Contrast Sensitivity Changes Due to Glaucoma and Normal Aging: Low-Spatial-Frequency Losses in Both Magnocellular and Parvocellular Pathways

Abstract

purpose. To explore the effects of glaucoma and aging on low-spatial-frequency contrast sensitivity by using tests designed to assess performance of either the magnocellular (M) or parvocellular (P) visual pathways.

methods. Contrast sensitivity was measured for spatial frequencies of 0.25 to 2 cyc/deg by using a published steady- and pulsed-pedestal approach. Sixteen patients with glaucoma and 16 approximately age-matched control subjects participated. Patients with glaucoma were tested foveally and at two midperipheral locations: (1) an area of early visual field loss, and (2) an area of normal visual field. Control subjects were assessed in matched locations. An additional group of 12 younger control subjects (aged 20–35 years) were also tested.

results. Older control subjects demonstrated reduced sensitivity relative to the younger group for the steady (presumed M)- and pulsed (presumed P)-pedestal conditions. Sensitivity was reduced foveally and in the midperiphery across the spatial frequency range. In the area of early visual field loss, the glaucoma group demonstrated further sensitivity reduction relative to older control subjects across the spatial frequency range for both the steady- and pulsed-pedestal tasks. Sensitivity was also reduced in the midperipheral location of “normal” visual field for the pulsed condition.

conclusions. Normal aging results in a reduction of contrast sensitivity for the low-spatial-frequency–sensitive components of both the M and P pathways. Glaucoma results in a further reduction of sensitivity that is not selective for M or P function. The low-spatial-frequency–sensitive channels of both pathways, which are presumably mediated by cells with larger receptive fields, are approximately equivalently impaired in early glaucoma.

There is histologic evidence for a reduction in the proportion of larger retinal ganglion cells (RGCs) in glaucomatous eyes, both in human ¹ and in experimental primate models of glaucoma. ^{2 3} Although many types of RGC have been identified, there are three major classes of retinal ganglion cells that are currently well understood: namely, the parvocellular (P), magnocellular (M), and koniocellular (K) RGCs (for review, see Ref. ⁴ ). These cell types can be distinguished both functionally and morphologically. ⁴ In terms of morphology, parvocellular RGCs are on average smaller than M or K neurons; hence, one possible inference from histologic studies of glaucomatous eyes is that people with glaucoma should show dysfunction on tasks mediated by M or K neurons but relative sparing of visual tasks dominated by P input. This theory has been widely posited. 

More recent histologic work demonstrates that the theory of selective large cell loss may be more complicated than initially thought. Several laboratories provide evidence for cell shrinkage (both of cell soma and dendritic tree) before cell death in experimental glaucoma. ^{5 6 7} Morgan et al. ⁵ classified primate RGCs as parasol (presumed M) or midget (presumed P) and found that the ratio of parasol to midget cells was relatively preserved in glaucomatous eyes, although the total number of cells was reduced. The same study found that ocular hypertension resulted in a significant reduction in cell soma size of both parasol and midget cells. Cell shrinkage would result, on average, in a reduction in cell size in glaucomatous eyes that might be misinterpreted as a loss of larger cells. ^{5 8} Weber et al., ⁶ using a primate model, and Shou et al., ⁷ using a feline model, also found evidence for cell soma and dendrite shrinkage, but both of these groups found relatively greater changes in parasol (alternately classified as Y neurons in cats) cells than in midget cells (X neurons in cats). Hence, these studies ^{6 7} suggest that both larger and smaller cell types shrink before death in glaucoma but that larger cells may be more susceptible to this process. 

An alternate histologic approach has explored the effects of glaucoma on the lateral geniculate nucleus (LGN) which receives direct projections from the RGCs. ^{9 10 11} Yucel et al. ^{10 11} have demonstrated approximately equivalent loss of the M and P layers of the LGN in experimental primate glaucoma; however Chaturvedi et al. ⁹ found a greater loss of neurons from M layers in autopsied human LGN. Crawford et al. ¹² studied metabolic activity of neurons in the LGN of primates with experimental glaucoma and found similar reductions in cytochrome oxidase reactivity in P and M layers. 

