that defects due to damage to this system should be found at lower degrees of damage. Moreover, temporally modulated ( motion or flicker) stimuli are robust against refractive error, eliminating a po­tential perimetric artifact and allowing portability of test­ing. Because it uses large stimuli in areas of visual damage, this type of perimetry should provide low variability ( 24). In motion perimetry, a patch of random dot motion appears on a background of static random dots ( Fig. 12). Threshold is found by changing stimulus size. Our studies show motion perimetry to be slightly more sensitive and of lower variability than conventional perimetry ( 51). For ex­ample, Figure 13 shows a glaucoma suspect with a large left optic cup. Humphrey 24- 2 results, including the total de­viation probability plots, are normal. Motion perimetry shows an inferior nasal defect in the OD. As a control for the motion perimetry method, luminance size threshold perim­etry ( 52) ( a test identical to motion perimetry except that the background is plain gray and stimuli are lighter gray circu­lar patches of different size) gives normal results. A longi­tudinal study, however, will be needed to determine its efficacy for following patients. CONCLUSION While major modifications in perimetric technology and statistical analysis have improved standard differential light sensitivity perimetry, the method most commonly used in clinical practice, it remains relatively insensitive 52 © 2004 Lippincott Williams & Wilkins What's New in Perimetry JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 Tote! Deviation n # a • i e l • i i n KM VBP Traces Right eye Amplitude deviation outside normal limits ( ASI 315) FIGURE 11. Conventional automated perimetry ( top) and mfVEP ( bottom) of the OD of a patient with anterior ischemic optic neuropathy. Note that mfVEP reasonably reproduces the defect found on conventional perimetry. m and plagued by high test- retest variability in damaged vi­sual fields. Optic nerve damage is not detected until at least 30% of the ganglion cells are damaged ( 53). Clearly, we can improve on the sensitivity of perimetry. Regarding perimet­ric variability using standard clinical testing with a size III stimulus, Spry et al. ( 54) estimate that in patients with vi­sual field damage, eight serial visual field examinations are needed to be confident that there is change over time. Newly developed tests show both improved sensitivity and substantially lower variability. It may not be long before these advances make their way into clinical practice. FIGURE 12. Motion perimetry video display. The small circle ( left) represents a motion target. A magnified view of the target ( right) shows that 50% of the dots are moving in random directions ( open circles) and 50% are moving right ( black circles). X denotes the fixation target. REFERENCES von Graefe A. Ueber die Untersuchung des Gesichtsfeldes bei am-blyopischen Affectionen. Arch Ophthalmol 1856; 2: 258- 298. Aubert H, Foerster R. Unter suchungen fiber raumsinn der retina. Albrecht Von Graefe's Arch Klin Experiment Ophthalmol 1857; 3: 1- 37. Bjerrum JP. Om en tilfoelse til saedvanllige synsfeltundersogelse samt om synfeltet ved glaukom. Nord Othth Tidsskr Kkjobenh 1889; 2: 144- 185. Ferree CE, Rand G. An illuminated perimeter with campimeter fea­tures. Am J Ophthalmol 1922; 5: 455^ 165. 53 JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 Wall o 0 0 o 0 o 0 o o 0 0 c o 0 o 0 o 0 0 0 0 c c 0 0 c 0 0 0 o 0 0 o Q(,) o 0 0 o Cf) 00 0 0 » 0 0 0 • • 0 0 0 0 0 0 0 0 0 • 0 • • 0 • 0 o » « 0 a 0 O 0 o 0 0 a • 0 0 o 0 . 0 0 HFA 24- 2 Motion Luminance Size Threshold w 0 o 0 0 0 o Q 0 \ J 0 0 0 o 0 0 0 0 o 0 0 0 0 0 0 Q .0 Q 0 0 0 0 Q o 0 0 0 o o o 0 o o p! o • 0 o 0 0 * 0 0 • • • • 0 0 • °_ o „ o 0 • • 0 o 0 - • 0 • o ° • 0 o » 0 FIGURE 13. Conventional auto­mated and motion perimetry in a patient with asymmetric optic disc cupping ( fourth row). Note that the inferior nasal nerve fiber bundle defect found with motion perim­etry ( second row) is not found with conventional automated pe­rimetry ( first row) or with lumi­nance size threshold perimetry ( third row). o c+ o o 10. 