Differential Changes of Magnocellular and Parvocellular Visual Function in Early- and Late-Onset Strabismic Amblyopia

Abstract

purpose. Studies in nonhuman primates show that monocular visual deprivation starting at different ages has different effects on cells in the parvocellular and magnocellular laminae of the lateral geniculate nucleus. The present study used color and luminance contrast sensitivity (CS) measurements to look for differences in parvocellular- and magnocellular-related visual function in human subjects with strabismic amblyopia.

methods. Fifteen subjects with early- and 14 with late-onset strabismic amblyopia and similar ranges of visual acuity were studied, together with 15 subjects with normal vision. Contrast sensitivities were measured to an equiluminant (L-M cone-modulated) grating with slow onset and an achromatic (L+M cone-modulated) 0.8-cpd grating with rapid onset using an adaptive method.

results. Luminance and color CS were lower in the amblyopic eyes than in the fellow eyes of all amblyopes. For luminance CS, this was due both to an increase in sensitivity of the fellow eye and to a reduction in sensitivity in the amblyopic eye. Color CS was greatly reduced in the amblyopic and fellow eyes of subjects with strabismic amblyopia of early- and late onset compared with subjects with normal vision. The reduction in color CS compared with luminance CS was significantly greater in eyes with late- rather than early-onset amblyopia.

conclusions. Parvocellular and magnocellular function are differentially affected in the amblyopic and fellow eyes of subjects with strabismic amblyopia. The difference is more marked in late-onset amblyopia than in early-onset amblyopia.

The central visual pathways in the primate contain two major anatomic subdivisions that transmit information from the retina to the visual cortex in parallel. The magnocellular (M) pathway has a synaptic relay in the magnocellular laminae of the lateral geniculate nucleus (LGN) and terminates mainly in layer IVcα of the primary visual cortex, whereas the parvocellular (P) pathway has a relay in the parvocellular LGN laminae and terminates mainly in layer IVcβ. ^{1 2} Cells in the magnocellular pathway respond best to achromatic stimuli of low spatial and high temporal frequencies, whereas cells in the parvocellular pathway respond best to chromatic stimuli of high spatial and low temporal frequencies. ^{3 4 5 6 7} It is thus possible to select stimuli that, at threshold, measure the sensitivity predominantly of one or the other pathway. By using such stimuli, the effects of amblyopia on the magnocellular and parvocellular pathways can be studied in humans. 

Studies in nonhuman primates have shown that monocular visual deprivation causes hypertrophy and shrinkage of neuronal cell bodies in the LGN. ^{8 9} After experimental monocular visual deprivation of early or late onset, the patterns of cell size change in the magnocellular and parvocellular LGN laminae of nonhuman primates differ, suggesting differential changes in sensitivity to deprivation with age in the two pathways and demonstrating the presence of two distinct periods of developmental sensitivity. ⁹ Changes in visual evoked potentials and in contrast sensitivity (CS) found in human subjects with strabismic amblyopia are different when the age of onset of amblyopia is before or after 18 months of age, indicating that there may also be more than one sensitive period for visual development in humans. ¹⁰ The stimuli used in that study were not selective for M and P pathways. The present study describes the psychophysical results of a similar study performed in subjects with strabismic amblyopia of early and late onset using stimuli biased to measure CS predominantly of either the M or the P pathway and to examine whether the two pathways are affected differently and whether the changes found differ between amblyopia of early or late onset. 

Subjects and Methods

Subjects

Fifteen subjects with strabismic amblyopia of early onset and 14 of late onset were recruited after attendance at a strabismus clinic (Table 1) . As in the previous study, subjects were assigned to early- or late-onset groups on the basis of a clear history of onset of amblyopia before or after 18 months of age. ¹⁰ Parental confirmation was obtained when necessary. The age of onset of squint was clearly remembered as, in most cases, the age at which patching was started. A number of the subjects remembered wearing their patches very poorly. It was not possible to establish reliable compliance data, but it is likely that those with poorer acuities had less successful patching. 