Because of the suggestion of greater neuronal loss of M and K visual pathways in glaucoma, clinical studies exploring the utility of selective functional assessment of these pathways are numerous. ^{13 14 15 16 17 18} Deficits have also been identified for tasks exploring presumed P performance, such as red-green chromatic tasks, ^{19 20} high-resolution perimetry, ²¹ and tasks exploring the phenomenon of spatial aliasing. ²² In contrast, Karwatsky et al. ²³ did not find evidence of red-green chromatic deficits in glaucoma (presumed P). As the procedures used to assess M and P performance often differ markedly in stimulus composition, retinal locations tested, and task complexity (for example, frequency-doubling perimetry [FDP] versus high-pass resolution perimetry), it can be difficult to compare the relative magnitude of deficits in these functions across studies and subjects. Several psychophysical studies in which the investigators have endeavored to minimize such differences have shown comparable loss of M and P function. ^{24 25 26}  

Psychophysically, M function is typically assessed with low-spatial-frequency stimuli that either flicker or move (for example, in FDP). Conversely, to differentiate maximally between the two pathways, the P pathway is usually assessed with higher-spatial-frequency static stimuli that are often chromatic. Although not normally assessed psychophysically, the P pathway is also capable of transmitting low-spatial-frequency information, ²⁷ with approximately 20% of achromatic LGN P cells having optimal spatial frequencies of <5 cyc/deg. ²⁸ Leonova et al. ²⁹ recently described a method for assessing contrast sensitivity that enables separate measurement of achromatic P and M contrast sensitivity functions at low spatial frequencies. The described technique displays the test stimulus on a luminance pedestal and biases detection to one or the other pathways through the use of differing interstimulus adaptation. ²⁹ A key advantage of using this technique is that the M and P pathways can be assessed using the same test stimulus: only the adaptation phase differs. This technique has been used to study M and P contrast sensitivity deficits in retinitis pigmentosa ³⁰ and melanoma-associated retinopathy. ³¹  

In this study, we applied the technique of Leonova et al. ²⁹ to the study of visual dysfunction in glaucoma, to explore for differential contrast sensitivity reductions in M and P pathways at low spatial frequencies. Although contrast sensitivity functions represent the upper envelope of mechanisms contributing to performance, it is presumed that the sensitivity for lower spatial frequencies is mediated by, on average, cells with larger receptive fields than those responsible for higher spatial frequency contrast sensitivity. We were interested in determining whether there was a spatial-frequency–dependent loss of contrast sensitivity in a group of patients with glaucoma relative to control subjects and in comparing the magnitude of loss in the M and P pathways when presumably the largest cells of each subtype were assessed. Although the main purpose of this study was to assess visual changes due to glaucoma, we also compared the performance of younger and older control subjects, to explore the effects of normal aging on the low-spatial-frequency end of the M and P contrast sensitivity functions. 

Methods

Subjects

In this research, we principally studied 16 subjects with primary open-angle glaucoma (POAG) and 16 control subjects of similar age. In the glaucoma group, subjects ranged in age from 52 to 87 years (mean, 72 ± 10 years), and in the control group, subjects ranged between 53 and 81 years (mean, 68 ± 7 years) of age. There was no significant difference in mean age between these groups (t ₍₃₀₎ = −1.32; P = 0.20). An additional group of 12 young adult control subjects was also included, ranging in age between 20 and 35 years (mean, 26 ± 5 years). Subjects with glaucoma were recruited either from the ophthalmology clinic of one of the authors (MJW) or from the Glaucoma Clinic of the Melbourne Optometric Clinic (Victorian College of Optometry, affiliated with the University of Melbourne). Control subjects were recruited from the Melbourne Optometric Clinic. 

Subjects with glaucoma were required to have a clinical diagnosis of POAG with a repeatable glaucomatous visual field loss. The visual field loss was documented on perimeters (the Medmont perimeter; Medmont Pty. Ltd., Camberwell, Australia; and the Humphrey Field Analyzer; Carl Zeiss Meditec, Dublin, CA). Visual fields were classified according to the staging system of Mills et al., ³² and all subjects with glaucoma were classified as having early loss. 

