HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Li, X. Articles by Hess, R. F. Search for Related Content PubMed PubMed Citation Articles by Li, X. Articles by Hess, R. F.

Cortical Deficits in Human Amblyopia: Their Regional Distribution and Their Relationship to the Contrast Detection Deficit

From the Department of Ophthalmology, McGill University, Montreal, Quebec, Canada.

	Abstract

Top Abstract Methods Results Discussion Appendix 1 References

 
PURPOSE. The understanding of the site and nature of the cortical processing deficit in human amblyopia awaits the resolution of three fundamental questions about which there is, at present, much controversy: First, is area V1 affected as the present animal models would predict, but some imaging studies argue against? Second, how extensive is the loss of extrastriate function, and does it simply follow as a consequence of an impaired V1 input? Third, does the brain imaging deficit, be it striate or extrastriate, correlate with the well-documented psychophysical loss?—a fundamental issue on which previous brain imaging studies are divided.

METHODS. A spatially broadband stimulus was used to determine the functional MRI responses from the different retinotopically identified visual cortical areas in a group of normal (n = 6) and a group of amblyopic (n = 11) observers. Responses were compared between the amblyopic and fellow fixing eyes of amblyopes and between the dominant and nondominant eyes of normal subjects, in central and peripheral parts of the visual field. Psychophysical acuity and contrast sensitivity was also measured and its correlation with the brain imaging deficit determined.

RESULTS. V1 was affected in most but not all cases; the brain-imaging deficit involved extensive regions of extrastriate cortex and, at least with the stimuli used in the study, correlated with the V1 loss, suggesting a strong V1 influence; and neither the striate nor the extrastriate deficits correlated with the psychophysical contrast threshold losses at either high or low spatial frequencies.

CONCLUSIONS. The results suggest that there are significant suprathreshold processing deficits that are not a consequence of the well-known threshold deficit. Our preoccupation over the past 30 years with the contrast detection deficit in amblyopia limited to the processing within a circumscribed part of V1 may have to be modified to include not only processing deficits for high-contrast stimuli but also the involvement of multiple extrastriate areas.

The psychophysical picture of human amblyopia is mainly in terms of contrast sensitivity. Amblyopes detect low-spatial-frequency stimuli with the same sensitivity as in normal subjects but require more contrast to detect higher spatial frequency stimuli.^{5 6 7} In strabismic amblyopia, this detection deficit involves mainly the central field.^{8 9} Contrast perception is normal however, even at high spatial frequencies, well above their raised detection thresholds.¹⁰ Thus, the visual problem in amblyopia on which our present animal models are based is limited to low contrast and high spatial frequencies and the central visual field.

Animal models of amblyopia have shown that strabismus or anisometropia produce a loss of binocularity of cortical cells,¹¹ an upset in the excitatory-inhibitory balance of binocular cells,^{12 13} reduced strength of cellular interactions¹⁴ and reduced mean sensitivity and mean spatial resolution of the foveal V1 cellular population as a whole.¹⁵ These latter findings have lead to the widely held belief that the contrast sensitivity or visibility deficit can be explained in terms of the V1 response, although it has been acknowledged that the V1 losses may not be the whole story; they are quantitatively less than the behavioral deficit across a group of amblyopic animals, and there is a great deal of interanimal variability.¹⁶ There is still some controversy concerning whether the number of cells driven by the amblyopic eye is different from that driven by the fellow fixing eye. In general it is not,^{11 14 17 18 19 20} but there is evidence that in the case of severe convergent strabismus there may be fewer cells driven by the deviated eye,^{11 15 17 18 21 22} and also it has been reported, in a subset of animals, that there may be less foveal V1 neurons driven by the amblyopic eye.^{15 23} Granted we are not at the stage where we can go from cellular responses in V1 to perception even in normal animals, but there is still a demonstrated correlation between the visibility defect in amblyopia as assessed by spatial resolution and peak contrast sensitivity and the responses of cells in V1 driven by the amblyopic eye. There is also a realization, however, that there could be more to the story either by a consideration of V1 processing beyond that of isolated cells or by considering responses in visual areas other than V1.