Most subjects underwent at least one operation for squint during childhood. Of those who initially had esotropia, one did not have surgery (primary esotropia), several still had a degree of esotropia after surgery (residual esotropia), and most became divergent (consecutive exotropia). Only three subjects in each group initially had exotropia. Three never had surgery (primary exotropia), whereas redivergence occurred in three subjects after surgery (residual exotropia). All subjects underwent full orthoptic and ophthalmic assessments before testing. Seventeen of the subjects had participated in the previous study. ¹⁰ Fifteen young adults with normal vision formed the control group. 

Stimulus Presentation

Stimuli were presented in the center of a monitor (MultiSync P1150; NEC Corp., Tokyo, Japan) at a mean luminance of 90 cd/m². Stimuli were surrounded by a luminance-matched field (10° × 8°). The room was darkened. CIE coordinates of the phosphors (x, y) were measured (Chromameter CS100; Minolta, Tokyo, Japan) and were found to be 0.616, 0.355 for the red gun, 0.297, 0.603 for the green gun, and 0.143, 0.060 for the blue gun. The background was kept constant with a medium gray. From the CIE 1931 tristimulus values, the L-, M-, and S-cone signals were obtained by the transformation suggested by Smith and Pokorny (for further details see Wuerger & Morgan ¹¹ ). M- and P-biased stimuli were generated on a personal computer using a 14-bit VSG card and MATLAB (Cambridge Research Ltd., Rochester, UK). 

Measurement of Contrast Sensitivity

CS was measured using stimuli designed to stimulate predominantly magnocellular (M) or parvocellular (P) pathways. These stimuli cannot be guaranteed to be “pure” M- and P- isolating stimuli, ¹² but at detection threshold the assumption is that they will be detected by the more sensitive mechanism. ¹³  

Magnocellular-Biased Stimulus

The M-biased stimulus was an achromatic, horizontally orientated 0.8-cpd sinusoidal grating pattern with spatial (σ_x = σ_y = 1.0 deg) Gaussian windows. This gave a stimulus subtending approximately 1° of visual angle at the viewing distance of 0.92 m. The stimulus was multiplied by a temporal Gaussian window, with SD σ_t = 10,000 ms producing a transient stimulus with essentially sharp temporal edges like a square wave. 

Parvocellular-Biased Stimulus

The P-biased stimulus was a red-green horizontally orientated 3.2 cpd sinusoidal grating patch presented with a spatial (σ_x = σ_y = 0.5°). The grating was modulated in the equiluminant L-M direction in cone-contrast space (for details, see Wuerger and Morgan ¹¹ ). The stimulus subtended a visual angle of approximately 2° at the viewing distance of 0.92 m and was multiplied by a temporal Gaussian envelope with σ_t = 125 ms, producing gradual onset and offset of the stimulus. 

Threshold Measurement

Monocular detection thresholds were measured for the stimuli using a temporal, two-alternative forced choice (2AFC) technique based on the Quest methodology. ¹⁴ A staircase procedure driven by the subject’s responses and controlled by computer determined the detection threshold. Each trial consisted of two presentations (cued by sounds), one of which contained the stimulus while the other was a blank field of the same space-averaged luminance. Assessment of threshold began with an estimate of threshold based on the previous testing experience of the investigator. Thresholds were determined from the mean of the probability density function (pdf). ¹⁵ The next trial was placed at the current most probable estimate of threshold using the mean of the posterior pdf. The session ended after 40 trials. The final estimate of threshold was given by the final mean of the posterior pdf. 

Statistical Analysis

Visual acuities between groups were compared using unpaired t tests after conversion to logMAR equivalents. Contrast sensitivity measurements were log transformed before statistical analysis. Comparisons between amblyopic and fellow eyes were made using paired t tests and between groups using unpaired t tests. 

