- Email: [email protected]
Rapid method for assessing rod function using recovery of spatial contrast thresholds following a bleach
Experimental Eye Research 125 (2014) 256e261
Contents lists available at ScienceDirect
Experimental Eye Research journal homepage: www.elsevier.com/locate/yexer
Rapid method for assessing rod function using recovery of spatial contrast thresholds following a bleach M. Cinta Puell a, Jeremiah M.F. Kelly b, *, Ian J. Murray b a b
n 118, Madrid 28037, Spain Applied Vision Research Group, Complutense University of Madrid, Av. Arcos de Jalo The Vision Centre, Carys Bannister Building, Faculty of Life Sciences, University of Manchester, Manchester M13 9PL, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 April 2014 Accepted in revised form 27 June 2014 Available online 9 July 2014
Poor vision in low light is a common complaint of elderly people. This poorly understood phenomenon is likely to involve both receptoral and post receptoral mechanisms. We investigated the recovery of contrast thresholds for sine-wave gratings of low spatial frequencies and low mean luminance as a function of time in darkness after photo pigment bleaching. Thirteen subjects aged 30.4 (±10.7) years took part in the study. Contrast thresholds were measured for 15 min following almost complete photo pigment bleaching. The stimuli were achromatic sinusoidal gratings of 0.5, 1 and 2 cycle per degree (cpd) generated on a CRT monitor. They had mean luminance 0.01 cd m2 and subtended 10 in diameter. The dynamics of the recovery at each spatial frequency were modelled using monophasic and biphasic exponential decay functions. The data were best modelled by a bi-phasic decay with a distinct transition point around 7 min after the bleach. Both phases followed an exponential decay. The time constant (mean, standard error) for the first phase was 0.35 (0.04) min while for the second phase it was 5.15 (0.27) min. This difference was statistically significant (p < 0.001). A control experiment revealed the second, slower phase was mediated by rod photoreceptors. Maximum contrast sensitivity was reached 15 min after a photic bleach. The dynamics of contrast sensitivity recovery follow two phases and these may be attributed to the cone and rod systems. © 2014 Elsevier Ltd. All rights reserved.
Keywords: dark adaptation recovery sensitivity contrast threshold photobleach mesopic contrast sensitivity cones rods
1. Introduction The recovery of sensitivity to a luminance target following exposure to a photo-bleach has been known for many years (Hecht et al., 1937; Lamb and Pugh, 2004; Reuter, 2011). This classical recovery function, usually plotted in terms of time following a desensitising bleach, is composed of several kinetically distinct components and is divided in to rod- and cone-mediated sections (Lamb, 1981). Many systemic conditions, for example liver disease, adversely affect the kinetics of this recovery. Sensitivity recovery is slowed in later life (Jackson et al., 1999) and has been shown to be abnormal in the presence of ocular diseases including diabetic retinopathy (Henson and North, 1979) and Age Related Macular Degeneration (Brown et al., 1986; Owsley et al., 2000; Owsley et al., 2001; Owsley et al., 2007; Dimitrov et al., 2008). The speed of dark adaptation is dependent on the regeneration of opsin to rhodopsin, the so-called retinoid cycle (Lamb and Pugh, 2004). However, it is still not clear whether rod-mediated
* Corresponding author. E-mail address: [email protected] (J.M.F. Kelly). http://dx.doi.org/10.1016/j.exer.2014.06.021 0014-4835/© 2014 Elsevier Ltd. All rights reserved.
functional deficits in AMD in particular, are primarily caused by a reduction in the number and/or sensitivity of photoreceptors, by post-receptoral abnormalities or by damage to other tissues, such as the Chorio-Bruch's RPE complex (CBRc) (Curcio et al., 1996). Feigl et al. (2007) proposed that most functional impairment in early AMD is post-receptoral, their findings were based on psychophysical and electrophysiological data. They suggested that changes in the CBRc induce a relative hypoxia in the intermediate layers of the retina. Bearing in mind this ischemia/post-receptoral hypothesis, it could be interesting to examine the recovery of the visual system to stimuli that depend on the activity of post-receptoral processing rather than the recovery of simple luminous thresholds as in the classical dark adaptation curve described above. Spatial contrast threshold detection is a good candidate for such a test, as it uses a stimulus with constant mean luminance and is mediated through post receptoral mechanisms via parvo-cellular (P) and magnocellular (M) pathways (Lee, 2011). It is thought the M pathway dominates detection of spatial patterns in the mesopic (rod- and cone-mediated) and scotopic (rod-dominated) ranges of retinal illumination. Purpura et al. (1988) used a primate model, to show that M cells, rather than P cells were sensitive to temporally modulated sine gratings at low
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
spatial frequency (0.6e1.6 cpd) when the mean retinal illumination was lower than 0.43 td which is equivalent to the low mesopic range in humans. D'Zmura and Lennie (1986) using a technique to isolate rod and cone systems, found that over most of the mesopic range, the spatial contrast sensitivity of the cone system was lower than that of the rod system at low spatial frequencies (1e3 cpd). If this is so, then it could be expected that recovery in contrast sensitivity following a bleach, would contain rod- and conemediated components having different time courses. To our knowledge there are very few studies that have used contrast threshold recovery after a photobleach to test this idea. Margrain and Thomson (1997) measured recovery of the luminance mechanism required to detect sinusoidal gratings (0.6e14 cpd) throughout the course of dark adaptation. The threshold over time for the lowest spatial frequency was qualitatively similar to a classical dark adaptation function, showing discrete rod and cone phases and for spatial frequencies above the rod spatial resolution limit (3.5 cpd), the rod phase was absent. Hahn and Geisler (1995) reported recovery of luminance sensitivity (log td) for sine-wave gratings targets (250 ms flash) ranging from 1 to 15 cpd during long-term dark adaptation following full bleaches. They found that the dark adaptation curves were similar in shape and time course throughout the course of long-term dark adaptation. Due to the experimental design, measurements were confined to the cone system. D'Zmura and Lennie, (1986) found that the recovery of contrast threshold after a bleach using a rod-isolating grating (1.38 cpd and mean illumination 9.3 td) had both cone and rod phases. In this study, we ask the question, how do the post-receptoral contrast extracting mechanisms respond to a photic bleach? The issue is important because it may shed light on why sensitivity recovery involving rods is slowed in older eyes and vulnerable to a wide range of systemic and ocular conditions. Many patients complain of poor night vision, but its investigation is regarded as excessively time consuming. For this reason there are comparatively few studies concentrating on scotopic function. One of the major advantages of the technique described here, is that it is much faster than the more conventional luminance-based approach which takes around 30e35 min. By studying the recovery of contrast thresholds for grating stimuli after photopigment bleaching, we aim to determine the temporal characteristics of post-receptoral mechanisms in the dark adaptation process.
