High-pass Resolution Perimetry in Eyes With Ocular Hypertension and Primary Open-angle Glaucoma Pamela A. S a m p l e , P h . D . , D a n i e l S. A h n , M.D., Paul C. Lee, B.S., and Robert N . Weinreb, M.D.
We analyzed the results of high-pass resolu tion perimetry on normal eyes, eyes with ocular hypertension, and eyes with primary open-angle glaucoma. The subjects were matched for age and lens density. We con trolled for refraction, visual acuity, pupil size, miotic medications, and learning effects. Under these conditions the glaucomatous eyes showed a significant reduction in overall reso lution threshold compared to both normal (P £ .01) and hypertensive eyes (P < .01). The hypertensive eyes were not significantly dif ferent from normal. Standard increment threshold fields and high-pass resolution vis ual fields had an agreement of abnormality of 67%. Resolution perimetry showed a 92% agreement with standard perimetry in the location of the defect, when the defect was apparent on both tests. These results indicat ed that resolution perimetry may be useful for diagnosis and management of primary openangle glaucoma.
AWARENESS OF THE PROBLEMS associated with standard perimetry for diagnosis and manage ment of glaucoma has recently increased. In some eyes, there is histologic evidence that a large percentage of optic nerve fibers become atrophied before functional changes are seen on the visual field.1 Additionally, other tests of peripheral visual function show marked abnor malities while standard visual fields remain normal. 24 Automated static perimetry has dis-
Accepted for publication Dec. 5, 1991. From the Department of Ophthalmology, University of California at San Diego, La Jolla, California. This study was supported in part by National Eye Institute grant EY-08208 (Dr. Sample). Reprint requests to Pamela A. Sample, Ph.D., Depart ment of Ophthalmology, 0946, University of California at San Diego, La Jolla, CA 92093-0946.
advantages that include a large test-retest vari ability in normal subjects, 5 which is even higher in patients with glaucoma, 6 and the long dura tion of this monotonous procedure with little or no feedback to subjects. High-pass resolution perimetry is a new test designed to address some of these problems. 7 While standard light-sense perimetry assesses the ability to detect luminance increments on a bright background, high-pass resolution perimetry measures resolution thresholds throughout the central 30 degrees of the visual field. The design of the targets and the thresh olding paradigm reduce the test time to approx imately six minutes. Reports have indicated low variability in normal eyes, eyes with ocular hypertension, and glaucomatous eyes, 8 and pa tient acceptance has been good.9"11 Theoretical ly, the test also has the potential to provide more information about underlying reductions in ganglion cell function. 12 The test has been evaluated with mixed re sults in eyes with glaucoma and ocular hyper tension. Wanger and Persson 13 found abnormal high-pass resolution perimetry results in a high percentage of eyes with suspected or early glau coma when compared to results in normal eyes; however, their normal subjects were on average 10 years younger. Dannheim, Abramo, and Verlohr14 compared automated light-sense pe rimetry with high-pass resolution perimetry in glaucoma. They found good agreement in the number of eyes detected as abnormal for each. In contrast, Lachenmayr 10 found high-pass res olution perimetry to be markedly less sensitive than automated light-sense or flicker perimetry in detecting glaucoma. In this study, we compared the results of high-pass resolution perimetry in eyes with ocular hypertension and primary open-angle glaucoma to those from healthy eyes matched for age and lens density. We also controlled for refraction, pupil size, use of miotic medica tions, and experience with the test procedures.