The research followed the tenets of the Declaration of Helsinki. Subjects gave informed consent after explanation of the nature and possible consequences of the study. The research was approved by the Ethics Committee of Moorfields Eye Hospital. 

Results

All subjects presented as adults with a manifest squint. All except three in each group had been esotropic as children, though a number became exotropic (Table 1) . 

Subjects with early- and late-onset amblyopia had a similar range of refractive errors. The mean spherical equivalent for the early-onset amblyopic eyes was +2.11 DS, and for early-onset fellow eyes it was −0.19 DS. For the late-onset amblyopic eyes, the mean spherical equivalent was +3.04 DS, and for the late-onset fellow eyes it was +1.17 DS. No significant difference was observed between the early- and late-onset amblyopic eyes (P = 0.43). Eight subjects with early- and five with late-onset amblyopia had anisometropia greater than 1.5 D of spherical equivalent. No control subject had significant anisometropia. 

Subjects in both the early- and the late-onset groups had a similar spread of acuities in their amblyopic eyes (logMAR equivalent means: early-onset 0.76, Snellen equivalent means ≈ 6/35, 20/115; late-onset 0.66, Snellen equivalent ≈ 6/27, 20/91; P = 0.32). There was no difference between groups in acuity of the fellow eye (logMAR equivalent means: early onset −0.05, Snellen equivalent means ≈ 6/5, 20/18; late onset −0.08, Snellen equivalent ≈ 6/5, 20/17; P = 0.74) nor was there any difference between them and the control group. All control subjects had a visual acuity of at least 6/6 (20/20) in each eye with binocular single vision. 

Most subjects had manifest squint with suppression of the amblyopic eye (Table 1) . Subjects 2 and 15 had some variability in their suppression, with diplopia at times; subjects 20 and 25 had diplopia; and subject 18 had fully accommodative esotropia. 

Color (L-M) Contrast Sensitivity

Results for color CS for one random control subject eye and for the amblyopic and fellow eyes of the groups are shown in Figure 1A . Mean CS of the amblyopic eye was worse than that of the fellow eye for all amblyopes taken together and for early- and late-onset amblyopes taken separately (Table 2 , final columns). When compared with those of the control subjects, the mean CS of the amblyopic eyes was markedly reduced for early-onset, late-onset, and all amblyopia taken together (Table 2) . There was a smaller, but highly significant, reduction of mean CS in the fellow eyes of all three groups compared with the control group (Table 2) . 

Luminance (L+M) Contrast Sensitivity

Results for luminance CS for one random control subject eye and for the amblyopic and fellow eyes of the groups are shown in Figure 1B . CS of the amblyopic eye was worse than that of the fellow eye for all amblyopes taken together and for late-onset amblyopes taken separately (Table 3 , final columns). Although the difference for early-onset amblyopes was similar in magnitude to that for the late-onset amblyopes, it did not reach significance at the 5% level. Comparison with control subjects showed that the difference between amblyopic and fellow eyes was as much due to an increase in the CS of the fellow eyes as to a reduction in the amblyopic eyes, particularly in the late-onset group, though the differences from control eyes did not individually reach significance (Table 3) . 

Relationship of Color to Luminance Contrast Sensitivity

Because the value of color CS was much smaller than that for luminance CS, the relationship between them has been examined by calculating the ratio (luminance CS −color CS)/luminance CS. As the color CS reduces in relation to the luminance CS, this ratio approaches 1. 

Ratios for one random eye of control subjects and for the amblyopic and fellow eyes of early- and late-onset amblyopes are plotted in Figure 1C . The ratio for the amblyopic eye was greater than that of the fellow eye for all amblyopes taken together and for early- and late-onset amblyopes taken separately, indicating a greater relative change in color CS (Table 4 , final columns). When compared with that of control subjects, the ratio for the amblyopic eyes was markedly greater for all groups (Table 4) . There was a smaller, but nevertheless highly significant, increase in the ratio for the fellow eyes of early-onset and late-onset amblyopic eyes and all fellow eyes taken together compared with control (Table 4) . 