2. Methods 2.1. Apparatus Sinusoidal gratings of 0.5, 1 and 2 cpd were generated on a calibrated, gamma-corrected high-resolution CRT monitor (Sony GDM-F500R, Tokyo, Japan). They were temporally modulated at 2 Hz. Michelson contrast ranged between 0.02 and 0.7 and was defined as
C¼
Lmax Lmin Lmax þ Lmin
The stimuli subtended 10 at the viewing distance of 75 cm as illustrated in the inset in Fig. 1. The mean screen luminance was reduced from 12.5 cd m2 to 1.3 102 cd m2 using neutral density filters (four filters were used, three of type 211 (0.9 log units) and one of type 209 (0.3log units) [LEE Filters Worldwide, Andover, Hampshire, UK]). A ViSaGe unit (CRS, UK) and a desktop PC (Dell, USA) with Windows XP operating system (Microsoft, USA) were used to control the experiment. The hardware was controlled
257
Fig. 1. The contrast sensitivity recovery function in log units (LU) as a function of time and the parameters of the bi exponential model. The parameters are explained in the text. Note the parameter j5 is the offset from the final phase to the early phase. The inset illustrates the size and location of the stimulus.
using a series of scripts written in MATLAB (2012b, The MathWorks, Inc., Natick, Mass, USA) and the psychophysics toolbox (Brainard, 1997). The scripts are available from the corresponding author. 2.2. Calibration of the monitor The calibration followed a procedure described by Parry et al. (2006). Initially the screen was auto calibrated using a Colorcal photometer (CRS Ltd, UK) and its associated software, by testing 128 voltages to obtain a gamma correction curve. The software used for stimulus generation (ViSaGe Desktop, CRS Ltd, UK) allows for correction values to compensate for intrinsic errors in the monitor for the red (R) green (G) and blue (B) phosphors and have default values of one. The CIE coordinates of x ¼ 0.31 and y ¼ 0.316 with luminance 12.5 cd m2 were entered into the software and the chromaticity coordinates at R, G and B were measured using the PR650 (Horiba UK). The voltages across the R, G and B guns were noted. The correction factors were adjusted until the monitor displayed the set luminance according to a PR1500 spot photometer (Photo Research USA). Contrast calibration was checked using a Matlab script that used a look up table to present two squares on the monitor, each with sides of 100 pixels separated by 20 pixels. The luminance of each square was measured with the PR1500 and the contrast calculated for a range of different mean luminances. The log10 (contrast) was linear for the look up table index (R2 ¼ 0.9996). 2.3. Subjects All observers had normal best-corrected visual acuity, no ocular pathology, established by an eye examination within the previous 12months. Thirteen naïve subjects were recruited (6 female) aged 30.4(22e55) yrs. They were given written and verbal information about the experiment and possible consequences of their
258
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
participation. Informed consent was obtained and the study was carried out in agreement with the tenets of the Declaration of Helsinki (World Medical Association: 1964e2008). This study was granted ethical approval by the Ethics Research Committee for the Experimentation on Humans at The University of Manchester (reference 09169). 2.4. Procedure An electronic flashgun (Nikon Speedlight SB-800, Tokyo, Japan) with energy 5.88 log scotopic troland.s, was used to produce an estimated 75e95% visual pigment bleach for subjects whose pupil diameters were in the range 4.5e6.5 mm (Rushton and Powell, 1972). Rather than have their pupils dilated, our observers sat in a darkened room for more than 5 min (whilst the kit was set up and the procedure was explained) and then, prior to the flash there was a period of total darkness of at least 30 s. This ensured adequate dilatation. Spectacles, if worn, were removed and the subject fixated the centre of the flash at a distance that ensured that the resulting after image included a region in space greater than the angular subtense of the stimulus. The subject viewed the monitor with their customary refractive correction. Immediately after the bleach, thresholds were set as follows: initially the contrast was set at 2.0 log units (LU). For each trial the subject was given an auditory cue and the stimulus presented on the screen for 340 ms then extinguished. The subject was instructed to report whether a grating was visible and if so whether the grating was vertical or horizontal. This response was entered into the computer program by the experimenter. If the subject's response was ‘absent’ or incorrect the contrast was increased by 0.15 LU and after one second the grating was again presented to the subject. If the subject correctly identified the stimulus orientation this was defined as a threshold and the contrast and time were noted, there was then a 10 s delay followed by the next presentation which had contrast 0.48 LU lower than the previous stimulus. Preliminary trials indicated that there was no reduction in threshold after 15 min, so for the main experiments, measurements were ended 15 min after the bleach. The experiment was performed over three visits on different days of the same week. Occasionally measurements were made on the same day and a washout period of at least 45 min was allowed between measurements. At each session only one spatial frequency was tested. A brief preliminary trial was performed to allow the subject to become familiar with the task. To isolate the contribution of the cone system the experiment was repeated with red sinusoidal wave gratings (0.5, 1 and 2 cpd) and mean luminance 0.01 cd m2. The software was modified so that the gratings were presented on the monitor using the red phosphor alone and the mean luminance was reset to 12.5 cd m2. In order to eliminate short wavelength energy emitted by the red phosphor, the gratings were viewed through a red filter (Lee, UK filter 026) mounted in a goggle. In this control experiment, the screen was attenuated by neutral density filters to give a mean luminance of 1 102 cd m2 as before. 2.5. Data analysis Contrast threshold recovery functions were analysed using scripts in R (R Core Team, 2012). Two models were considered for fitting the sensitivity recovery functions; a three-parameter single exponential decay and a six-parameter bi-exponential decay. The parameters were determined using nonlinear regression and are presented as mean (standard error) unless otherwise indicated.
2.6. Modelling The recovery of contrast sensitivity following a photo-bleach can be modelled by an exponential function representing a single mechanism
t CS ¼ j1 þ j2 $exp j3
(1)
Alternatively, a bi-exponential model can be fitted, allowing for the possibility of two mechanisms
t hðt; j4 Þ þ j5 $exp CS ¼ j1 þ j2 $exp j3 j6
(2)
where t ¼ time (minutes). The value of j1 is the absolute threshold log10(contrast), j2 the threshold elevation above the absolute threshold (j1 ) at time zero and j3 the time constant (minutes) of the exponential decay. If two phases are present, j4 is a putative transition point between them and the function h(t,j4 ) is zero when t < j4 and has the value tj4 for t > j4 . The first phase threshold is given by j1 þ j5, and j6 is the time constant of the second phase (minutes). Fig. 1 illustrates the parameters of the bi-exponential model using simulated data. A similar model was used by Dimitrov et al. (2008) to model cone and rod luminance sensitivity dynamics. 3. Results 3.1. Preliminary investigation of contrast sensitivity recovery In a series of pilot experiments, the recovery time course of contrast sensitivity for three spatial frequencies, 0.5, 1.0 and 2.0 cpd, following a photo-bleach in excess of 70% was measured in a female subject aged 34 years. The measurements were obtained for over twenty-five minutes. Initial examination of the data suggested a simple monophasic exponential decay and therefore a threeparameter model was fitted to the data, see Equation (1). The parameters of the model of best fit are shown in Table 1. The data are presented in Fig. 2. The model fit through the data is shown as a solid line. The paler horizontal line is the model threshold at 15 min. A linear regression of the post fifteen-minute thresholds against time found that the slope after this time was not significantly different from zero (p ¼ 0.1, 0.8, 0.5 for 0.5, 1.0 and 2.0 cpd respectively). The data confirm our initial observation that sensitivity recovery is complete after 15 min. For this reason all subsequent contrast sensitivity measurements were confined to the first 15 min following the bleach. Following the pilot experiments, contrast thresholds were measured for at least fifteen minutes in a group of thirteen healthy subjects with normal vision. Throughout the experiments observers were instructed to respond when they detected the entire grating stimulus, thus detecting contrast not luminance. We used
Table 1 Parameters of the simple exponential model fitted to the preliminary data for one observer. The absolute contrast threshold is j1 , the threshold elevation (with respect to putative absolute threshold j1 ) at time zero j2 , and the time constant of decay j3 . j1
j2
j3
Absolute threshold
Threshold elevation
Time constant
cpd
Log10(contrast)
Log10(contrast)
Min
0.5 1 2
1.68 1.73 1.55
0.99 1.09 0.88
2.61 1.73 2.73
Spatial frequency
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
259
Fig. 2. Recovery of log contrast threshold following photo bleach as a function of time in darkness for single observer. Note that recovery is complete after 15 min.