©AMERICAN JOURNAL OF OPHTHALMOLOGY 113:309-316, M A R C H , 1992
309
310
AMERICAN JOURNAL OF OPHTHALMOLOGY
Material and Methods High-pass resolution perimetry was per formed with the Ophthimus High-pass Resolu tion Perimeter, Version 2 (HighTech Vision, Malmo, Sweden), which uses a multiphase test strategy in a portable computer-controlled test display. The test has been described in detail elsewhere. 7 Briefly, high-pass resolution perim etry evaluates 50 locations within the central 30 degrees of visual field, excluding the central most 5 degrees (because of monitor limita tions). It uses high-pass spatially filtered, ringshaped targets that are equiluminous over their extent with the display background of 20 candelas/m 2 (calibrated before each session with the light meter provided by Hightech Vision and verified with a photometer). The targets consist of a light ring with dark inner and outer borders. These targets are de signed to equate thresholds for detection and resolution as nearly as possible. Fourteen dif ferent sized targets vary in angular size from the smallest target, which subtends about 0.8 degree of visual angle and is arbitrarily desig nated as size 0. The targets are spaced with approximately 0.10 log unit between neighbor ing sizes. Threshold is defined as the smallest discernible target to the nearest 0.10 log unit, and is recorded as ring size, ranging between 1 and 14. A larger ring size corresponds to a poorer threshold. Presentation time is 165 mil liseconds. One eye was selected randomly from each of 14 normal subjects, 16 subjects with ocular hypertension, and 27 subjects with primary open-angle glaucoma. The three groups were matched for age, which is known to affect the results, 15 and lens density. Ages ranged from 55 to 74 years (64.64 ± 5.92; mean ± standard deviation) for normal subjects, 57 to 73 years (65.00 ± 5.39) for glaucoma suspects, and 50 to 77 years (65.96 ± 7.80) for subjects with glau coma. To obtain an index of lens density for each eye, we used a previously described and validated procedure. 16 Values of lens density were reported in units relative to 0.00 lens density on the lens density index, the value for a clear lens. Higher values corresponded to increasing opacity of the lens. A one-unit change in the index was equivalent to a 0.1-log unit change in lens density. Subjects with eyes with a lens density index of 1.50 or more were
March, 1992
excluded from the study. Lens densities ranged from 0.37 to 1.50 log units (0.80 ± 0.41 log unit) for normal subjects, 0.32 to 1.27 log units (0.80 ± 0.28 log unit) for glaucoma suspects, and 0.00 to 1.50 log units (0.80 ± 0.44 log unit) for subjects with glaucoma. Only subjects with eyes with a best-corrected visual acuity of 2 0 / 20 or better were included. Patients using miotic medications underwent a 24-hour washout period before testing and subjects with eyes with pupils less than 3 mm in diameter were excluded. All subjects underwent a complete ophthalmologic examination including best-corrected visual acuity, slit-lamp biomicroscopy, applanation tonometry, and ophthalmoscopy. All subjects were optimally refracted for all tests. Proper refraction is necessary for accurate test results. 17 Subjects with ocular disease other than primary open-angle glaucoma, or a history of congenital color-vision loss were excluded from the study. Only subjects with intraocular pressures less than 21 mm Hg, normal optic nerve heads, normal standard visual fields, and no family history of glaucoma were included in the normal group. Glaucoma suspects all had intraocular pressures exceeding 24 mm Hg on at least two separate occasions. They had nor mal standard visual fields and normal-appear ing optic disks. Eyes with primary open-angle glaucoma exhibited glaucomatous optic nerve head abnormalities, characteristic standard vis ual field loss, and intraocular pressure exceed ing 24 mm Hg on at least two occasions. Only subjects experienced with automated visual field testing were included in this study. Each subject was given a two-minute demon stration of high-pass resolution perimetry be fore the test to reduce learning effects. These effects are reported to be small, though signifi cant, and are similar in normal and glaucoma tous eyes.18 No subject had previously under gone high-pass resolution perimetry. Standard visual fields were obtained with program 24-2, using a Humphrey Visual Field Analyzer 620. Because standard visual fields measure incre ment luminance thresholds and high-pass reso lution visual fields measure resolution thresh olds for size, absolute quantitative comparisons between them are not valid. Hence, we qualita tively compared the two fields by determining whether high-pass resolution results were ab normal in eyes with standard visual field loss and whether the location of abnormality was consistent. To do this, we eliminated the two
High-pass Resolution Perimetry in Glaucoma
Vol. 113, No. 3
points most commonly associated with the blind spot, central fixation, and the uppermost four points of the standard visual field, as they are not duplicated on resolution perimetry. The remaining 48 threshold locations on the stan dard visual field and 50 locations on the resolu tion visual field were compared. To determine abnormality on resolution pe rimetry, we used a conservative criterion of a cluster of at least three points equal to or greater than 2 standard deviations from our normal subjects grouped by age in decades. The Humphrey internal normal values were used to determine abnormality on the standard visual field with a cluster of three or more points equal to or greater than 5 dB below normal required to qualify as abnormal. Glaucomatous eyes with isolated defective points greater than 10 dB and eyes with diffuse loss, therefore, did not meet the criteria denned by this analysis. This project was approved by the Human Subjects Committee at the University of Cali fornia, San Diego. The nature of the procedures was fully explained and informed consent was obtained from each subject. Differences in mean log threshold between groups were eval uated using analysis of variance. A P value of less than .05 was considered significant.
