Simulations for FASTPAC and the standard 4-2 dB full-threshold strategy of the Humphrey Field Analyzer.

PURPOSE: This study evaluates the accuracy, reproducibility, and efficiency of FASTPAC, a new fast strategy for automated perimetry using 3-dB steps with single threshold crossing, compared to the standard 4-2 dB full threshold strategy by means of computer simulations. METHOD: An "artificial patient" module was developed to create responses to stimuli by a Monte-Carlo technique from a given probability distribution. The authors performed 10,200 simulations with threshold values ranging from 0 to 50 dB. RESULTS: Results demonstrate an 18% decrease in the number of presentations per threshold determination, which is equal to a similar reduction in testing time. For both strategies, there is a considerable influence of the starting deviation (difference between starting value and actual threshold) on threshold error (difference between estimated threshold and actual threshold): negative starting deviations lead to negative threshold errors and vice versa. This relationship is more pronounced for FASTPAC (slope 0.18 dB/dB, P < 0.0001) than for the full-threshold strategy (slope 0.13 dB/dB, P < 0.0001). In addition, fluctuations of the determined thresholds, described as the distance between the 16th and 84th percentiles of the threshold errors, increase with increasing absolute starting deviations. This is particularly true of FASTPAC: CONCLUSIONS: The influence of the starting value on the threshold determination may lead to a considerable underestimation of visual field defects, accompanied by a higher fluctuation. This is an intrinsic property of both staircase procedures. FASTPAC, however, is more affected than the standard 4-2 dB full-threshold strategy. FASTPAC, therefore, provides time reduction at the expense of accuracy and reliability.

Glaucomatous visual fields. FASTPAC versus full threshold strategy of the Humphrey Field Analyzer.

PURPOSE: This study examined how FASTPAC, a fast strategy for the Humphrey Field Analyzer (HFA), compares to the standard 4/2-dB full threshold strategy in patients with glaucoma. METHOD: The author tested one eye each of 50 patients with glaucoma (23 with only relative defects, 27 with at least one absolute defect; age range, 25 to 79 years; median age, 61.5 years) with both strategies using program 30-2 of the HFA (stimulus size III). Global visual field indices as calculated by Statpac, unweighted mean sensitivity, and examination time were compared. In addition, a point by point analysis of the sensitivity values was performed. RESULTS: Between the two strategies, no significant differences were found for indices directly related to the sensitivity values (mean sensitivity [MS], mean deviation [MD]) or for the actual thresholds. For the entire population, FASTPAC showed higher mean short-term fluctuation (SF; P = 0.018), lower mean pattern standard deviation (PSD; P = 0.006), and mean corrected PSD (CPSD; P < 0.001) values and a time reduction of more than 30% (P < 0.001) compared to the standard strategy. These differences between the strategies were found to be independent of the amount of field loss, except for PSD (P = 0.001). No age influence was present. Statistically significant linear regression (r > 0.70, P < 0.001) was found between the FASTPAC values of MS, MD, PSD, CPSD, and the corresponding values determined with the standard strategy. No correlation exists for SF (r < 0.46, P > 0.030). CONCLUSIONS: FASTPAC provides a considerable time reduction at the cost of higher threshold fluctuation. This may lead to problems in detecting relative defects or changes in differential light threshold in follow-up examinations.

Points of a normal visual field are not statistically independent.

The aim of the present study was to determine whether the sensitivity values at individual test locations in a normal visual field are statistically independent. A total of 144 eyes of 144 normal subjects were tested with the Humphrey Field Analyzer, program 30-2. For each point in the visual field the pairwise correlation of sensitivity values to all other points was calculated together with the angular distance between points. The overall 2775 correlation coefficients were plotted as a function of angular distance. With increasing distance the correlation coefficients decrease continually from 0.63 (distance 6 degrees) to approx. 0.5 (distance 30 degrees and above). The strong relationship between adjacent points that is present at up to 30 degrees is destroyed by intraindividual randomisation of the visual field data. Interindividual randomisation abolishes any correlation. Adjacent locations in a normal visual field are strongly statistically related to each other. Two patterns may be separated; part of the correlation is an intrinsic neighbourhood effect present up to 30 degrees and part is due to the observation that the sensitivity values of a specific visual field are obtained from the same subject and are thus not independent. Therefore, for the calculation of normal values, procedures have to be developed that take the relationship between neighbouring points into consideration.