the absolute thresholds to calculate a contrast sensitivity function for all 13 observers, and used this to consider the possibility that our observers were responding only to either the dark or light bar of the stimulus by comparing maximum contrast thresholds for different spatial frequencies. The resulting contrast sensitivity function is presented in Fig. 3. The solid black lines are the medians and the horizontal lines at each end of the box plots are the 25th and 75th percentiles. The obvious low pass nature of the curve is similar to that found in many other studies, confirming the classical linear signature of rod contrast sensitivity (van Nes et al., 1967). The slope of contrast sensitivity against spatial frequency from 0.5 cpd to 2 cpd was non-zero, a least squares linear regression was found to be significant (p < 0.001). This is strong evidence that sensitivity was mediated by contrast-based, rather than luminance-based mechanisms. 3.2. Evidence for two phases of contrast sensitivity recovery Following a series of preliminary trials to allow the subjects to familiarise themselves with the task, contrast thresholds were
Fig. 3. Maximum contrast sensitivity (inverse of absolute contrast threshold) at each spatial frequency for all observers. Filled circles are potential outliers, horizontal lines the median, the grey box encloses the interquartile range (IQR) and the dashed line indicates data points within 1.5 times the IQR.
measured for at least fifteen minutes. Note that in Fig. 2 there is a hint of an inflexion at 7e8 min in the 0.5 and 2.0 cpd data. In order to determine whether this is a genuine phenomenon the recovery of contrast sensitivity measured in the thirteen healthy subjects was modelled by a bi-phasic model The pooled data are summarised in Fig. 4. For all three spatial frequencies there is clear evidence of a transition between two detection mechanisms. This occurs earlier in the recovery function for 2 cpd than the other two spatial frequencies. The solid line is a fit with a single exponential decay model and the dashed line is the best fit with a bi-phasic model (see Equation (2)). Mean and standard errors of the parameters of the biexponential decay model fitted to the binned data are shown in Table 2. Although visual inspection of the pooled data suggests that the first phase threshold (j1 þ j5 ), see Fig. 1, seems to rise with spatial frequency, no significant spatial frequency effect was found (p ¼ 0.4). In fact there was no spatial frequency effect for any parameter (all p values >0.15). To consider whether the simple exponential decay or a biexponential model best represent the results, an extra sum of squares analysis was applied to the binned data for each spatial frequency. The technique measures the marginal reduction in error sum of squares when one or more predictor variables are added to the model. This showed that the bi-exponential model was a significant improvement over the simple exponential model (F-statistics; 14.88, 33.45, 27.34, df ¼ 3 and 14, p-value <0.001). To ensure that this was not an artefact of aggregating the data, an analysis of the raw data was carried out and again found to be significant (Fstatistics; 243, 536, 447 on df ¼ 3 and 138, 133, 138 for 0.5, 1.0 and 2.0 cpd respectively; p-value <0.001). Notice the between groups degrees of freedom reflects the change in parameters between the two models, while the within group variation is due to the small variation in the number of observations for each spatial frequency. For the two lower spatial frequencies the time constants of the first section of the curve are slightly longer than for the higher spatial frequency. The inflexion point (6.37 min) is reached earlier for the higher spatial frequency. Both these observations suggest that, as expected, increasing spatial frequency leads to more cones being recruited to detect the stimulus. When contrast sensitivity recovery was considered by subject across all spatial frequencies, the time constant for the first phase was 0.35(0.04) min while for the second phase it was 5.15(0.27) min. The difference between the time constants of the two phases was subject to a paired t- test and found to be significant (t ¼ 6.30, df ¼ 38, p-value <0.001), compelling evidence that the two phases represent separate mechanisms. Hence the contrast sensitivity recovery functions at all spatial frequencies suggest the operation of two distinct mechanisms. To
260
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
Fig. 4. Contrast sensitivity recovery following photo bleach as a function of time in darkness in thirteen observers. The solid curve is the best-fit single exponential decay model. The dashed line is the best fit with the bi-exponential decay model (bi-phasic model). Filled circles are the data collected into one-minute bins; vertical bars are the interquartile range for each bin.
thresholds were 1.68, 1.73, and 1.54 log units. Note that a simple, monophasic exponential decay function is fit to the data.
test whether the rod system is responsible for the second phase the experiment was repeated using conditions designed to silent rod responses as described in the following section.
4. Discussion 3.3. Contrast sensitivity recovery using a long wavelength stimulus A simple, relatively rapid method for assessing the integrity of rod-mediated visual function, based on contrast sensitivity, is described. Like luminance detection, contrast sensitivity recovery after a desensitizing bleach, is composed of two phases, one mediated by cones and the other by rods. Contrast detection recovers to maximum sensitivity much more quickly than luminance detection; thresholds reach a minimum around 15 min after the bleach, compared with 25e35 min for luminance detection (Lamb and Pugh, 2006). The transition time from cone to rod-mediated detection is similar to that for luminance detection at around 8 min. As expected contrast sensitivity reaches a plateau rapidly because initial sensitivity is quite poor under our low luminance
In a second series of experiments, to confirm the role of rods in the second stage of the sensitivity recovery functions, an observer wore goggles fitted with Lee filter (No. 026), which attenuates all wavelengths less than 588 nm to less than 1%. Contrast sensitivity recovery functions were obtained following the bleach as previously. The results for each spatial frequency are shown in Fig. 5. The data are clearly mono-phasic. There is no evidence of a second phase of sensitivity improvement. The final thresholds were elevated by a factor of ~0.8 log units, to 0.88, 0.98 and 0.83 log units for each spatial frequency, when compared to the absolute thresholds for the same subject in Fig. 1, where the absolute
Table 2 Across subject pooled data for the parameters (mean (standard error)) of the bi-exponential model fitted to the sensitivity recovery functions. Spatial frequency (cpd)
j1
j2
j3
j4
j1 þ j5
j6
Absolute threshold
Threshold elevation
Cone time constant
Inflexion
Phase 1 threshold
Rod time constant
Min
Min
Log10(contrast)
Min
0.82 (0.07) 0.92 (0.16) 0.66 (0.04)
8.39 (0.60) 8.58 (0.25) 6.37 (0.57)
1.19 (0.13) 1.23 (0.16) 1.05 (0.12)
8.37 (0.98) 14.34 (1.68) 12.85 (2.26)
Log10(contrast) 0.5 1.0 2.0
1.60 (0.13) 1.61 (0.04) 1.45 (0.12)
1.68 (0.22) 1.73 (0.18) 1.69 (0.18)
Fig. 5. Contrast sensitivity recovery for a red stimulus is described by a simple monophasic function.