The ratio for the amblyopic eyes of the late-onset group was significantly greater than that for the early-onset group, indicating a greater reduction in color CS relative to luminance CS in the late-onset amblyopes. 

For the early-onset amblyopic eyes, the relative reduction in color CS relative to luminance CS was significantly greater in eyes with poorer acuity (Fig. 2 ; linear regression of [luminance CS − color CS]/luminance CS to logMAR equivalent acuity; r ² = 0.36; P = 0.018). For the late-onset amblyopic eyes, the ratio was larger, which means that color CS was relatively worse at the better acuities than for early-onset amblyopic eyes. Hence, the regression line was flatter and the relationship did not reach significance (Fig. 2 ; r ² = 0.15; P = 0.17). 

Discussion

This study showed that color and luminance CS of the amblyopic eyes of subjects with strabismic amblyopia are reduced in comparison with their fellow eyes. However, it also demonstrated that the changes in color and luminance CS differ from each other, that substantial changes occur in the CS of the fellow eye, and that there are differences in the patterns of change that relate to the age of onset of the amblyopia. 

Reduction in sensitivity for red/green chromatic stimuli has been described in amblyopic eyes, ^{16 17} though interestingly there is much less effect for blue/yellow stimuli. ^{16 18} The present findings using a red/green isoluminant stimulus are in general accord with these previous studies. 

For early- and late onset amblyopes, the magnitude of the differences in CS between amblyopic and fellow eyes are similar for color and luminance. However, comparisons between the amblyopes and control subjects reveal a more complex picture. Color CS in amblyopic and fellow eyes of amblyopes was reduced compared with that in control subjects because the reduction in the CS was greater for amblyopic eyes than for fellow eyes. However, for luminance CS, the difference between amblyopic and fellow eyes was as much caused by an increase in the CS of the fellow eye compared to the control eye as to a decrease in the CS of the amblyopic eye. Analysis of the (luminance CS − color CS)/luminance CS ratio confirmed that color CS was significantly more reduced than luminance CS for amblyopic and fellow eyes. 

In the present study the luminance stimulus had a rapid onset, whereas the color stimulus had a slow onset. Although it is not possible to completely isolate M and P responses, the luminance CS measured in this way at threshold can be expected to predominantly reflect the sensitivity of the M pathway and the color CS that of the P pathway. ¹³ Thus it is reasonable to infer that sensitivity of the P pathway is reduced in relation to that of the M pathway for amblyopic and fellow eyes. 

Some previous studies have described an increase in achromatic CS in the fellow eye of amblyopes with strabismus compared with control eyes, ^{19 20} whereas others have described reduced function. ^{21 22} The present data suggest a complex abnormality of fellow eye function in which the sensitivity of the P pathway is reduced relative to that of the M pathway. The differences in M and P sensitivity make it probable that the temporal and spatial characteristics of the stimuli used will affect the results. For example, CS measured using a brief computer-generated stimulus may give results different from those using contrast charts or low-contrast optotypes. 

There are interesting parallels between the findings of this study and previous work on the changes in LGN cell size in a nonhuman primate model of deprivation amblyopia. ^{8 9} In the primate, in addition to the relative shrinkage of deprived LGN cells compared with undeprived cells, an increase in the ratio of M to P cell size was observed for deprived eyes after long-term deprivation and undeprived eyes. The relative reduction of P compared with M function observed for amblyopic and fellow eyes in the present study parallels these changes and suggest a similar underlying pathophysiology in both instances. 