Both control subjects and those with glaucoma were required to have best corrected visual acuity of 6/7.5 or better, to be free from systemic disease known to affect visual function, and to have refractive errors in the range of ±5 D sphere and ±2 D of cylinder. Subjects with glaucoma were required to be free from other ocular disease. Control subjects were required to have normal findings in a comprehensive eye examination (including slit lamp biomicroscopy and ophthalmoscopy of the macula and optic nerve) and intraocular pressure <21 mm Hg measured with applanation tonometry. 

Before participation, all subjects provided written informed consent in accordance with a protocol approved by our institutional human research ethics committee and in accordance with the tenets of the Declaration of Helsinki. 

Stimuli and Equipment

The stimuli were similar to those used by Alexander et al. ^{30 31} and are illustrated in Figure 1 . The test stimuli were Gabor patches (sine wave presented in a Gaussian envelope of SD 2.66°) and were presented in the center of an 8° square luminance pedestal of 12.5 cd/m². The background was 25 cd/m² and subtended 38.3° × 30.5° of visual angle. 

Stimuli were generated with a commercial system (ViSaGe system; Cambridge Research Systems, Ltd., Kent, UK) and presented on a γ-corrected 21-in. monitor (frame rate: 100 Hz; G520 Trinitron; Sony, Tokyo, Japan). A chin and forehead rest was used to view the monitor, and subjects were refractively corrected for the 50-cm viewing distance. Subject responses were collected with a button box (CB6; Cambridge Research Systems). 

The steady- and pulsed-pedestal conditions are illustrated in Figure 1 . For the steady condition (Fig. 1a) , the 12.5 cd/m² luminance pedestal was presented continuously. The test stimulus was presented for 30 ms against the luminance pedestal, after which the adapting pedestal remained alone. For the pulsed-pedestal condition (Fig. 1b) , both the luminance pedestal and the test stimulus were presented simultaneously for 30 ms during the test period. The adaptation interval between stimulus presentations was 1.5 seconds. Detailed exploration of these techniques by Leonova et al. ²⁹ demonstrated that the steady-pedestal condition favors the M pathway. Steady adaptation to the pedestal alters the response and gain of units that are stimulated by the prolonged pedestal presentation (predominantly P-cells), whereas the M-cells respond to the brief stimulus presentation. The pulsed-pedestal condition favors the P pathway as the abrupt onset of the luminance pedestal saturates the M pathway. ²⁹ The contrast response properties of these tasks have been shown to be consistent with those described for primate M and P pathways. ²⁹  

Subjects were required to adapt to the background (and pedestal for the steady condition) for 1 minute before commencing each test run. In each trial, the test stimulus was randomly chosen to be oriented at either 45° or 135°. The subject was required to identify the orientation of the test stimulus (a two-alternative, forced choice). Three correct responses in a row resulted in a 20% decrease in the contrast of the Gabor, whereas every incorrect response resulted in a 20% increase in contrast. The contrast (C) of the Gabor was defined as  

\[C{=}\ \frac{L_{\mathrm{peak}}{-}L_{\mathrm{pedestal}}}{L_{\mathrm{pedestal}}},\]

where L _peak is the maximum luminance of the Gabor and L _pedestal is the luminance of the pedestal. This three-down, one-up staircase strategy converges approximately on the 79% correct response level. ³³ Each staircase was terminated after six reversals, with the staircase result being calculated as the mean of the last four reversals. 

Subjects were tested foveally and at two midperipheral test locations. At each location, contrast sensitivity was measured for the following spatial frequencies: 0.25, 0.5, 1, and 2 cyc/deg. Higher spatial frequencies were not tested, as previous studies show little separation between performance on the steady and pulsed-pedestal conditions at 4 cyc/deg and beyond. ^{29 30 31} The midperipheral locations were placed so that the corner of the pedestal square closest to the fovea was located at 10° on a 45° diagonal line from the fovea. For the glaucomatous observers, the peripheral locations were chosen such that one was placed in an area of reduced visual field sensitivity, and the other was placed in an area of normal visual field. The visual field quadrant was defined as normal if locations were classified as within normal limits on the total deviation plot of their most recently measured visual field. To be classified as abnormal, the total deviation plot had to have two or more points in the area of interest flagged at P < 0.05. An example is shown in Figure 2 . Older control subjects were also tested foveally and at two peripheral locations, where the peripheral quadrants were chosen to match those of the glaucoma group. The younger control subjects were tested foveally and at a single peripheral location, matched to one of the quadrants tested for the older control subjects. 