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
viewing conditions. Effectively, this means that a large part of the dynamic range of the contrast detecting mechanism has been eliminated, leaving less than 1 log unit of recovery after the inflexion point before maximum sensitivity is reached. We argue that ganglion cell mediated contrast mechanisms, rather than rod mediated luminance mechanisms, are responsible for the recovery functions. Firstly, the data in Fig. 3 show classical rod mediated contrast sensitivity characteristics. van Nes et al. (1967) showed that rod mediated contrast thresholds do not exhibit a reduction in sensitivity at low spatial frequencies. Second, they and others since have reported that contrast sensitivity is quite poor under low luminance conditions, exactly as we find in our experiments. Note that if the observers had been detecting only the light bars of the stimulus there would not have been such a substantial loss of sensitivity at the higher spatial frequency, convincing evidence of contrast detecting mechanisms at work. In humans, the macular region is composed of a small cone-rich region of diameter 1.25 , surrounded by a much larger rod dominated parafovea. It has been estimated that the macular contains around 9 as many rods as cones (Curcio et al., 1990). The human retina loses rods steadily with age; according to donor eye studies, normal older eyes lose around 30% of macula rods throughout adulthood (Curcio et al., 1996). This selective rod atrophy is not correlated with rod density which peaks at around 15 . Instead, it occurs in an annular region between 0.3 mm (c.1.1 ) and 3 mm (c.10 ). As rods are lost, the outer diameter of cones expands to fill the gaps, so that photoreceptor coverage is maintained. This raises the possibility that post receptoral factors, rather than simple rod loss, or reduced retinoid availability might account for scotopic impairment in older eyes and in ocular disease. There is no convincing evidence for an age-related decrease in the rhodopsin content of human retina, as might be expected from loss of rods, suggesting that the remaining rods are somehow compensating. Hence it is important to determine whether post-receptoral factors contribute to age related scotopic sensitivity loss. The technique used in the present experiments is ideally suited to test this possibility for three reasons. First, the stimulus is detected by post receptoral contrast mechanisms. This is evident from the fact that the data exhibit a low pass spatial frequency function; maximum sensitivity is higher for 0.5 and 1.0 cpd than for 2.0 c/deg see Fig. 3. Second, the test activates many more rods than cones (~2.7 M vs 0.4 M respectively; (Curcio et al., 1990)). Finally, because it is easy to implement and convenient for patients, it can easily be performed in clinics and on a large sample of observers. Apart from these operational advantages the method tests vision under low mesopic conditions, exactly those typically described by patients as causing difficulties. The complaint of poor night vision rarely applies to total darkness. Typically, older persons and those suffering from early AMD, for example, encounter problems negotiating dimly lit steps or stairs where illumination is low. The ability to detect objects against a dimly lit, but non-zero background is mediated by post receptoral contrast detecting mechanisms. It is these circumstances that the present technique is designed to emulate. Unsurprisingly most difficulties occur when moving from high to low luminance conditions when the slowed dynamics of sensitivity recovery are particularly evident. There is a clear need for a clinically implementable procedure for rapidly assessing scotopic visual function. The technique
261
described here, will enable the investigation of the many patients who report night vision problems that are almost certainly related to slowed dark adaptation. This includes, but is not necessarily restricted to, patients with early AMD.
Grant information Spanish Ministry of Education, ref PR2011-0203, UK College of Optometrists, Postgraduate Scholarship.
References Brainard, D.H., 1997. The psychophysics toolbox. Spat. Vis. 10 (4), 433e436. Brown, B., Adams, A.J., Coletta, N.J., Haegerstrom-Portnoy, G., 1986. Dark adaptation in age-related maculopathy. Ophthalmic Physiol. Opt. 6 (1), 81e84. Curcio, C.A., Medeiros, N.E., Millican, C.L., 1996. Photoreceptor loss in age-related macular degeneration. Investig. Ophthalmol. Vis. Sci. 37 (7), 1236e1249. Curcio, C.A., Sloan, K.R., Kalina, R.E., Hendrickson, A.E., 1990. Human photoreceptor topography. J. Comp. Neurol. 292 (4), 497e523. D'Zmura, M., Lennie, P., 1986. Shared pathways for rod and cone vision. Vis. Res. 26 (8), 1273e1280. Dimitrov, P.N., Guymer, R.H., Zele, A.J., Anderson, A.J., Vingrys, A.J., 2008. Measuring rod and cone dynamics in age-related maculopathy. Investig. Ophthalmol. Vis. Sci. 49 (1), 55e65. Feigl, B., Brown, B., Lovie-Kitchin, J., Swann, P., 2007. Functional loss in early agerelated maculopathy: the ischaemia postreceptoral hypothesis. Eye (Lond) 21 (6), 689e696. Hahn, L.W., Geisler, W.S., 1995. Adaptation mechanisms in spatial visioneI. Bleaches and backgrounds. Vis. Res. 35 (11), 1585e1594. Hecht, S., Haig, C., Chase, A.M., 1937. The influence of light adaptation on subsequent dark adaptation of the eye. J. Gen. Physiol. 20 (6), 831e850. Henson, D.B., North, R.V., 1979. Dark adaptation in diabetes mellitus. Br. J. Ophthalmol. 63 (8), 539e541. Jackson, G.R., Owsley, C., McGwin Jr., G., 1999. Aging and dark adaptation. Vis. Res. 39 (23), 3975e3982. Lamb, T.D., 1981. The involvement of rod photoreceptors in dark adaptation. Vis. Res. 21 (12), 1773e1782. Lamb, T.D., Pugh Jr., E.N., 2004. Dark adaptation and the retinoid cycle of vision. Prog. Retin. Eye Res. 23 (3), 307e380. Lamb, T.D., Pugh Jr., E.N., 2006. Phototransduction, dark adaptation, and rhodopsin regeneration the proctor lecture. Investig. Ophthalmol. Vis. Sci. 47 (12), 5137e5152. Lee, B.B., 2011. Visual pathways and psychophysical channels in the primate. J. Physiol. 589 (Pt 1), 41e47. Margrain, T.H., Thomson, W.D., 1997. Recovery of spatial vision during dark adaptation in normal subjects. Ophthalmic Physiol. Opt. 17 (6), 509e515. Owsley, C., Jackson, G.R., Cideciyan, A.V., Huang, Y., Fine, S.L., Ho, A.C., Maguire, M.G., Lolley, V., Jacobson, S.G., 2000. Psychophysical evidence for rod vulnerability in age-related macular degeneration. Investig. Ophthalmol. Vis. Sci. 41 (1), 267e273. Owsley, C., Jackson, G.R., White, M., Feist, R., Edwards, D., 2001. Delays in rodmediated dark adaptation in early age-related maculopathy. Ophthalmology 108 (7), 1196e1202. Owsley, C., McGwin Jr., G., Jackson, G.R., Kallies, K., Clark, M., 2007. Cone- and rodmediated dark adaptation impairment in age-related maculopathy. Ophthalmology 114 (9), 1728e1735. Parry, N.R., McKeefry, D.J., Murray, I.J., 2006. Variant and invariant color perception in the near peripheral retina. J. Opt. Soc. Am. A Opt. Image Sci. Vis. 23 (7), 1586e1597. Purpura, K., Kaplan, E., Shapley, R.M., 1988. Background light and the contrast gain of primate P and M retinal ganglion cells. Proc. Natl. Acad. Sci. U. S. A. 85 (12), 4534e4537. R Core Team, 2012. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Reuter, T., 2011. Fifty years of dark adaptation 1961e2011. Vis. Res. 51 (21e22), 2243e2262. Rushton, W.A., Powell, D.S., 1972. The rhodopsin content and the visual threshold of human rods. Vis. Res. 12 (6), 1073e1081. van Nes, F.L., Koenderink, J.J., Nas, H., Bouman, M.A., 1967. Spatiotemporal modulation transfer in the human eye. J. Opt. Soc. Am. 57 (9), 1082e1088.