Although early- and late-onset amblyopes with strabismus had a similar range of acuities in their amblyopic eyes, the reduction in P relative to M sensitivity was significantly greater in the late-onset group. This confirms the findings of a previous study in humans ¹⁰ and anatomic studies of monocular deprivation in primates ⁹ indicating that the underlying pathophysiology is different in amblyopia of early and late onset. One possible explanation is that M pathways simply mature and lose plasticity earlier than P pathways. However, studies in nonhuman primates showed a peak in the sensitivity of P cells to deprivation at 6 to 9 months of age, as judged by the shrinkage caused by 2 months of monocular deprivation begun at different ages. ⁸ This indicates a real increase in parvocellular sensitivity to deprivation between the ages equivalent to our early- and late-onset groups. 

Most subjects with strabismic amblyopia rely on their fellow eyes for their everyday functional vision; the amblyopic eye is suppressed. The present study indicates complex changes in visual function that affect the fellow eye of such subjects. Most of our subjects had had patching of their fellow eyes, with limited and variable compliance and effect as judged by their present acuities, so it is not possible to exclude this as a cause of the changes in their fellow eyes. However, the striking size changes seen in undeprived LGN cells in studies of nonhuman primates occur in the absence of any direct deprivation of that eye. ⁹ The impact of these changes in the fellow eye pathway on visual perception merits further study. 

 Supported by the Special Trustees, Moorfields Eye Hospital; the Ann Allerton Fund, Royal College of Ophthalmologists; a Wellcome Trust Basic Science Training Fellowship (ARD); and a grant from the Biotechnology and Biological Sciences Research Council (MJM).

 Submitted for publication April 5, 2006; revised June 26, 2006; accepted September 14, 2006.

 Disclosure: A.R. Davis, None; J.J. Sloper, None; M.M. Neveu, None; C.R. Hogg, None; M.J. Morgan, None; G.E. Holder, 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.

 

Table 1.

 

Clinical Features of Early- and Late-Onset Strabismic Amblyopes

Subject	Age (years)	Diagnosis	Snellen Acuity			Anisometropia > 1.50 DS	Binocular Status
								Amblyopic Eye	Fellow Eye
Early onset
1	18	30Δ Residual ET	6/36	6/9	N	Constant diplopia
2	33	25Δ Residual ET	6/24	6/6	N	Variable suppression
3	22	60Δ Consecutive XT	6/9	6/4	Y	Suppression
4	24	4Δ Residual XT	6/18	6/5	Y	Suppression
5	33	35Δ Residual XT	6/60	6/5	Y	Suppression
6	17	45Δ Residual XT	1/60	6/5	Y	Suppression
7	41	40Δ Consecutive XT	HM	6/4	Y	Suppression
8	40	45Δ Residual ET	6/9	6/6	Y	Suppression
9	33	25Δ Residual ET	6/9	6/4	N	Suppression
10	50	35Δ Consecutive XT	6/24	6/5	N	Suppression
11	43	25Δ Consecutive XT	6/24	6/5	N	Suppression
12	21	12Δ Consecutive XT	6/18	6/4	N	Suppression
13	19	35Δ Consecutive XT	6/24	6/6	Y	Suppression
14	32	14Δ Residual ET	6/9	6/4	N	Suppression
15	23	8Δ Residual ET	6/24	6/9	Y	Variable suppression
Mean	29.9
Late onset
16	36	35Δ Primary ET	6/9	6/5	N	Suppression
17	49	25Δ Primary XT	6/9	6/4	Y	Suppression
18	39	Fully Accommodative ET	6/12	6/6	N	BSV
19	60	25Δ Residual XT	6/12	6/6	Y	Suppression
20	29	40Δ Consecutive XT	6/24	6/5	Y	Constant diplopia
21	30	45Δ Consecutive XT	6/24	6/5	N	Suppression
22	27	30Δ Residual ET	6/24	6/5	N	Suppression
23	34	40Δ Consecutive XT	6/36	6/6	N	Suppression
24	49	60Δ Consecutive XT	6/60	6/5	N	Suppression
25	36	50Δ Consecutive XT	6/18	6/5	N	Constant diplopia
26	31	45Δ Consecutive XT	2/60	6/4	Y	Suppression
27	30	40Δ Consecutive XT	1/60	6/4	Y	Suppression
28	20	25Δ Residual XT	6/6	6/4	N	Suppression
29	47	45Δ Consecutive XT	6/60	6/6	N	Suppression
Mean	36.9

 Δ, prism diopters; ET, esotropia; XT, exotropia.