Results Analysis of variance indicated a significant difference in mean ring threshold between the normal and glaucomatous eyes at all locations of the visual field (Fig. 1). The superior nasal quadrant (P s .01) and the overall visual field (P s .01) differed slightly more than the inferior nasal (P < .02), inferior temporal (P < .02), and the superior temporal quadrants (P < .04). The ocular hypertensive and glaucomatous eyes also differed significantly in the superior nasal (P < .01), superior temporal (P ^ .01), and overall visual field (P s .01) and slightly less in the inferior nasal (P s .04) or temporal quad rants (P s .02). A fairly consistent reduction in sensitivity was noted with eccentricity for glau comatous eyes relative to age-matched and lens-density-matched normal eyes (Fig. 1). Dif ferences in ring threshold between normal and glaucomatous eyes remained significant throughout the visual field (Fig. 2). Ring thresh old was not significantly different between nor-
SUPERIOR NASAL
311
SUPERIOR TEMPORAL
9
NORMAL SUSPECT POAG
8" 7 6 LLT N CO
C3
5
Z.
4
% i
3
en
CO LU CE X I-
z
A
B
C
INFERIOR NASAL
A
B
C
INFERIOR TEMPORAL
9
76 5-
A
B
C
A
B
C
LOCATION IN ECCENTRICITY (DEGREES)
Fig. 1 (Sample and associates). Mean ring size increases with increasing eccentricity (A = 5 to 10 degrees, B = 10 to 20 degrees, and C = 20 to 28 degrees) for each of the four visual field quadrants. Error bars denote standard error of the mean. POAG indicates subjects with primary open-angle glauco ma. mal eyes and ocular hypertensive eyes for the overall visual field or any quadrant. Twenty of 27 of the glaucomatous eyes (74%) met our criteria for abnormal standard visual fields. Of these 20 eyes, 13 (65%) were also abnormal on resolution perimetry. Two of the remaining seven eyes showed several unclustered abnormal points on resolution perimetry. Of the seven eyes that did not meet our criteria on standard perimetry, five (71%) also did not meet criteria on resolution perimetry, giving an overall agreement of 67% (18 of 27 eyes). Of the 13 eyes that were abnormal on both tests, 12 (92%) had overlapping locations of visual field defect. Of the 16 glaucoma suspect eyes, all were normal by definition on the standard visual fields and only one met the criteria for abnor mality on resolution perimetry.
312
March, 1992
AMERICAN JOURNAL OF OPHTHALMOLOGY
5 4 6 4 4 6
4 4 5
4 4 5
Fig. 2 (Sample and associates). Mean ring size with resolution perimetry for each of the 50 visual field locations. The top number in each cluster denotes the mean for normal eyes, the middle number is for glauco ma suspect eyes, and the bottom value is for glaucomatous eyes.
4 4 5
Temporal
Case Reports Case 1 A 67-year-old normal subject (Fig. 3) had no evidence of defect on standard or ring visual fields. The optic nerve head was within normal limits. Ring size increased at a normal rate with eccentricity, as did thresholds for the standard visual field. Case 2 A 69-year-old patient with glaucoma had mild changes in the visual fields (Fig. 4). The optic nerve head had an inferotemporal notch that corresponded to the depressed thresholds in the superior nasal quadrant on both standard and ring perimetry. Case 3 A 76-year-old patient with glaucoma had considerable impairment of the superior visual field (Fig. 5). Results of standard and ring pe rimetry were similar and showed significantly reduced thresholds superiorly. At several
points there was no response to the brightest stimulus on standard perimetry or to the largest ring size on ring perimetry. The optic nerve head was excavated to the rim inferiorly.
Discussion Standard light-sense perimetry gives little quantitative information about the relationship between the measured threshold values and the state of the visual system. A 5-dB loss of sensi tivity does not give a meaningful estimate of optic nerve damage. Several receptive visual fields fall under each static flash, providing a great deal of redundancy to the system. Hence, one can only speculate about the extent of damage related to a given amount of sensitivity loss. This could account for the insensitivity of standard perimetry in detecting glaucomatous damage. 1 High-pass resolution perimetry was designed to establish a proportional relation ship between the resolution threshold and the underlying ganglion cell density so that each
U
26
24
25
28 do
27
26
24
32 33
26 :S
29 29
2?