Deficits in perimetric performance in patients with symptomatic human immunodeficiency virus infection or acquired immunodeficiency syndrome.

PURPOSE: We measured the perimetric performance in patients with either acquired immunodeficiency syndrome (AIDS) or human immunodeficiency virus (HIV) disease but without AIDS. METHODS: Light-difference sensitivity in the central field was measured in 74 eyes of 37 patients. The Humphrey Field Analyzer 640, program 30-2 was used. Additionally, 143 eyes of 143 normal control subjects were studied. RESULTS: Mean deviation was significantly reduced in patients with HIV disease compared with control subjects (mean +/- S.E.M., -4.30 +/- 0.52 vs -0.77 +/- 0.15, respectively; P < .0001). Analysis of subgroups demonstrated that patients with lymphadenopathy syndrome or AIDS-related complex (N = 40 eyes; -3.52 +/- 0.41; P < .0001) as well as patients with AIDS (N = 34 eyes; -5.23 +/- 0.97; P < .0001) had a reduced mean deviation. Those comparisons remained significant (P < .0001) when data were analyzed independently for the right eyes and for the left eyes. Corrected pattern standard deviation (3.15 +/- 0.30 vs 1.39 +/- 0.09; P < .0001) was higher in patients with HIV disease compared with control subjects. Again, analysis of subgroups disclosed a significant increase in patients with lymphadenopathy syndrome or AIDS-related complex (2.55 +/- 0.36; P < .0001) as well as in patients with AIDS (3.85 +/- 0.51; P < .0001). Both comparisons remained significant when data were analyzed independently for the right and left eyes. CONCLUSIONS: This study demonstrates visual dysfunction despite normal visual acuity in patients with HIV disease. Our results are consistent with the hypothesis of damage at the neuroretinal level.

The different effects of aging on normal sensitivity in flicker and light-sense perimetry.

PURPOSE: To verify whether or not an accelerated loss at an older age for normal sensitivity in the central visual field is present when using the stimulus configuration of conventional white/white automated light-sense perimetry and the stimulus configuration of the automated flicker perimeter developed by one of the authors (BJL). METHODS: One hundred thirty eyes of 130 normal subjects aged 9 to 86 years were tested with the Humphrey-Field-Analyzer 640, program 30-2, and our automated flicker perimeter. In addition, short introductory learning programs were used for both techniques. All tests were performed in random order. RESULTS: Mean critical flicker fusion frequency shows a linear loss over the entire age range (r = -0.5546, P < 0.0001, slope a = -0.3820 dB/decade), whereas mean light difference sensitivity decreases only slightly up to 46 years of age (r = -0.0118, P = 0.9226, slope a = -0.0153 dB/decade), with a marked acceleration above 46 years of age (r = -0.7304, P < 0.0001, slope a = -2.0640 dB/decade). CONCLUSIONS: The absence of an accelerated loss at an older age for critical flicker fusion frequency (CFF) and the presence of such a loss for light-difference sensitivity (LDS) might be attributed to the independence of a flickering stimulus from distributing effects induced by the ocular media at an older age as proposed by one of the authors. The different age effects for CFF and LDS could also be explained by different age-related losses at different sites and for different neuronal populations throughout the visual pathways.

The role of temporal threshold criteria in psychophysical testing in glaucoma.

The magnicellular system seems to be affected early and specifically by glaucomatous damage. It shows high temporal resolution and good motion perception. Psychophysical tests investigating the temporal transfer characteristics of the afferent visual system thus are of special interest for the detection of early glaucomatous damage. Flicker perimetry is largely independent from disturbances of the ocular media and provides a means to identify eyes with pressure-produced functional loss. Defects in temporal modulation perimetry may precede the onset or progression of defects in conventional light-sense perimetry. Motion sensitivity seems to be affected early by glaucomatous damage. The problem of noisefield campimetry is the lack of quantification and reliability control of the patients' responses. Large-field or Ganzfeld stimulators can separate glaucomatous from normal eyes with high sensitivity and specificity; their value for the identification of eyes at risk to develop progression of functional loss still must be demonstrated.