11. 12. 13. Ferree CE, Rand G, Sloan LL. Selected cases showing the advan- 14. tages of a combined tangent screen and perimeter. Arch Ophthalmol 1933; 10: 166- 184. 15. Sloan LL. Area and luminance of test object as variables in exami­nation of the visual field by projection perimetry. Vision Res 1961: 1: 121- 138. 16. Aulhorn E, Harms H. Visual Perimetry. In: Jameson D, Hurvich LM, eds. Visual psychophysics. Handbook of Sensory Physiology. Vol 4. Berlin: Springer- Verlag, 1972: 102- 145. 17. Goldmann H. Ein selbstregistrierendes Projektion skulperimeter. Ophthalmologica 1945; 109: 71- 79. Goldmann H. Grundlagen exakter Perimetrie. Ophthalmologica 1945; 109: 57- 70. 18. Lynn JR. Principles of Quantitiative Perimetry: testing and inter­preting the visual field. New York: Grune and Stratton; 1977. Fankhauser F, Koch P, Roulier A. On automation of perimetry. Al- 19. brecht Von Graefe's Arch Klin Experiment Ophthalmol 1972; 184: 126- 150. 20. Heijl A. Studies on computerized perimetry. Acta Ophthalmol Suppl ( Copenh) 1977; 132: 1^ 12. 21. Heijl A. Computer test logics for automatic perimetry. Acta Oph­thalmol ( Copenh) 1977; 55: 837- 853. 22. Krakau CE. Aspects on the design of a automated perimeter. Acta Ophthalmol ( Copenh) 1978; 56: 389- 405. Trobe JD, Acosta PC, Shuster JJ, et al. An evaluation of the accu­racy of community- based perimetry. Am J Ophthalmol 1980; 90: 654- 660. Kutzko KE, Brito CF, Wall M. Effect of instructions on conven­tional automated perimetry. Invest Ophthalmol Vis Sci 2000; 41: 2006- 2013. Keltner JL, Johnson CA, Beck RW, et al. Quality control functions of the Visual Field Reading Center ( VFRC) for the Optic Neuri­tis Treatment Trial ( ONTT). Control Clin Trials 1993; 14: 143- 159. Johnson CA, Nelson- Quigg JM. A prospective three- year study of response properties of normal subjects and patients during auto­mated perimetry. Ophthalmology 1993; 100: 269- 274. Katz J, Sommer A. Reliability indexes of automated perimetric tests. Arch Ophthalmol 1988; 106: 1252- 1254. Sanabria O, Feuer WJ, Anderson DR. Pseudo- loss of fixation in automated perimetry. Ophthalmology 1991; 98: 76- 78. Zalta AH. Lens rim artifact in automated threshold perimetry. Oph­thalmology 1989; 96: 1302- 1311. Rebolleda G, Munoz FJ, Fernandez Victorio JM, et al. Effects of 54 © 2004 Lippincott Williams & Wilkins What's New in Perimetry JNeuro- Ophthalmol, Vol. 24, No. 1, 2004 pupillary dilation on automated perimetry in glaucoma patients re­ceiving pilocarpine. Ophthalmology 1992; 99: 418- 423. 23. Bengtsson B, Olsson J, Heijl A, et al. A new generation of algo­rithms for computerized threshold perimetry, SITA. Acta Ophthal­mol Scand 1997; 75: 368- 375. 24. Wall M, Kutzko KE, Chauhan BC. Variability in patients with glau­comatous optic nerve damage is reduced using size V stimuli. Invest Ophthalmol Vis Sci 1997; 38: 426- 435. 24a. Wall M, Woodward KR, Brito CF. The effect of attention on con­ventional automated perimetry and luminance size threshold perim­etry. Invest Ophthalmol Vis Sci 2004; 45: 342- 350. 25. Wall M, Johnson CA, Kutzko KS, et al. Long- and short- term vari­ability of automated perimetry results in patients with optic neuritis and healthy subjects. Arch Ophthalmol 1998; 116: 53- 61. 26. Wall M, Lefante J, Conway M. Variability of high- pass resolution perimetry in normals and patients with idiopathic intracranial hy­pertension. Invest Ophthalmol Vis Sci 1991; 32: 3091- 3095. 27. Chauhan BC, House PH, McCormick TA, et al. Comparison of con­ventional and high- pass resolution perimetry in a prospective study of patients with glaucoma and healthy controls. Arch Ophthalmol 1999; 117: 24- 33. 28. Sample PA, Taylor JD, Martinez G, et al. Short- wavelength color visual fields in glaucoma suspects at risk. Am J Ophthalmol 1993; 115: 225- 233. 29. Drum B, Breton M, Massof R, et al. Pattern discrimination perim­etry: a new concept in visual field testing. Doc OphthalmolProc Ser 1987; 49: 433- 440. 