Contents lists available at ScienceDirect
Experimental Eye Research journal homepage: www.elsevier.com/locate/yexer
Rapid method for assessing rod function using recovery of spatial contrast thresholds following a bleach M. Cinta Puell a, Jeremiah M.F. Kelly b, *, Ian J. Murray b a b
n 118, Madrid 28037, Spain Applied Vision Research Group, Complutense University of Madrid, Av. Arcos de Jalo The Vision Centre, Carys Bannister Building, Faculty of Life Sciences, University of Manchester, Manchester M13 9PL, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 April 2014 Accepted in revised form 27 June 2014 Available online 9 July 2014
Poor vision in low light is a common complaint of elderly people. This poorly understood phenomenon is likely to involve both receptoral and post receptoral mechanisms. We investigated the recovery of contrast thresholds for sine-wave gratings of low spatial frequencies and low mean luminance as a function of time in darkness after photo pigment bleaching. Thirteen subjects aged 30.4 (±10.7) years took part in the study. Contrast thresholds were measured for 15 min following almost complete photo pigment bleaching. The stimuli were achromatic sinusoidal gratings of 0.5, 1 and 2 cycle per degree (cpd) generated on a CRT monitor. They had mean luminance 0.01 cd m2 and subtended 10 in diameter. The dynamics of the recovery at each spatial frequency were modelled using monophasic and biphasic exponential decay functions. The data were best modelled by a bi-phasic decay with a distinct transition point around 7 min after the bleach. Both phases followed an exponential decay. The time constant (mean, standard error) for the first phase was 0.35 (0.04) min while for the second phase it was 5.15 (0.27) min. This difference was statistically significant (p < 0.001). A control experiment revealed the second, slower phase was mediated by rod photoreceptors. Maximum contrast sensitivity was reached 15 min after a photic bleach. The dynamics of contrast sensitivity recovery follow two phases and these may be attributed to the cone and rod systems. © 2014 Elsevier Ltd. All rights reserved.
Keywords: dark adaptation recovery sensitivity contrast threshold photobleach mesopic contrast sensitivity cones rods
1. Introduction The recovery of sensitivity to a luminance target following exposure to a photo-bleach has been known for many years (Hecht et al., 1937; Lamb and Pugh, 2004; Reuter, 2011). This classical recovery function, usually plotted in terms of time following a desensitising bleach, is composed of several kinetically distinct components and is divided in to rod- and cone-mediated sections (Lamb, 1981). Many systemic conditions, for example liver disease, adversely affect the kinetics of this recovery. Sensitivity recovery is slowed in later life (Jackson et al., 1999) and has been shown to be abnormal in the presence of ocular diseases including diabetic retinopathy (Henson and North, 1979) and Age Related Macular Degeneration (Brown et al., 1986; Owsley et al., 2000; Owsley et al., 2001; Owsley et al., 2007; Dimitrov et al., 2008). The speed of dark adaptation is dependent on the regeneration of opsin to rhodopsin, the so-called retinoid cycle (Lamb and Pugh, 2004). However, it is still not clear whether rod-mediated
* Corresponding author. E-mail address: [email protected] (J.M.F. Kelly). http://dx.doi.org/10.1016/j.exer.2014.06.021 0014-4835/© 2014 Elsevier Ltd. All rights reserved.
functional deficits in AMD in particular, are primarily caused by a reduction in the number and/or sensitivity of photoreceptors, by post-receptoral abnormalities or by damage to other tissues, such as the Chorio-Bruch's RPE complex (CBRc) (Curcio et al., 1996). Feigl et al. (2007) proposed that most functional impairment in early AMD is post-receptoral, their findings were based on psychophysical and electrophysiological data. They suggested that changes in the CBRc induce a relative hypoxia in the intermediate layers of the retina. Bearing in mind this ischemia/post-receptoral hypothesis, it could be interesting to examine the recovery of the visual system to stimuli that depend on the activity of post-receptoral processing rather than the recovery of simple luminous thresholds as in the classical dark adaptation curve described above. Spatial contrast threshold detection is a good candidate for such a test, as it uses a stimulus with constant mean luminance and is mediated through post receptoral mechanisms via parvo-cellular (P) and magnocellular (M) pathways (Lee, 2011). It is thought the M pathway dominates detection of spatial patterns in the mesopic (rod- and cone-mediated) and scotopic (rod-dominated) ranges of retinal illumination. Purpura et al. (1988) used a primate model, to show that M cells, rather than P cells were sensitive to temporally modulated sine gratings at low
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
spatial frequency (0.6e1.6 cpd) when the mean retinal illumination was lower than 0.43 td which is equivalent to the low mesopic range in humans. D'Zmura and Lennie (1986) using a technique to isolate rod and cone systems, found that over most of the mesopic range, the spatial contrast sensitivity of the cone system was lower than that of the rod system at low spatial frequencies (1e3 cpd). If this is so, then it could be expected that recovery in contrast sensitivity following a bleach, would contain rod- and conemediated components having different time courses. To our knowledge there are very few studies that have used contrast threshold recovery after a photobleach to test this idea. Margrain and Thomson (1997) measured recovery of the luminance mechanism required to detect sinusoidal gratings (0.6e14 cpd) throughout the course of dark adaptation. The threshold over time for the lowest spatial frequency was qualitatively similar to a classical dark adaptation function, showing discrete rod and cone phases and for spatial frequencies above the rod spatial resolution limit (3.5 cpd), the rod phase was absent. Hahn and Geisler (1995) reported recovery of luminance sensitivity (log td) for sine-wave gratings targets (250 ms flash) ranging from 1 to 15 cpd during long-term dark adaptation following full bleaches. They found that the dark adaptation curves were similar in shape and time course throughout the course of long-term dark adaptation. Due to the experimental design, measurements were confined to the cone system. D'Zmura and Lennie, (1986) found that the recovery of contrast threshold after a bleach using a rod-isolating grating (1.38 cpd and mean illumination 9.3 td) had both cone and rod phases. In this study, we ask the question, how do the post-receptoral contrast extracting mechanisms respond to a photic bleach? The issue is important because it may shed light on why sensitivity recovery involving rods is slowed in older eyes and vulnerable to a wide range of systemic and ocular conditions. Many patients complain of poor night vision, but its investigation is regarded as excessively time consuming. For this reason there are comparatively few studies concentrating on scotopic function. One of the major advantages of the technique described here, is that it is much faster than the more conventional luminance-based approach which takes around 30e35 min. By studying the recovery of contrast thresholds for grating stimuli after photopigment bleaching, we aim to determine the temporal characteristics of post-receptoral mechanisms in the dark adaptation process.