Influence of stimulus size and contrast on the temporal parameters of saccadic eye movements: implications for road traffic.

Saccadic eye movements are required for the recognition of peripheral objects in road traffic. Their latencies largely determine reaction time in emergency situations. The aim of this study was to investigate the influence of stimulus and surround parameters on the temporal characteristics of saccadic eye movements under conditions of object size, contrast and luminance corresponding to nighttime traffic. Square stimuli of 1 degree or 5 degrees size were presented under an eccentricity of 5 degrees and 15 degrees. The luminance of the surround was 4 x 10(-4) cd/m2 and 1 cd/m2. More than approx. 2,000 saccades of 7 normal subjects were registered and evaluated with respect to the latency, maximal velocity, and frequency of secondary saccades. At high stimulus contrast, latency approaches a minimum of approx. 200 ms. Latency increases with decreasing contrast up to maximal values of more than 600 ms. Transforming the contrast values into a relative decibel scale shows that this increase in latency occurs at significantly higher relative contrast values under scotopic as compared with photopic conditions. Our results demonstrate that an overall latency of 200-300 ms is not adequate for the assessment of accidents, especially under the stimulus conditions of nighttime traffic.

[Visual acuity perimetry. Detection of glaucomatous visual field defects]. / Sehschärfeperimetrie. Erfassung glaukomatöser Gesichtsfelddefekte.

50 eyes of 50 patients with different types of glaucoma (27 eyes with POAG, 14 eyes with glaucoma due to pseudoexfoliation, 8 eyes with chronic narrow-angle glaucoma, 1 eye with normal-tension glaucoma) were examined with light-sense perimetry (Humphrey-Field-Analyzer, program 30-2) and resolution perimetry according to Frisén [2-7]. As the field area tested by Frisén's ring perimeter is smaller than that of the HFA, the comparison was restricted to the area of the ring perimeter, so that 50 out of 77 test locations of the HFA were included. In addition to a quantitative comparison with statistical criteria based on age-corrected normal values, a subjective qualitative assessment was established. The quantitative comparison of all hemifields showed good agreement in 61% of eyes, the ring perimeter indicating moderately and markedly increased numbers of defects in 13% and 4% of eyes and the HFA in 18% and 4% of eyes, respectively. Qualitative comparison of the entire fields revealed good agreement in 46% of eyes, with moderately and markedly more defects in 18% and 10% for the ring perimeter and 12% and 14% for the HFA, respectively. The average time needed for examination was 6.0 +/- 1.1 min for the ring perimeter and 17.2 +/- 2.7 min for the HFA. The Global Deviation of the ring perimeter correlates linearly with the Mean Deviation of the HFA (r = 0.4824, P < 0.001).

Flicker perimetry resists retinal image degradation.

The influence of refractive defocus and artificial media opacities on perimetric thresholds in automated light-sense and flicker perimetry was investigated in 20 eyes of 20 normal subjects. Thresholds were determined at 13 locations in the central visual field up to 25 degrees. Refractive defocus was induced by blurring with glasses of +1, +2, +3, +6 and +9 diopters spherical. Three diffusers were used as artificial media opacities, causing a mean reduction of visual acuity to 0.46, 0.08, and 0.02. Blurring of the retinal image by a small defocus or by slight artificial media opacities causes a measurable reduction of light-difference sensitivity. Mean sensitivity (MS) and defocus are related logarithmically (log(MS)/defocus, r = -0.9297; P < 0.0001). The correlation between MS and the luminance factor l15, characterizing the artificial media opacities, is linear (MS/l15, r = -0.9736; P < 0.0001). Flicker fusion frequency resists retinal image degradation much better. Mean flicker frequency (MF) and defocus are related logarithmically (log(MF)/defocus, r = -0.4960; P < 0.0001). The correlation between mean flicker frequency (MF) and the luminance factor l15 is nonlinear (MF/[l15]2, r = 0.8693; P < 0.0001). The results of the present study show that perimetric methods that use temporal threshold criteria, such as flicker fusion frequency, should be more suitable than methods that use static criteria for detecting neuronal damage in the presence of factors that disturb retinal image quality.

Central function and visual field damage in glaucoma.