A separate test run was conducted for each spatial frequency and pedestal condition (steady or pulsed). Within each run, the three test locations were tested in an interleaved fashion, and two staircases were completed for each location. Only the locations were interleaved, not the pedestal conditions or the spatial frequencies. The final contrast sensitivity estimate for each spatial frequency at each location was determined as the mean of the results of the two staircases. For each subject, all measures were made within a single test session of approximately 1.5 hours’ duration, with rest breaks allowed as required. 

Results

Comparison of the Performance of Older and Younger Control Subjects

Figure 3shows the foveal performance for younger control subjects (Fig. 3a)and older control subjects (Fig. 3b) . Mean data (±SE) is shown for both steady (solid symbols) and pulsed (open symbols) conditions. Older observers performed more poorly than the younger cohort for both steady (F_(1,26) = 30.1, P < 0.001)- and pulsed (F_(1,26) = 28.3, P < 0.001)-pedestal conditions. For both pedestal subtypes, there was a statistically significant interaction between spatial frequency and group (steady: F_(3,78) = 5.82, P < 0.001; pulsed: F_(3,78) = 2.73, P = 0.05). 

Figure 3also confirms that at the lowest spatial frequencies there was marked separation between the thresholds obtained for the steady and pulsed conditions for both younger and older groups. This finding is consistent with those in previous reports, hence validating our experimental setup and supports the notion that different mechanisms (M and P) govern performance for the two pedestal conditions. ²⁹ As our older control subjects (and glaucoma group) were more elderly than previous subjects studied with this task, ^{29 30 31} it was important to demonstrate separation between the steady and pulsed thresholds in the older cohort. Separation in thresholds decreased with increasing spatial frequency and was less in older than in younger control subjects at the higher spatial frequencies. Indeed, there was no difference between sensitivity for the steady and pulsed pedestals at 2 cyc/deg in the older control subjects, implying either that some combination of M and P processing is involved in detection for both pedestal conditions at this spatial frequency or that the sensitivity of the isolated pathways is similar. 

From the group data, it is not possible to tell whether, in individual subjects, sensitivity was consistently higher for the steady pedestal across the range of spatial frequencies tested. Figures 3c and 3dshow the difference between sensitivity in the pulsed-pedestal and the steady-pedestal conditions. In the younger control subjects, sensitivity was highest in the steady-pedestal condition, with the exception of two individuals for 2 cyc/deg stimuli. There was much higher variability in individual performance in the older group; nevertheless, in all older individuals, sensitivity was higher for the steady pedestal than the pulsed pedestal at spatial frequencies of 0.25 and 0.5 cyc/deg (and in all but two individuals, at 1 cyc/deg). 

Figure 4compares the midperipheral performance of younger and older control subjects in the same format as Figure 3 . Older control subjects were tested at two midperipheral locations. Data for the location that matched that of the younger control subjects are shown in Figure 4 . Older control subjects had significantly lower sensitivity in both the steady (F_(1,26) = 49.23, P < 0.001) and pulsed (F_(1,26) = 24.53, P < 0.001) conditions. There was a statistically significant interaction between spatial frequency and group in the pulsed condition (F_(3,78) = 7.25, P < 0.001) but not in the steady condition (F_(3,78) = 1.74, P = 0.17). Figures 4c and 4dshow that for stimuli of 0.25 and 0.5 cyc/deg, sensitivity was highest for the steady pedestal condition in most subjects, regardless of age group. However, many subjects had better sensitivity for the pulsed condition at the two higher spatial frequencies. This result was evident in both groups but was more apparent for older control subjects. 

The statistical analysis of the data presented in Figures 3 and 4revealed that the difference between the younger and older groups was dependent on spatial frequency for all conditions except the steady pedestal task viewed peripherally. For each task, individual t-tests comparing younger to older groups for each spatial frequency demonstrated statistically significant differences in group performance (defined as P < 0.0125 after Bonferroni correction for multiple comparisons) for all spatial frequencies with the exception of the 0.5 cyc/deg pulsed pedestal stimulus in the periphery. Inspection of Figure 4breveals that the nonsignificant finding at this spatial frequency arises as the peak of the curve shifts to lower spatial frequencies in older observers. 