30. Hart WM, Jr., Hartz RK, Hagen RW, et al. Color contrast perimetry. Invest Ophthalmol Vis Sci 1984; 25: 400- 413. 31. Phelps CD. Acuity perimetry and glaucoma. Trans Am Ophthalmol Soc 1984; 82: 753- 791. 32. Aulhorn E, Kost G. Noise field campimetry. A new perimetric method ( snow campimetry). In: Heijl A, ed. Perimetry Update 1988/ 89. Proc VHIth International Perimetric Society Meeting. 1989: 331- 336. 33. Kardon R, Haupert C, Thompson HS. Automated pupil perimetry. Invest Ophthalmol Vis Sci 1990; 31( suppl): 504. 34. Lachenmayr BJ, Drance SM, Chauhan BC, et al. Diffuse and local­ized glaucomatous field loss in light- sense, flicker and resolution perimetry. Graefes Arch Clin Exp Ophthalmol 1991; 229: 267- 273. 35. Frisen L. A computer graphics visual field screener using high- pass spatial frequency resolution targets and multiple feedback devices. Doc Ophthalmol Proc Ser 1987; 49: 441^ 146. 36. Frisen L. Acuity perimetry: estimation of neural channels. Int Oph­thalmol 1988; 12: 169- 174. 37. Frisen L. High- pass resolution targets in peripheral vision. Ophthal­mology 1987; 94: 1104- 1108. 38. Frisen L. High- pass resolution perimetry. A clinical review. Doc Ophthalmol 1993; 83: 1- 25. 39. Wall M. High- pass resolution perimetry in optic neuritis. Invest Ophthalmol Vis Sci 1991; 32: 2525- 2529. 40. Wall M, Conway MD, House PH, et al. Evaluation of sensitivity and specificity of spatial resolution and Humphrey automated perimetry in pseudotumor cerebri patients and normal subjects. Invest Oph­thalmol Vis Sci 1991; 32: 3306- 3312. 41. Chauhan BC, House PH. Intratest variability in conventional and high- pass resolution perimetry. Ophthalmology 1991; 98: 79- 83. 42. Casson R, James B, Rubinstein A, et al. Clinical comparison of fre­quency doubling technology perimetry and Humphrey perimetry. Brit J Ophthalmol 2001 ; 85: 360- 362. 43. Cello KE, Nelson- Quigg JM, Johnson CA. Frequency doubling technology perimetry for detection of glaucomatous visual field loss. Am J Ophthalmol 2000; 129: 314- 322. 44. Wall M, Neahring RK, Woodward KR. Sensitivity and specificity of frequency doubling perimetry in neuro- ophthalmic disorders: a comparison with conventional automated perimetry. Invest Oph­thalmol Vis Sci 2002; 43: 1277- 1283. 45. Quigley HA. Identification of glaucoma- related visual field abnor­mality with the screening protocol of frequency doubling technol­ogy. Am J Ophthalmol 1998; 125: 819- 829. 46. Frisen L. New, sensitive window on abnormal spatial vision: rarebit probing. Vision Res 2002; 42: 1931- 1939. 47. Slotnick SD, Klein SA, Carney T, et al. Using multi- stimulus VEP source localization to obtain a retinotopic map of human primary visual cortex. Clin Neurophysiol 1999; 110: 1793- 1800. 48. Hood DC, Greenstein VC. Multifocal VEP and ganglion cell dam­age: applications and limitations for the study of glaucoma. Prog Retin Eye Res 2003; 22: 201- 251. 49. Graham SL, Klistorner A, Grigg JR, et al. Objective perimetry in glaucoma: recent advances with multifocal stimuli. Surv Ophthal­mol 1999; 43( suppl): 199- 209. 50. Johnson CA. Selective versus nonselective losses in glaucoma. J Glaucoma 1994; 3( suppl): S32- S44. 51. Wall M, Brito C, Kutzko K. Motion perimetry: properties and re­sults. In: Perimetry Update 1996/ 1997, Proceedings of the 10th In­ternational Perimetric Society Meeting. 1997: 21- 33. 52. Wall M, Jennisch CJ, Munden PM. Motion perimetry identifies nerve fiber bundle- like defects in ocular hypertension. Arch Oph­thalmol 1997'; 115: 26- 33. 53. Harwerth RS, Carter- Dawson L, Shen F, et al. Ganglion cell losses underlying visual field defects from experimental glaucoma. Invest Ophthalmol Vis Sci 1999; 40: 2242- 2250. 54. Spry PG, Bates AB, Johnson CA, et al. Simulation of longitudinal threshold visual field data. Invest Ophthalmol Vis Sci 2000; 41: 2192- 2200. 55. Hood DC, Zhang X, Greenstein VC, et al. An interocular compari­son of the multifocal VEP: a possible technique for detecting local damage to the optic nerve. Invest Ophthalmol Vis Sci 2000 ; 41: 1580- 1587. 55