82 eyes of 82 patients with different types of glaucoma were examined with various psychophysical tests assessing central and paracentral function, including foveal temporal contrast sensitivity function, FM 100-Hue test, and foveal and parafoveal blue-on-yellow-sensitivity. For all eyes visual field tests were performed with the Humphrey-Field-Analyzer, program 30-2. Global visual field indices were calculated as follows: Mean Sensitivity MS, Mean Deviation MD, and Corrected Pattern Standard Deviation CPSD. Linear regression analysis and multiple regression analysis correcting for a possible influence of age between the central and paracentral criteria and the global indices was performed. For the entire study population highly significant correlations are present between foveal and parafoveal blue-on-yellow-sensitivity and MS, MD and CPSD. Whereas in the Normal Tension Glaucoma subgroup (19/82 eyes) no significant correlations are found, the subgroup of 35/82 eyes with markedly elevated intraocular pressure (> or = 30 mmHg) shows highly statistically significant correlations between the low- and high-frequency end of the foveal temporal contrast sensitivity function and foveal and parafoveal blue-on-yellow-sensitivity and the global field indices. The results of the present study support the idea that there are two different mechanisms of glaucomatous damage, one which is pressure-dependent and one which may be pressure-independent. The pressure-dependent mechanism is responsible for deficits of central or paracentral function which are correlated to overall visual field damage.

[Glaucoma parameters of the papilla. Comparison of clinical and stereophotography methods with laser scanning tomography]. / Glaukomparameter an der Papille. Vergleich klinischer und stereophotographischer Methoden mit der Laser-Scanning-Tomographie.

The observations recorded by laser tomographic scanner (LTS) and with clinical and stereophotographic evaluation of 40 optic discs in the same glaucomatous patients were compared. Quantitative parameters (vertical and horizontal cup-to-disc ratio) and qualitative parameters (nasal excavation, notch, pseudo-pit, overpass, bayonetting) were evaluated. The correlation coefficients for the quantitative parameter ranged from 0.59 to 0.73. The definition of excavation used meant that the values found with the LTS were consistently higher. The qualitative parameter did not differ significantly with the method. Therefore, the LTS can be considered equivalent to the standard methods.

[Light perception and flicker perimetry. Effect of refractive error, artificial media opacities and pupillary size]. / Lichtsinn- und Flimmerperimetrie. Einfluss von Fehlrefraktion, artifiziellen Medientrübungen und Pupillenweite.

The influence of defocus, artificial media opacities and pupil size on perimetric thresholds in automated light sense and flicker perimetry was investigated in 20 eyes of 20 normal subjects. Thresholds were determined at 13 locations in the central visual field. Blurring the retinal image by a small defocus or by slight artificial media opacities causes a measurable reduction in light-difference sensitivity. Flicker fusion frequency, however, is much more resistant to degradation of the retinal image. Artificial pupil size has a similar effect on both light-difference sensitivity and flicker fusion frequency. The present study shows that perimetric methods using temporal threshold criteria should be more suitable for the detection of neuronal damage in the presence of factors disturbing the quality of the retinal image than methods using static criteria.

Diffuse field loss and diffuse retinal nerve-fiber loss in glaucoma.

A total of 61 eyes of 61 patients were examined by light-sense, flicker and resolution perimetry. All fields were classified as being normal or showing diffuse loss, localized defects or diffuse plus localized loss. For the assessment of diffuse field loss, cumulative defect curves were derived. Retinal nerve-fiber layer (RNFL) loss was assessed by a semiquantitative scoring technique. In light-sense and flicker perimetry, the total diffuse and total overall RNFL scores for eyes with diffuse field loss were approximately twice those obtained for eyes without diffuse field loss. These differences were highly statistically significant (P = 0.0055 and 0.0087 for light-sense perimetry, P = 0.0033 and 0.0042 for flicker perimetry). In resolution perimetry, eyes with diffuse field loss also showed higher RNFL scores; the differences, however, were not statistically significant. The results show that diffuse field loss in light-sense and flicker perimetry is accompanied by diffuse nerve-fiber loss.

The selective effects of elevated intraocular pressure on temporal resolution.