To compare the magnitude of the difference between the older and younger subjects across conditions, effect sizes (Cohen’s d) were calculated and are presented in Table 1 . Effect size (d) was calculated as d = (μ_y − μ_o)/ς_pooled, where ς_pooled is the root mean square of the standard deviations of younger and older control groups and μ_y and μ_oare the means of the younger and older groups, respectively. ³⁴ At the lowest spatial frequency, where the presumed M and P pathways are most clearly separated, the magnitude of the difference between the younger and older groups was similar in the steady and pulsed conditions. Table 1shows the smallest difference between the groups with the 0.5-cyc/deg stimulus. 

Performance of the Glaucoma Group

Figure 5compares the group performance of the older control and glaucoma groups for the steady pedestal condition. A repeated-measures ANOVA (within factors: location, spatial frequency; between factor: group) showed that the glaucoma group performed significantly more poorly than the control subjects (F_(1,30) = 8.83, P = 0.006). There was no significant three-way interaction between location, spatial frequency, and group (F_(6,180) = 1.44, P = 0.20). The two-way interaction between location and group was significant (F_(2,60) = 6.10, P = 0.004); however, the two-way interaction between spatial frequency and group was not (F_(3,90) = 0.57, P = 0.63), implying that the difference between the groups depended on the location tested but not the spatial frequency. To analyze the performance at each visual field location, we performed a repeated-measures ANOVA for each location, with a Bonferroni-corrected P = 0.016 being considered significant. The glaucoma group performed significantly worse than the control subjects in the area of abnormal visual field (F_(1,30) = 25.61, P < 0.001), but not at the other two locations (fovea: F_(1,30) = 1.48; P = 0.23; normal peripheral location: F_(1,30) = 3.18, P = 0.09). 

Figure 6compares the performance of glaucoma and older control groups for the pulsed-pedestal condition. A repeated-measures ANOVA demonstrated that the glaucoma group had significantly reduced sensitivity relative to control subjects (F_(1,30) = 9.96, P = 0.004). There was no significant three-way interaction between location, group, and spatial frequency (F_(6,180) = 1.79, P = 0.10). The two-way interaction between location and group was significant (F_(2,60) = 6.22, P = 0.004), demonstrating that the difference between the groups varied, depending on the location. Inspection of Figure 6shows the smallest difference between groups foveally and the largest difference in the midperipheral location with abnormal visual field sensitivity. There was no significant two-way interaction between spatial frequency and group (F_(3,90) = 1.73, P = 0.17). The glaucoma group performed significantly worse (P < 0.016 after Bonferroni correction) than control subjects in both peripheral areas of visual field (abnormal area: F_(1,30) = 12.15, P = 0.002; “normal” area: F_(1,30) = 8.78, P = 0.006) but not foveally (F_(1,30) = 2.62; P = 0.12). 

Effect sizes were determined, to compare across tasks the magnitude of the difference in performance between control and glaucoma groups (Table 2) . Figures 3d and 4dshow that the pulsed- and steady-pedestal tasks only clearly measure performance for separate mechanisms for the two lowest spatial frequencies (0.25 and 0.5 cyc/deg). Hence, to compare the magnitude of the effect within the presumed M and P pathways, we determined the average effect size for the 0.25 and 0.5 cyc/deg stimuli only and present these in Table 2 . Although the interpretation of effect sizes is intended to be somewhat qualitative, d = 0.5 is considered a medium effect size, and d ≥ 0.8 is considered a large effect. ³⁴ As expected, Table 2demonstrates that the largest difference between the control and glaucoma groups was present at the location of abnormal visual field sensitivity. The effect size for the pulsed-pedestal task was more consistent in magnitude across the three locations tested. This could be interpreted as indicating that the pulsed task has some advantages in the detection of diffuse loss or in the detection of very early loss, as the magnitude of effect was moderate both in the fovea and in the area of normal peripheral field. In contrast, the steady task showed its greatest effect in the area of abnormal visual field, possibly implying that the task was better able to discriminate focal loss. This interpretation should be considered cautiously, as effect sizes are not intended to be strictly quantitatively compared and the sample size of the present study was relatively small. 