23
i l i (3D 30)
29
28
30
jf
We analyzed the results of high-pass resolu tion perimetry on normal eyes, eyes with ocular hypertension, and eyes with primary open-angle glaucoma. The subjects were matched for age and lens density. We con trolled for refraction, visual acuity, pupil size, miotic medications, and learning effects. Under these conditions the glaucomatous eyes showed a significant reduction in overall reso lution threshold compared to both normal (P £ .01) and hypertensive eyes (P < .01). The hypertensive eyes were not significantly dif ferent from normal. Standard increment threshold fields and high-pass resolution vis ual fields had an agreement of abnormality of 67%. Resolution perimetry showed a 92% agreement with standard perimetry in the location of the defect, when the defect was apparent on both tests. These results indicat ed that resolution perimetry may be useful for diagnosis and management of primary openangle glaucoma.
AWARENESS OF THE PROBLEMS associated with standard perimetry for diagnosis and manage ment of glaucoma has recently increased. In some eyes, there is histologic evidence that a large percentage of optic nerve fibers become atrophied before functional changes are seen on the visual field.1 Additionally, other tests of peripheral visual function show marked abnor malities while standard visual fields remain normal. 24 Automated static perimetry has dis-
Accepted for publication Dec. 5, 1991. From the Department of Ophthalmology, University of California at San Diego, La Jolla, California. This study was supported in part by National Eye Institute grant EY-08208 (Dr. Sample). Reprint requests to Pamela A. Sample, Ph.D., Depart ment of Ophthalmology, 0946, University of California at San Diego, La Jolla, CA 92093-0946.
advantages that include a large test-retest vari ability in normal subjects, 5 which is even higher in patients with glaucoma, 6 and the long dura tion of this monotonous procedure with little or no feedback to subjects. High-pass resolution perimetry is a new test designed to address some of these problems. 7 While standard light-sense perimetry assesses the ability to detect luminance increments on a bright background, high-pass resolution perimetry measures resolution thresholds throughout the central 30 degrees of the visual field. The design of the targets and the thresh olding paradigm reduce the test time to approx imately six minutes. Reports have indicated low variability in normal eyes, eyes with ocular hypertension, and glaucomatous eyes, 8 and pa tient acceptance has been good.9"11 Theoretical ly, the test also has the potential to provide more information about underlying reductions in ganglion cell function. 12 The test has been evaluated with mixed re sults in eyes with glaucoma and ocular hyper tension. Wanger and Persson 13 found abnormal high-pass resolution perimetry results in a high percentage of eyes with suspected or early glau coma when compared to results in normal eyes; however, their normal subjects were on average 10 years younger. Dannheim, Abramo, and Verlohr14 compared automated light-sense pe rimetry with high-pass resolution perimetry in glaucoma. They found good agreement in the number of eyes detected as abnormal for each. In contrast, Lachenmayr 10 found high-pass res olution perimetry to be markedly less sensitive than automated light-sense or flicker perimetry in detecting glaucoma. In this study, we compared the results of high-pass resolution perimetry in eyes with ocular hypertension and primary open-angle glaucoma to those from healthy eyes matched for age and lens density. We also controlled for refraction, pupil size, use of miotic medica tions, and experience with the test procedures.