Twenty-seven eyes of 27 patients with normal tension glaucoma (NTG), 68 eyes of 68 patients with primary open-angle glaucoma (POAG), and 11 eyes of 11 patients with special forms of high-tension glaucoma (SHTG) including (pigmentary glaucoma, angle-closure glaucoma, secondary glaucoma, etc.), were examined by automated light-sense and temporal resolution perimetry. Light-sense perimetry was performed with the Humphrey Field Analyzer, using program 30-2. Temporal resolution perimetry was conducted with a system developed by Lachenmayr. Global field indices were calculated for all visual fields and modalities: mean defect (MD) for light-sense perimetry and mean flicker defect (FD) for flicker perimetry. For each of the three glaucoma groups there was a significant linear correlation of FD over MD. The slope of the regression line (regression coefficient a) increased with increasing maximal pressure level of the glaucomatous group: for NTG a = 0.6671, for POAG a = -1.2413, and for SHTG a = -2.235. The differences of the regression coefficients between NTG and POAG and between NTG and SHTG were statistically significant (P = 0.0366 and P = 0.0046, respectively). The results of the present study provide evidence that the relative amount of damage to flicker compared to light-sense perimetry increases with increasing maximal pressure level of a glaucomatous population. Thus, flicker perimetry may be a tool for the identification of eyes with pressure-produced damage.

Diffuse field loss and central visual function in glaucoma.

Eighty-two eyes of 82 patients with different types of glaucoma were examined with program 30-2 of the Humphrey Field Analyzer. Forty-six of 82 eyes (56.1%) had diffuse field loss and 36/82 (43.9%) eyes had no diffuse field loss according to the cumulative defect curves. Various foveal threshold criteria including temporal transfer and color vision were tested in all eyes. All criteria showed a loss of sensitivity in the presence of diffuse field loss. The loss of sensitivity was statistically significant for the high-frequency end of the foveal temporal contrast-sensitivity function (P = 0.0498), for foveal flicker-fusion frequency (P = 0.0275) and for foveal and parafoveal blue-on-yellow sensitivity (fovea: P = 0.0009; 4 parafoveal points: P = 0.0001; 16 points in the central 10 degrees: P < 0.0001). The loss in sensitivity was not statistically significant for the low and intermediate temporal frequencies of the foveal temporal contrast-sensitivity function and for the FM 100-Hue loss score. All temporal threshold criteria showed a low sensitivity (25-37%) and a low negative predictive value (48-50%) for the presence of diffuse field loss. Their specificity (75-94%) and positive predictive value (65-83%) are high, however. Thus, whenever one of the foveal temporal threshold criteria as used in the present study is abnormal, the probability is very high that there is some diffuse field loss.

Multiple-stimulus presentation and voice control in automated perimetry.

In all, 55 eyes of 55 patients were examined prospectively in random order with the Humphrey field analyzer [central field 76 points, full threshold strategy, single-stimulus presentation, response-button control (HFA 1); central field 76 points, defect-depth strategy, response-button control (HFA 2)] and the Dicon TKS-4000 [central field 76 points, defect-depth strategy, multiple-stimulus presentation, response-button control (DIC 1); central field 76 points, defect-depth strategy, multiple-stimulus presentation, voice control (DIC 2)]. Some 26 patients (47%) had glaucomatous field defects, 7 patients (13%) had lesions of the visual pathway, 5 patients (9%) had normal fields. The other 17 patients (31%) had field defects due to vascular lesions of the retina or the optic nerve, retrobulbar neuritis, cataract, dysthyroid optic neuropathy, disorders of the macula, or human immunodeficiency virus (HIV) retinopathy. The mean testing time for the whole study population was 5.2 +/- 2.7 min for DIC 1. The difference from the mean testing time for HFA 2 (6.4 +/- 2.7 min) is statistically significant (p = 0.0013). DIC 2 reduces the mean testing time to 4.9 +/- 2.6 min. The difference from DIC 1 is not statistically significant (p = 0.8110). A multiple-stimulus presentation and voice control seem to be useful methods to reduce the testing time in automated perimetry without a loss of accuracy. Voice control, as used in the DICON TKS 4000, still has to be improved, however.

Correlation of retinal nerve-fiber-layer loss, changes at the optic nerve head and various psychophysical criteria in glaucoma.