Discussion

In this study, the steady and pulsed pedestal tasks of Leonova et al. ²⁹ were used to explore the effects of aging and of glaucoma on the contrast sensitivity of presumed M and P pathways. Our data for young control subjects compare well with those reported previously, ^{29 30 31} and the separation between the sensitivity measures for the steady and pulsed conditions at low spatial frequencies supports the suggestion that the tasks are measuring the performance of separate mechanisms. 

The effect of aging on subject performance for the pedestal tasks was not easily predictable. Aging has been shown to decrease contrast sensitivity for stationary grating stimuli of medium to high spatial frequencies, in the presence of intact sensitivity for spatial frequencies <2 cyc/deg. ³⁵ With increasing temporal frequency, performance is also reduced at low spatial frequencies. ³⁶ The effect of mean luminance on contrast sensitivity differs between older and younger observers, ³⁷ and there is evidence of differences in contrast gain in older individuals for stimuli of low contrast and short duration. ³⁸ Given that the stimuli used herein exploit adaptation to assess contrast sensitivity of briefly presented stimuli, it was not straightforward to predict the effect of aging on task performance. 

Our older participants performed significantly worse than their younger counterparts on both the steady- and pulsed-pedestal tasks, with significant age-related effects being present across the spatial frequency range tested. Studies have demonstrated that optical factors associated with aging reduce contrast sensitivity for higher spatial frequencies. ³⁷ Given that none of our subjects had significant media opacity, all had excellent visual acuity, and only low spatial frequencies were tested, previous psychophysical literature would support a neural basis for our measured contrast sensitivity reduction (for review, see Ref. ³⁹ ). It is not clear whether the neural changes underlying contrast sensitivity loss with age arise at the retina or more centrally. A study of the effects of aging on primate LGN showed very few changes in spatial and temporal processing of either M or P neurons, implying a more central basis. ²⁸  

For our older control subjects, sensitivity to the steady and pulsed tasks was only separable for stimuli of 0.25 and 0.5 cyc/deg; hence, it is only at these lowest spatial frequencies that we can confidently infer that different pathways were assessed. The specific stimulus parameters (for example: test pulse duration, luminance of the pedestal, and background luminance) were chosen to match those of Alexander et al., ^{30 31} as these were demonstrated by Leonova et al. ²⁹ to enable separate assessment of the M and P pathways for spatial frequencies below approximately 4 cyc/deg. In those studies, younger subjects were assessed foveally. It is possible that different parameters would result in better separation between M and P pathway responses for older subjects and for midperipheral testing. However, optimizing the test stimulus falls outside the scope of the present study. 

An advantage of the steady- and pulsed-pedestal techniques is that measurement of the M and P pathway contributions to contrast sensitivity is enabled using identical test stimuli; only the adaptation phase differs. This minimizes the potential for relative performance differences on the two tasks to arise from nonvisual factors, such as task complexity. As a group, the glaucoma participants demonstrated reduced sensitivity for both the steady and pulsed-pedestal conditions. As the magnitude of the reduction was similar between adaptation conditions for the 0.25- and 0.5-cyc/deg stimuli, we conclude that similar levels of loss were present in both M and P pathways at these spatial frequencies. The glaucoma group also had reduced sensitivity for the 1- and 2-cyc/deg stimuli, which is likely to represent a combination of reduction of both M and P processing. In summary, we did not find evidence of selective loss of sensitivity within either pathway, nor for greater loss of sensitivity from the presumed largest neurons tested (lowest spatial frequencies); however, only a narrow range of spatial frequencies was included herein. The finding of reduced sensitivity of both M and P processing is consistent with recent psychophysical evidence that identified different types of loss in individual patients as a consequence of early glaucoma. ⁴⁰  