©AMERICAN JOURNAL OF OPHTHALMOLOGY 113:309-316, M A R C H , 1992
309
310
AMERICAN JOURNAL OF OPHTHALMOLOGY
Material and Methods High-pass resolution perimetry was per formed with the Ophthimus High-pass Resolu tion Perimeter, Version 2 (HighTech Vision, Malmo, Sweden), which uses a multiphase test strategy in a portable computer-controlled test display. The test has been described in detail elsewhere. 7 Briefly, high-pass resolution perim etry evaluates 50 locations within the central 30 degrees of visual field, excluding the central most 5 degrees (because of monitor limita tions). It uses high-pass spatially filtered, ringshaped targets that are equiluminous over their extent with the display background of 20 candelas/m 2 (calibrated before each session with the light meter provided by Hightech Vision and verified with a photometer). The targets consist of a light ring with dark inner and outer borders. These targets are de signed to equate thresholds for detection and resolution as nearly as possible. Fourteen dif ferent sized targets vary in angular size from the smallest target, which subtends about 0.8 degree of visual angle and is arbitrarily desig nated as size 0. The targets are spaced with approximately 0.10 log unit between neighbor ing sizes. Threshold is defined as the smallest discernible target to the nearest 0.10 log unit, and is recorded as ring size, ranging between 1 and 14. A larger ring size corresponds to a poorer threshold. Presentation time is 165 mil liseconds. One eye was selected randomly from each of 14 normal subjects, 16 subjects with ocular hypertension, and 27 subjects with primary open-angle glaucoma. The three groups were matched for age, which is known to affect the results, 15 and lens density. Ages ranged from 55 to 74 years (64.64 ± 5.92; mean ± standard deviation) for normal subjects, 57 to 73 years (65.00 ± 5.39) for glaucoma suspects, and 50 to 77 years (65.96 ± 7.80) for subjects with glau coma. To obtain an index of lens density for each eye, we used a previously described and validated procedure. 16 Values of lens density were reported in units relative to 0.00 lens density on the lens density index, the value for a clear lens. Higher values corresponded to increasing opacity of the lens. A one-unit change in the index was equivalent to a 0.1-log unit change in lens density. Subjects with eyes with a lens density index of 1.50 or more were
March, 1992
excluded from the study. Lens densities ranged from 0.37 to 1.50 log units (0.80 ± 0.41 log unit) for normal subjects, 0.32 to 1.27 log units (0.80 ± 0.28 log unit) for glaucoma suspects, and 0.00 to 1.50 log units (0.80 ± 0.44 log unit) for subjects with glaucoma. Only subjects with eyes with a best-corrected visual acuity of 2 0 / 20 or better were included. Patients using miotic medications underwent a 24-hour washout period before testing and subjects with eyes with pupils less than 3 mm in diameter were excluded. All subjects underwent a complete ophthalmologic examination including best-corrected visual acuity, slit-lamp biomicroscopy, applanation tonometry, and ophthalmoscopy. All subjects were optimally refracted for all tests. Proper refraction is necessary for accurate test results. 17 Subjects with ocular disease other than primary open-angle glaucoma, or a history of congenital color-vision loss were excluded from the study. Only subjects with intraocular pressures less than 21 mm Hg, normal optic nerve heads, normal standard visual fields, and no family history of glaucoma were included in the normal group. Glaucoma suspects all had intraocular pressures exceeding 24 mm Hg on at least two separate occasions. They had nor mal standard visual fields and normal-appear ing optic disks. Eyes with primary open-angle glaucoma exhibited glaucomatous optic nerve head abnormalities, characteristic standard vis ual field loss, and intraocular pressure exceed ing 24 mm Hg on at least two occasions. Only subjects experienced with automated visual field testing were included in this study. Each subject was given a two-minute demon stration of high-pass resolution perimetry be fore the test to reduce learning effects. These effects are reported to be small, though signifi cant, and are similar in normal and glaucoma tous eyes.18 No subject had previously under gone high-pass resolution perimetry. Standard visual fields were obtained with program 24-2, using a Humphrey Visual Field Analyzer 620. Because standard visual fields measure incre ment luminance thresholds and high-pass reso lution visual fields measure resolution thresh olds for size, absolute quantitative comparisons between them are not valid. Hence, we qualita tively compared the two fields by determining whether high-pass resolution results were ab normal in eyes with standard visual field loss and whether the location of abnormality was consistent. To do this, we eliminated the two
High-pass Resolution Perimetry in Glaucoma
Vol. 113, No. 3
points most commonly associated with the blind spot, central fixation, and the uppermost four points of the standard visual field, as they are not duplicated on resolution perimetry. The remaining 48 threshold locations on the stan dard visual field and 50 locations on the resolu tion visual field were compared. To determine abnormality on resolution pe rimetry, we used a conservative criterion of a cluster of at least three points equal to or greater than 2 standard deviations from our normal subjects grouped by age in decades. The Humphrey internal normal values were used to determine abnormality on the standard visual field with a cluster of three or more points equal to or greater than 5 dB below normal required to qualify as abnormal. Glaucomatous eyes with isolated defective points greater than 10 dB and eyes with diffuse loss, therefore, did not meet the criteria denned by this analysis. This project was approved by the Human Subjects Committee at the University of Cali fornia, San Diego. The nature of the procedures was fully explained and informed consent was obtained from each subject. Differences in mean log threshold between groups were eval uated using analysis of variance. A P value of less than .05 was considered significant.