2. Methods 2.1. Apparatus Sinusoidal gratings of 0.5, 1 and 2 cpd were generated on a calibrated, gamma-corrected high-resolution CRT monitor (Sony GDM-F500R, Tokyo, Japan). They were temporally modulated at 2 Hz. Michelson contrast ranged between 0.02 and 0.7 and was defined as
C¼
Lmax Lmin Lmax þ Lmin
The stimuli subtended 10 at the viewing distance of 75 cm as illustrated in the inset in Fig. 1. The mean screen luminance was reduced from 12.5 cd m2 to 1.3 102 cd m2 using neutral density filters (four filters were used, three of type 211 (0.9 log units) and one of type 209 (0.3log units) [LEE Filters Worldwide, Andover, Hampshire, UK]). A ViSaGe unit (CRS, UK) and a desktop PC (Dell, USA) with Windows XP operating system (Microsoft, USA) were used to control the experiment. The hardware was controlled
257
Fig. 1. The contrast sensitivity recovery function in log units (LU) as a function of time and the parameters of the bi exponential model. The parameters are explained in the text. Note the parameter j5 is the offset from the final phase to the early phase. The inset illustrates the size and location of the stimulus.
using a series of scripts written in MATLAB (2012b, The MathWorks, Inc., Natick, Mass, USA) and the psychophysics toolbox (Brainard, 1997). The scripts are available from the corresponding author. 2.2. Calibration of the monitor The calibration followed a procedure described by Parry et al. (2006). Initially the screen was auto calibrated using a Colorcal photometer (CRS Ltd, UK) and its associated software, by testing 128 voltages to obtain a gamma correction curve. The software used for stimulus generation (ViSaGe Desktop, CRS Ltd, UK) allows for correction values to compensate for intrinsic errors in the monitor for the red (R) green (G) and blue (B) phosphors and have default values of one. The CIE coordinates of x ¼ 0.31 and y ¼ 0.316 with luminance 12.5 cd m2 were entered into the software and the chromaticity coordinates at R, G and B were measured using the PR650 (Horiba UK). The voltages across the R, G and B guns were noted. The correction factors were adjusted until the monitor displayed the set luminance according to a PR1500 spot photometer (Photo Research USA). Contrast calibration was checked using a Matlab script that used a look up table to present two squares on the monitor, each with sides of 100 pixels separated by 20 pixels. The luminance of each square was measured with the PR1500 and the contrast calculated for a range of different mean luminances. The log10 (contrast) was linear for the look up table index (R2 ¼ 0.9996). 2.3. Subjects All observers had normal best-corrected visual acuity, no ocular pathology, established by an eye examination within the previous 12months. Thirteen naïve subjects were recruited (6 female) aged 30.4(22e55) yrs. They were given written and verbal information about the experiment and possible consequences of their
258
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
participation. Informed consent was obtained and the study was carried out in agreement with the tenets of the Declaration of Helsinki (World Medical Association: 1964e2008). This study was granted ethical approval by the Ethics Research Committee for the Experimentation on Humans at The University of Manchester (reference 09169). 2.4. Procedure An electronic flashgun (Nikon Speedlight SB-800, Tokyo, Japan) with energy 5.88 log scotopic troland.s, was used to produce an estimated 75e95% visual pigment bleach for subjects whose pupil diameters were in the range 4.5e6.5 mm (Rushton and Powell, 1972). Rather than have their pupils dilated, our observers sat in a darkened room for more than 5 min (whilst the kit was set up and the procedure was explained) and then, prior to the flash there was a period of total darkness of at least 30 s. This ensured adequate dilatation. Spectacles, if worn, were removed and the subject fixated the centre of the flash at a distance that ensured that the resulting after image included a region in space greater than the angular subtense of the stimulus. The subject viewed the monitor with their customary refractive correction. Immediately after the bleach, thresholds were set as follows: initially the contrast was set at 2.0 log units (LU). For each trial the subject was given an auditory cue and the stimulus presented on the screen for 340 ms then extinguished. The subject was instructed to report whether a grating was visible and if so whether the grating was vertical or horizontal. This response was entered into the computer program by the experimenter. If the subject's response was ‘absent’ or incorrect the contrast was increased by 0.15 LU and after one second the grating was again presented to the subject. If the subject correctly identified the stimulus orientation this was defined as a threshold and the contrast and time were noted, there was then a 10 s delay followed by the next presentation which had contrast 0.48 LU lower than the previous stimulus. Preliminary trials indicated that there was no reduction in threshold after 15 min, so for the main experiments, measurements were ended 15 min after the bleach. The experiment was performed over three visits on different days of the same week. Occasionally measurements were made on the same day and a washout period of at least 45 min was allowed between measurements. At each session only one spatial frequency was tested. A brief preliminary trial was performed to allow the subject to become familiar with the task. To isolate the contribution of the cone system the experiment was repeated with red sinusoidal wave gratings (0.5, 1 and 2 cpd) and mean luminance 0.01 cd m2. The software was modified so that the gratings were presented on the monitor using the red phosphor alone and the mean luminance was reset to 12.5 cd m2. In order to eliminate short wavelength energy emitted by the red phosphor, the gratings were viewed through a red filter (Lee, UK filter 026) mounted in a goggle. In this control experiment, the screen was attenuated by neutral density filters to give a mean luminance of 1 102 cd m2 as before. 2.5. Data analysis Contrast threshold recovery functions were analysed using scripts in R (R Core Team, 2012). Two models were considered for fitting the sensitivity recovery functions; a three-parameter single exponential decay and a six-parameter bi-exponential decay. The parameters were determined using nonlinear regression and are presented as mean (standard error) unless otherwise indicated.
2.6. Modelling The recovery of contrast sensitivity following a photo-bleach can be modelled by an exponential function representing a single mechanism
t CS ¼ j1 þ j2 $exp j3
(1)
Alternatively, a bi-exponential model can be fitted, allowing for the possibility of two mechanisms
t hðt; j4 Þ þ j5 $exp CS ¼ j1 þ j2 $exp j3 j6
(2)
where t ¼ time (minutes). The value of j1 is the absolute threshold log10(contrast), j2 the threshold elevation above the absolute threshold (j1 ) at time zero and j3 the time constant (minutes) of the exponential decay. If two phases are present, j4 is a putative transition point between them and the function h(t,j4 ) is zero when t < j4 and has the value tj4 for t > j4 . The first phase threshold is given by j1 þ j5, and j6 is the time constant of the second phase (minutes). Fig. 1 illustrates the parameters of the bi-exponential model using simulated data. A similar model was used by Dimitrov et al. (2008) to model cone and rod luminance sensitivity dynamics. 3. Results 3.1. Preliminary investigation of contrast sensitivity recovery In a series of pilot experiments, the recovery time course of contrast sensitivity for three spatial frequencies, 0.5, 1.0 and 2.0 cpd, following a photo-bleach in excess of 70% was measured in a female subject aged 34 years. The measurements were obtained for over twenty-five minutes. Initial examination of the data suggested a simple monophasic exponential decay and therefore a threeparameter model was fitted to the data, see Equation (1). The parameters of the model of best fit are shown in Table 1. The data are presented in Fig. 2. The model fit through the data is shown as a solid line. The paler horizontal line is the model threshold at 15 min. A linear regression of the post fifteen-minute thresholds against time found that the slope after this time was not significantly different from zero (p ¼ 0.1, 0.8, 0.5 for 0.5, 1.0 and 2.0 cpd respectively). The data confirm our initial observation that sensitivity recovery is complete after 15 min. For this reason all subsequent contrast sensitivity measurements were confined to the first 15 min following the bleach. Following the pilot experiments, contrast thresholds were measured for at least fifteen minutes in a group of thirteen healthy subjects with normal vision. Throughout the experiments observers were instructed to respond when they detected the entire grating stimulus, thus detecting contrast not luminance. We used
Table 1 Parameters of the simple exponential model fitted to the preliminary data for one observer. The absolute contrast threshold is j1 , the threshold elevation (with respect to putative absolute threshold j1 ) at time zero j2 , and the time constant of decay j3 . j1
j2
j3
Absolute threshold
Threshold elevation
Time constant
cpd
Log10(contrast)
Log10(contrast)
Min
0.5 1 2
1.68 1.73 1.55
0.99 1.09 0.88
2.61 1.73 2.73
Spatial frequency
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
259
Fig. 2. Recovery of log contrast threshold following photo bleach as a function of time in darkness for single observer. Note that recovery is complete after 15 min.