Clinical assessment of M pathway function (for example, frequency-doubling perimetry) has been popular not only because of the notion that larger RGCs may be lost first in glaucoma, but also because of the suggestion that assessment of pathways with reduced redundancy is more likely to manifest a deficit. ⁴¹ This theory proposes that the ability to detect loss is governed in part by the number of cells contributing to the particular percept and by the degree of retinal overlap of the same cells. ⁴¹ Our results suggest that specific measures of P-pathway function may be equally as effective as assessing M function for the detection of early glaucoma. This finding is consistent with a number of previous studies that have identified visual deficits in people with early glaucoma using tasks that are understood to be processed predominantly by the parvocellular pathways. ^{19 20 21 22 24 25 26} These findings do not contest the logic that testing pathways with reduced redundancy is a sound clinical approach in glaucoma. They do, however, suggest that probing a select sample of P-cells may be as effective as the more common approach of targeting M-cell–related function. For example, the stimulus used in this study is likely to have been detected by as few as 20% of achromatic P-cells. ²⁸ Further research is needed to determine the diagnostic accuracy of the tasks described herein; however, the research points to broader options for the clinical detection of glaucomatous visual dysfunction and possibly for monitoring its progression. 

 Supported by National Health and Medical Research Council [NHMRC] Project Grant 353567 (AMM, DRB).

 Submitted for publication October 9, 2006; revised December 6, 2006; accepted February 22, 2007.

 Disclosure: A.M. McKendrick, None; G.P. Sampson, None; M.J. Walland, None; D.R. Badcock, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Schematic of the stimulus display. (a) Steady-pedestal procedure: A pedestal square of 12.5 cd/m² was presented continuously on the 25 cd/m² background. The test stimulus (Gabor) was presented for 30 ms, followed by further adaptation to the decrement pedestal. (b) Pulsed-pedestal procedure: The adapting phase showed the 25 cd/m² background. The test interval consisted of a 30-ms simultaneous display of the decrement pedestal and the Gabor. For both procedures, the Gabor was oriented at either 45° or 135°, chosen randomly in each trial.

Figure 2.

 

Illustration of stimulus placement on a representative glaucomatous visual field (HFAII, SITA Standard). Left: sensitivity at each location (in decibels); right: total deviation probability plot. Shaded circles: test locations relative to the patient’s visual field. Each patient was assessed centrally, and at two peripheral locations. One peripheral location was chosen to be in an area of abnormal visual field (two or more points in the area of interest flagged at P < 0.05 on the total deviation plot; in this example, inferonasal), whereas the second location was chosen to be in an area of normal visual field (no locations flagged at P < 0.05 on the total deviation plot; in this example, superior-temporal).

Figure 3.

 

Comparison between younger and older control groups of contrast sensitivity for the steady- and pulsed-pedestal conditions when viewed foveally. (a, b) Mean data ± SEM; (c, d) the difference between the log contrast sensitivity in the pulsed condition and that in the steady condition. All individual data are shown.

Figure 4.

 

Comparison between younger and older control groups of contrast sensitivity for steady- and pulsed-pedestal conditions when viewed in the mid-periphery. Figuredetails are the same as described in Figure 3 .

Table 1.

 

Effect Size Statistics (Cohen’s d) Comparing the Magnitude of the Difference between the Younger and Older Control Groups for Each of the Stimulus Types and Eccentricities

Stimulus Type and Eccentricity	0.25 cyc/deg	0.5 cyc/deg	1 cyc/deg	2 cyc/deg
Steady pedestal (M) fovea	1.28	0.88	1.30	1.45
Steady pedestal (M) periphery	1.53	1.20	1.36	1.26
Pulsed pedestal (P) fovea	1.61	0.38	1.39	1.40
Pulsed pedestal (P) periphery	0.91	0.34	1.23	0.93

 The data show the effect size (Cohen’s d) for the difference between older and younger control subjects.

Figure 5.

 

Comparison of contrast sensitivity for the steady pedestal condition between the older control and glaucoma groups. Mean performance ± SEM is shown for each of the tested locations: (a) fovea and (b) normal and (c) abnormal periphery.

Figure 6.

 

Comparison of contrast sensitivity for the pulsed-pedestal condition. Data are as described in Figure 5 .

Table 2.

 

Effect Size Statistics for the Steady-Pedestal and Pulsed-Pedestal Tasks Averaged for Spatial Frequencies of 0.25 and 0.5 cyc/deg

	Effect Size
		Steady Pedestal (M)	Pulsed Pedestal (P)
Fovea	0.18	0.40
Normal periphery	0.44	0.60
Abnormal periphery	0.94	0.70

The authors thank Mitchell Anjou for assistance with recruitment procedures for participants from the Melbourne Optometric Clinic.