In 61 eyes of 61 patients with glaucoma, semiquantitative assessment of retinal nerve-fiber-layer (RNFL) loss and neuroretinal rim measurement of the optic nerve head by means of the Optic Nerve Head Analyzer were correlated to the outcomes of automated light-sense, flicker and resolution perimetry and the Farnsworth-Munsell (FM) 100-Hue test. A significant influence of age on total RNFL and total diffuse RNFL scores was found, but there was no measurable effect of age on neuroretinal rim area. Total RNFL and total diffuse RNFL scores showed a good correlation to the various visual field indices: total RNFL score vs mean flicker frequency as determined by flicker perimetry, r = -0.606, P less than 0.0001; total RNFL score vs mean sensitivity as determined by light-sense perimetry, r = -0.385, P = 0.002; and total RNFL score vs mean ring score as determined by resolution perimetry, r = 0.341, P = 0.007. There was no significant correlation between RNFL scores and the FM 100-Hue score. Correlation between the neuroretinal rim area and the various psychophysical indices was poor and mostly not statistically significant. The high correlation of flicker scores with RNFL loss provides interest for future applications of this perimetric technique.

Light-sense, flicker and resolution perimetry in glaucoma: a comparative study.

A total of 106 eyes of 106 patients with different types of glaucoma were examined by automated light-sense, flicker and resolution perimetry (Humphrey Field Analyzer, program 30-2; flicker perimeter as described by Lachenmayr [16, 18]; resolution perimeter as devised by Frisén [4, 6, 8-11]). The fields were classified in a masked fashion as being normal or as having purely diffuse loss, purely localized loss or diffuse as well as localized loss. As compared with light-sense perimetry, resolution perimetry had a markedly lower sensitivity in the detection of glaucomatous damage (77%) but a high specificity (93%); the comparison of resolution perimetry with flicker perimetry showed similar results (sensitivity, 75%; specificity, 85%). When flicker perimetry was compared with light-sense perimetry and vice versa, the sensitivity was high (95% and 94%, respectively), but the specificity was low (57% and 62%, respectively). The prevalence of detection of diffuse loss by both light-sense and resolution perimetry was related to visual acuity, whereas flicker perimetry did not show such a relationship.

Diffuse and localized glaucomatous field loss in light-sense, flicker and resolution perimetry.

A total of 75 eyes in 75 patients with different types of glaucoma (21 eyes with low-tension glaucoma, 49 with primary open-angle glaucoma and 5 with pigmentary glaucoma) were examined by automated light-sense, flicker and resolution perimetry. All fields were classified in a masked fashion as being "normal" (N) or as having "diffuse loss" (D), "localized loss" (L) or "diffuse plus localized loss" (DL). The frequency distributions for the various field loss categories were plotted against the highest intraocular pressure ever reported in the patients' records. The frequency distribution for the purely localized defects showed a peak at 20 mmHg and were markedly skewed to low pressure values, whereas those for both diffuse plus localized damage and purely diffuse loss peaked at about 30 mmHg. The data suggest that diffuse field loss may be an indicator of pressure-induced damage.

[Comparative study of various perimetry strategies]. / Vergleichsstudie verschiedener perimetrischer Strategien.

The conventional 4/2 dB bracketing strategy with fixed steps was compared with the "dynamic step unit strategy" (DSU strategy), which is characterized by an increase of step size with defect depth. In order to analyze the intraindividual variability, 6 eyes of 3 normal subjects were tested ten times with both strategies using the PERISTAT 433. The reproducibility of the DSU strategy is better by a factor of approximately 1.1 The comparative examination of 40 eyes of 40 patients with visual field defects resulting from various pathologic conditions using both the Humphrey-Field-Analyzer (program 30-2, 4/2 dB bracketing strategy) and the Peristat 433 (DSU strategy) yielded a clear advantage (by a factor of 3) in time and a lower number of exposures for the DSU strategy. In order to analyze the reproducibility, 15 eyes of 15 patients were tested three times with both methods. The reproducibility of the DSU strategy for relative defects is worse by a factor of 1.6. Overall, the DSU strategy is substantially faster than the conventional 4/2 dB bracketing strategy; this is achieved, however, at the expense of accuracy in relative defects.