Figure 1.

 

Mean values for log color CS (A), log luminance CS (B), and the ratio of color to luminance CS (C) calculated as (luminance CS − color CS)/luminance CS) for control eyes and early- and late-onset amblyopic and fellow eyes. Error bars, ±1 SE.

Table 2.

 

Color Contrast Sensitivity

	Amblyopic Eye vs. Normal Eye							Fellow Eye vs. Normal Eye							Amblyopic Eye vs. Fellow Eye
		Color Contrast Sensitivity	Difference from Normal (%)	t	P	Color Contrast Sensitivity	Difference from Normal (%)	t	P	Difference (%)	t	P
Normal (n = 15)	7.45	—	—	—	7.45	—	—	—	—	—	—
All amblyopes (n = 29)	2.35	−68	3.83	0.0004	5.13	−31	3.49	0.0011	−54	4.41	0.0001
Early-onset amblyopes (n = 15)	2.67	−64	3.31	0.0026	5.50	−26	2.93	0.0066	−51	2.80	0.0141
Late-onset amblyopes (n = 14)	2.01	−73	12.04	<0.0001	4.73	−37	11.54	<0.0001	−57	3.36	0.0052

Table 3.

 

Luminance Contrast Sensitivity

	Amblyopic Eye vs. Normal Eye							Fellow Eye vs. Normal Eye							Amblyopic Eye vs. Fellow Eye
		Luminance Contrast Sensitivity	Difference from Normal (%)	t	P	Luminance Contrast Sensitivity	Difference from Normal (%)	t	P	Difference (%)	t	P
Normal (n = 15)	36.6	—	—	—	36.6	—	—	—	—	—	—
All amblyopes (n = 29)	33.1	−10	0.83	0.4086	47.1	+29	1.42	0.1644	−3	3.13	0.0044
Early-onset amblyopes (n = 15)	29.8	−19	1.10	0.2801	43.9	+20	0.74	0.4646	−32	1.92	0.0759
Late-onset amblyopes (n = 14)	36.7	0	0.33	0.7408	50.5	+38	1.70	0.0997	−27	2.56	0.0236

Table 4.

 

Ratio of Luminance to Color Contrast Sensitivity

	Amblyopic Eye vs. Normal Eye							Fellow Eye vs. Normal Eye							Amblyopic Eye vs. Fellow Eye
		Luminance to Color Ratio*	Difference from Normal (%)	t	P	Luminance to Color Ratio*	Difference from Normal (%)	t	P	Difference (%)	t	P
Normal (n = 15)	0.78	—	—	—	0.78	—	—	—	—	—	—
All amblyopes (n = 29)	0.93	+19	7.03	<0.0001	0.86	+11	3.55	0.0010	+8	4.20	0.0002
Early-onset amblyopes (n = 15)	0.90	+16	4.77	<0.0001	0.84	+8	2.24	0.0329	+8	2.55	0.0222
Late-onset amblyopes, † (n = 14)	0.95	+22	7.36	<0.0001	0.89	+14	4.01	0.0004	+7	3.52	0.0038

*  Calculated as (luminance CS − color CS)/luminance CS.

†  Eyes with late-onset amblyopia significantly different from eyes with early-onset amblyopia (t = 2.24; P = 0.034; unpaired t test).

Figure 2.

 

The ratio of color CS to luminance CS calculated as (luminance CS − color CS)/luminance CS for early- and late-onset amblyopic eyes plotted against logMAR equivalent acuity. For the early-onset amblyopic eyes with better acuities, the ratio is significantly lower, indicating that in them color CS is relatively better in relation to luminance CS, but this is not so for late-onset amblyopic eyes.