the absolute thresholds to calculate a contrast sensitivity function for all 13 observers, and used this to consider the possibility that our observers were responding only to either the dark or light bar of the stimulus by comparing maximum contrast thresholds for different spatial frequencies. The resulting contrast sensitivity function is presented in Fig. 3. The solid black lines are the medians and the horizontal lines at each end of the box plots are the 25th and 75th percentiles. The obvious low pass nature of the curve is similar to that found in many other studies, confirming the classical linear signature of rod contrast sensitivity (van Nes et al., 1967). The slope of contrast sensitivity against spatial frequency from 0.5 cpd to 2 cpd was non-zero, a least squares linear regression was found to be significant (p < 0.001). This is strong evidence that sensitivity was mediated by contrast-based, rather than luminance-based mechanisms. 3.2. Evidence for two phases of contrast sensitivity recovery Following a series of preliminary trials to allow the subjects to familiarise themselves with the task, contrast thresholds were
Fig. 3. Maximum contrast sensitivity (inverse of absolute contrast threshold) at each spatial frequency for all observers. Filled circles are potential outliers, horizontal lines the median, the grey box encloses the interquartile range (IQR) and the dashed line indicates data points within 1.5 times the IQR.
measured for at least fifteen minutes. Note that in Fig. 2 there is a hint of an inflexion at 7e8 min in the 0.5 and 2.0 cpd data. In order to determine whether this is a genuine phenomenon the recovery of contrast sensitivity measured in the thirteen healthy subjects was modelled by a bi-phasic model The pooled data are summarised in Fig. 4. For all three spatial frequencies there is clear evidence of a transition between two detection mechanisms. This occurs earlier in the recovery function for 2 cpd than the other two spatial frequencies. The solid line is a fit with a single exponential decay model and the dashed line is the best fit with a bi-phasic model (see Equation (2)). Mean and standard errors of the parameters of the biexponential decay model fitted to the binned data are shown in Table 2. Although visual inspection of the pooled data suggests that the first phase threshold (j1 þ j5 ), see Fig. 1, seems to rise with spatial frequency, no significant spatial frequency effect was found (p ¼ 0.4). In fact there was no spatial frequency effect for any parameter (all p values >0.15). To consider whether the simple exponential decay or a biexponential model best represent the results, an extra sum of squares analysis was applied to the binned data for each spatial frequency. The technique measures the marginal reduction in error sum of squares when one or more predictor variables are added to the model. This showed that the bi-exponential model was a significant improvement over the simple exponential model (F-statistics; 14.88, 33.45, 27.34, df ¼ 3 and 14, p-value <0.001). To ensure that this was not an artefact of aggregating the data, an analysis of the raw data was carried out and again found to be significant (Fstatistics; 243, 536, 447 on df ¼ 3 and 138, 133, 138 for 0.5, 1.0 and 2.0 cpd respectively; p-value <0.001). Notice the between groups degrees of freedom reflects the change in parameters between the two models, while the within group variation is due to the small variation in the number of observations for each spatial frequency. For the two lower spatial frequencies the time constants of the first section of the curve are slightly longer than for the higher spatial frequency. The inflexion point (6.37 min) is reached earlier for the higher spatial frequency. Both these observations suggest that, as expected, increasing spatial frequency leads to more cones being recruited to detect the stimulus. When contrast sensitivity recovery was considered by subject across all spatial frequencies, the time constant for the first phase was 0.35(0.04) min while for the second phase it was 5.15(0.27) min. The difference between the time constants of the two phases was subject to a paired t- test and found to be significant (t ¼ 6.30, df ¼ 38, p-value <0.001), compelling evidence that the two phases represent separate mechanisms. Hence the contrast sensitivity recovery functions at all spatial frequencies suggest the operation of two distinct mechanisms. To
260
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
Fig. 4. Contrast sensitivity recovery following photo bleach as a function of time in darkness in thirteen observers. The solid curve is the best-fit single exponential decay model. The dashed line is the best fit with the bi-exponential decay model (bi-phasic model). Filled circles are the data collected into one-minute bins; vertical bars are the interquartile range for each bin.
thresholds were 1.68, 1.73, and 1.54 log units. Note that a simple, monophasic exponential decay function is fit to the data.
test whether the rod system is responsible for the second phase the experiment was repeated using conditions designed to silent rod responses as described in the following section.
4. Discussion 3.3. Contrast sensitivity recovery using a long wavelength stimulus A simple, relatively rapid method for assessing the integrity of rod-mediated visual function, based on contrast sensitivity, is described. Like luminance detection, contrast sensitivity recovery after a desensitizing bleach, is composed of two phases, one mediated by cones and the other by rods. Contrast detection recovers to maximum sensitivity much more quickly than luminance detection; thresholds reach a minimum around 15 min after the bleach, compared with 25e35 min for luminance detection (Lamb and Pugh, 2006). The transition time from cone to rod-mediated detection is similar to that for luminance detection at around 8 min. As expected contrast sensitivity reaches a plateau rapidly because initial sensitivity is quite poor under our low luminance
In a second series of experiments, to confirm the role of rods in the second stage of the sensitivity recovery functions, an observer wore goggles fitted with Lee filter (No. 026), which attenuates all wavelengths less than 588 nm to less than 1%. Contrast sensitivity recovery functions were obtained following the bleach as previously. The results for each spatial frequency are shown in Fig. 5. The data are clearly mono-phasic. There is no evidence of a second phase of sensitivity improvement. The final thresholds were elevated by a factor of ~0.8 log units, to 0.88, 0.98 and 0.83 log units for each spatial frequency, when compared to the absolute thresholds for the same subject in Fig. 1, where the absolute
Table 2 Across subject pooled data for the parameters (mean (standard error)) of the bi-exponential model fitted to the sensitivity recovery functions. Spatial frequency (cpd)
j1
j2
j3
j4
j1 þ j5
j6
Absolute threshold
Threshold elevation
Cone time constant
Inflexion
Phase 1 threshold
Rod time constant
Min
Min
Log10(contrast)
Min
0.82 (0.07) 0.92 (0.16) 0.66 (0.04)
8.39 (0.60) 8.58 (0.25) 6.37 (0.57)
1.19 (0.13) 1.23 (0.16) 1.05 (0.12)
8.37 (0.98) 14.34 (1.68) 12.85 (2.26)
Log10(contrast) 0.5 1.0 2.0
1.60 (0.13) 1.61 (0.04) 1.45 (0.12)
1.68 (0.22) 1.73 (0.18) 1.69 (0.18)
Fig. 5. Contrast sensitivity recovery for a red stimulus is described by a simple monophasic function.