Results Analysis of variance indicated a significant difference in mean ring threshold between the normal and glaucomatous eyes at all locations of the visual field (Fig. 1). The superior nasal quadrant (P s .01) and the overall visual field (P s .01) differed slightly more than the inferior nasal (P < .02), inferior temporal (P < .02), and the superior temporal quadrants (P < .04). The ocular hypertensive and glaucomatous eyes also differed significantly in the superior nasal (P < .01), superior temporal (P ^ .01), and overall visual field (P s .01) and slightly less in the inferior nasal (P s .04) or temporal quad rants (P s .02). A fairly consistent reduction in sensitivity was noted with eccentricity for glau comatous eyes relative to age-matched and lens-density-matched normal eyes (Fig. 1). Dif ferences in ring threshold between normal and glaucomatous eyes remained significant throughout the visual field (Fig. 2). Ring thresh old was not significantly different between nor-
SUPERIOR NASAL
311
SUPERIOR TEMPORAL
9
NORMAL SUSPECT POAG
8" 7 6 LLT N CO
C3
5
Z.
4
% i
3
en
CO LU CE X I-
z
A
B
C
INFERIOR NASAL
A
B
C
INFERIOR TEMPORAL
9
76 5-
A
B
C
A
B
C
LOCATION IN ECCENTRICITY (DEGREES)
Fig. 1 (Sample and associates). Mean ring size increases with increasing eccentricity (A = 5 to 10 degrees, B = 10 to 20 degrees, and C = 20 to 28 degrees) for each of the four visual field quadrants. Error bars denote standard error of the mean. POAG indicates subjects with primary open-angle glauco ma. mal eyes and ocular hypertensive eyes for the overall visual field or any quadrant. Twenty of 27 of the glaucomatous eyes (74%) met our criteria for abnormal standard visual fields. Of these 20 eyes, 13 (65%) were also abnormal on resolution perimetry. Two of the remaining seven eyes showed several unclustered abnormal points on resolution perimetry. Of the seven eyes that did not meet our criteria on standard perimetry, five (71%) also did not meet criteria on resolution perimetry, giving an overall agreement of 67% (18 of 27 eyes). Of the 13 eyes that were abnormal on both tests, 12 (92%) had overlapping locations of visual field defect. Of the 16 glaucoma suspect eyes, all were normal by definition on the standard visual fields and only one met the criteria for abnor mality on resolution perimetry.
312
March, 1992
AMERICAN JOURNAL OF OPHTHALMOLOGY
5 4 6 4 4 6
4 4 5
4 4 5
Fig. 2 (Sample and associates). Mean ring size with resolution perimetry for each of the 50 visual field locations. The top number in each cluster denotes the mean for normal eyes, the middle number is for glauco ma suspect eyes, and the bottom value is for glaucomatous eyes.
4 4 5
Temporal
Case Reports Case 1 A 67-year-old normal subject (Fig. 3) had no evidence of defect on standard or ring visual fields. The optic nerve head was within normal limits. Ring size increased at a normal rate with eccentricity, as did thresholds for the standard visual field. Case 2 A 69-year-old patient with glaucoma had mild changes in the visual fields (Fig. 4). The optic nerve head had an inferotemporal notch that corresponded to the depressed thresholds in the superior nasal quadrant on both standard and ring perimetry. Case 3 A 76-year-old patient with glaucoma had considerable impairment of the superior visual field (Fig. 5). Results of standard and ring pe rimetry were similar and showed significantly reduced thresholds superiorly. At several
points there was no response to the brightest stimulus on standard perimetry or to the largest ring size on ring perimetry. The optic nerve head was excavated to the rim inferiorly.
Discussion Standard light-sense perimetry gives little quantitative information about the relationship between the measured threshold values and the state of the visual system. A 5-dB loss of sensi tivity does not give a meaningful estimate of optic nerve damage. Several receptive visual fields fall under each static flash, providing a great deal of redundancy to the system. Hence, one can only speculate about the extent of damage related to a given amount of sensitivity loss. This could account for the insensitivity of standard perimetry in detecting glaucomatous damage. 1 High-pass resolution perimetry was designed to establish a proportional relation ship between the resolution threshold and the underlying ganglion cell density so that each
U
26
24
25
28 do
27
26
24
32 33
26 :S
29 29
2?
23
i l i (3D 30)
29
28
30
jf