M.C. Puell et al. / Experimental Eye Research 125 (2014) 256e261
viewing conditions. Effectively, this means that a large part of the dynamic range of the contrast detecting mechanism has been eliminated, leaving less than 1 log unit of recovery after the inflexion point before maximum sensitivity is reached. We argue that ganglion cell mediated contrast mechanisms, rather than rod mediated luminance mechanisms, are responsible for the recovery functions. Firstly, the data in Fig. 3 show classical rod mediated contrast sensitivity characteristics. van Nes et al. (1967) showed that rod mediated contrast thresholds do not exhibit a reduction in sensitivity at low spatial frequencies. Second, they and others since have reported that contrast sensitivity is quite poor under low luminance conditions, exactly as we find in our experiments. Note that if the observers had been detecting only the light bars of the stimulus there would not have been such a substantial loss of sensitivity at the higher spatial frequency, convincing evidence of contrast detecting mechanisms at work. In humans, the macular region is composed of a small cone-rich region of diameter 1.25 , surrounded by a much larger rod dominated parafovea. It has been estimated that the macular contains around 9 as many rods as cones (Curcio et al., 1990). The human retina loses rods steadily with age; according to donor eye studies, normal older eyes lose around 30% of macula rods throughout adulthood (Curcio et al., 1996). This selective rod atrophy is not correlated with rod density which peaks at around 15 . Instead, it occurs in an annular region between 0.3 mm (c.1.1 ) and 3 mm (c.10 ). As rods are lost, the outer diameter of cones expands to fill the gaps, so that photoreceptor coverage is maintained. This raises the possibility that post receptoral factors, rather than simple rod loss, or reduced retinoid availability might account for scotopic impairment in older eyes and in ocular disease. There is no convincing evidence for an age-related decrease in the rhodopsin content of human retina, as might be expected from loss of rods, suggesting that the remaining rods are somehow compensating. Hence it is important to determine whether post-receptoral factors contribute to age related scotopic sensitivity loss. The technique used in the present experiments is ideally suited to test this possibility for three reasons. First, the stimulus is detected by post receptoral contrast mechanisms. This is evident from the fact that the data exhibit a low pass spatial frequency function; maximum sensitivity is higher for 0.5 and 1.0 cpd than for 2.0 c/deg see Fig. 3. Second, the test activates many more rods than cones (~2.7 M vs 0.4 M respectively; (Curcio et al., 1990)). Finally, because it is easy to implement and convenient for patients, it can easily be performed in clinics and on a large sample of observers. Apart from these operational advantages the method tests vision under low mesopic conditions, exactly those typically described by patients as causing difficulties. The complaint of poor night vision rarely applies to total darkness. Typically, older persons and those suffering from early AMD, for example, encounter problems negotiating dimly lit steps or stairs where illumination is low. The ability to detect objects against a dimly lit, but non-zero background is mediated by post receptoral contrast detecting mechanisms. It is these circumstances that the present technique is designed to emulate. Unsurprisingly most difficulties occur when moving from high to low luminance conditions when the slowed dynamics of sensitivity recovery are particularly evident. There is a clear need for a clinically implementable procedure for rapidly assessing scotopic visual function. The technique
261
described here, will enable the investigation of the many patients who report night vision problems that are almost certainly related to slowed dark adaptation. This includes, but is not necessarily restricted to, patients with early AMD.
Grant information Spanish Ministry of Education, ref PR2011-0203, UK College of Optometrists, Postgraduate Scholarship.
References Brainard, D.H., 1997. The psychophysics toolbox. Spat. Vis. 10 (4), 433e436. Brown, B., Adams, A.J., Coletta, N.J., Haegerstrom-Portnoy, G., 1986. Dark adaptation in age-related maculopathy. Ophthalmic Physiol. Opt. 6 (1), 81e84. Curcio, C.A., Medeiros, N.E., Millican, C.L., 1996. Photoreceptor loss in age-related macular degeneration. Investig. Ophthalmol. Vis. Sci. 37 (7), 1236e1249. Curcio, C.A., Sloan, K.R., Kalina, R.E., Hendrickson, A.E., 1990. Human photoreceptor topography. J. Comp. Neurol. 292 (4), 497e523. D'Zmura, M., Lennie, P., 1986. Shared pathways for rod and cone vision. Vis. Res. 26 (8), 1273e1280. Dimitrov, P.N., Guymer, R.H., Zele, A.J., Anderson, A.J., Vingrys, A.J., 2008. Measuring rod and cone dynamics in age-related maculopathy. Investig. Ophthalmol. Vis. Sci. 49 (1), 55e65. Feigl, B., Brown, B., Lovie-Kitchin, J., Swann, P., 2007. Functional loss in early agerelated maculopathy: the ischaemia postreceptoral hypothesis. Eye (Lond) 21 (6), 689e696. Hahn, L.W., Geisler, W.S., 1995. Adaptation mechanisms in spatial visioneI. Bleaches and backgrounds. Vis. Res. 35 (11), 1585e1594. Hecht, S., Haig, C., Chase, A.M., 1937. The influence of light adaptation on subsequent dark adaptation of the eye. J. Gen. Physiol. 20 (6), 831e850. Henson, D.B., North, R.V., 1979. Dark adaptation in diabetes mellitus. Br. J. Ophthalmol. 63 (8), 539e541. Jackson, G.R., Owsley, C., McGwin Jr., G., 1999. Aging and dark adaptation. Vis. Res. 39 (23), 3975e3982. Lamb, T.D., 1981. The involvement of rod photoreceptors in dark adaptation. Vis. Res. 21 (12), 1773e1782. Lamb, T.D., Pugh Jr., E.N., 2004. Dark adaptation and the retinoid cycle of vision. Prog. Retin. Eye Res. 23 (3), 307e380. Lamb, T.D., Pugh Jr., E.N., 2006. Phototransduction, dark adaptation, and rhodopsin regeneration the proctor lecture. Investig. Ophthalmol. Vis. Sci. 47 (12), 5137e5152. Lee, B.B., 2011. Visual pathways and psychophysical channels in the primate. J. Physiol. 589 (Pt 1), 41e47. Margrain, T.H., Thomson, W.D., 1997. Recovery of spatial vision during dark adaptation in normal subjects. Ophthalmic Physiol. Opt. 17 (6), 509e515. Owsley, C., Jackson, G.R., Cideciyan, A.V., Huang, Y., Fine, S.L., Ho, A.C., Maguire, M.G., Lolley, V., Jacobson, S.G., 2000. Psychophysical evidence for rod vulnerability in age-related macular degeneration. Investig. Ophthalmol. Vis. Sci. 41 (1), 267e273. Owsley, C., Jackson, G.R., White, M., Feist, R., Edwards, D., 2001. Delays in rodmediated dark adaptation in early age-related maculopathy. Ophthalmology 108 (7), 1196e1202. Owsley, C., McGwin Jr., G., Jackson, G.R., Kallies, K., Clark, M., 2007. Cone- and rodmediated dark adaptation impairment in age-related maculopathy. Ophthalmology 114 (9), 1728e1735. Parry, N.R., McKeefry, D.J., Murray, I.J., 2006. Variant and invariant color perception in the near peripheral retina. J. Opt. Soc. Am. A Opt. Image Sci. Vis. 23 (7), 1586e1597. Purpura, K., Kaplan, E., Shapley, R.M., 1988. Background light and the contrast gain of primate P and M retinal ganglion cells. Proc. Natl. Acad. Sci. U. S. A. 85 (12), 4534e4537. R Core Team, 2012. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Reuter, T., 2011. Fifty years of dark adaptation 1961e2011. Vis. Res. 51 (21e22), 2243e2262. Rushton, W.A., Powell, D.S., 1972. The rhodopsin content and the visual threshold of human rods. Vis. Res. 12 (6), 1073e1081. van Nes, F.L., Koenderink, J.J., Nas, H., Bouman, M.A., 1967. Spatiotemporal modulation transfer in the human eye. J. Opt. Soc. Am. 57 (9), 1082e1088.