Brain imaging in humans with amblyopia has the potential of providing additional information on the site of the deficit. It has the advantage of not requiring the intermediate animal model stage and reflecting a population measure of cortical function. Its main disadvantage is that it is a specific and indirect measure of neural function.²⁴ The results to date have been inconsistent; some studies have argued for normal V1 function, with the disturbance being exclusively restricted to the extrastriate cortex (Sireteanu R et al. IOVS 1998;39:ARVO Abstract 4186).^{25 26 27} Others have argued that V1 activation is reduced.^{28 29 30 31 32 33 34 35 36 37} Of the studies that report reduced V1 activation, some report a pattern of inactivation that is consistent with the psychophysical loss,^{30 34} whereas others argue that this is not the case.^{31 32 33} In particular, Anderson et al.,³¹ using magnetoencephalography (MEG), and Barnes et al.,³³ using fMRI, argue that the reduced activation does not correlate with the contrast sensitivity loss. In both of these studies, not only was reduced activation found at low spatial frequencies where contrast sensitivity is rarely affected in amblyopia, but also any reduced activation found at high spatial frequencies did not correlate with the loss in contrast sensitivity. Another finding in most imaging studies is that the extrastriate cortex is also affected (Sireteanu R et al. IOVS 1998;39:ARVO Abstract 4186),^{25 26 27 33 35} though the extent of the deficit is unknown, as is whether it simply follows from the V1 loss.

There are three questions that must be resolved to further our understanding of the cortical deficit in humans with amblyopia. First, is V1 affected? Animal studies suggest it is,^{15 23} but some recent brain imaging studies argue that it is not (Sireteanu R et al. IOVS 1998;39:ARVO Abstract 4186).^{25 26 27} Second, how extensive is the loss of extrastriate function, and does it simply follow as a consequence of an impaired V1 input? Animal studies have traditionally been limited to V1^{15 23} and provide little elucidation. Brain imaging studies suggest that there is extrastriate dysfunction (Sireteanu R et al. IOVS 1998;39:ARVO Abstract 4186)^{25 26 27 28 29 30 31 32 33 34 35 36} but it is presently unresolved how extensive this is and, more important, whether it simply follows as a simple consequence of the V1 loss. If the extrastriate losses are shown to correlate with the V1 loss, then it suggests that the main site of dysfunction in amblyopia may be in V1. If the extrastriate losses are shown not to correlate with the V1 loss, then it suggests that there are independent, primary deficits in the extrastriate cortex. Third, does the brain imaging deficit, be it striate or extrastriate, correlate with the well-documented psychophysical contrast detection loss described earlier?—a fundamental question on which previous brain imaging studies are divided. Previous studies that have used spatially and temporally narrowband (i.e., sinusoidal grating) stimuli have suggested no correlation between reduced activation and either contrast sensitivity or grating acuity, whereas the studies that provide support for the relationship have used spatially and temporally broadband stimuli.^{30 34} This in itself is intriguing because one would have thought the opposite to be the case: The more spatially narrowband the stimuli, the great the correlation with the loss of only a subset of cells. Its importance is that it raises the intriguing question as to whether there are processing deficits not related solely to detection (i.e., deficits for well detectable, high-contrast stimuli).

In this investigation, we set out to answer these important questions by an assessment of the magnitude and extent of the cortical deficit in retinotopically mapped regions of the striate and extrastriate cortex in humans with strabismic amblyopia. We used a spatially and temporally broadband mapping stimulus to assess whether any reduced cortical dysfunction correlates with the visibility loss, because the only evidence suggesting such a correlation has involved the use of broadband stimulation.^{30 34} Such a mapping approach allows the assessment of the regional extent (i.e., visual field) of the functional deficit in different visual areas as well as the extent to which the extrastriate deficit is correlated with either the visibility loss or the V1 loss. The picture that emerges is one of a relatively consistent, correlated striate and extrastriate functional loss for suprathreshold spatiotemporal broadband stimuli that cannot be simply explained by the known psychophysical contrast detection deficit. We conclude that the cortical processing deficit in amblyopia is extensive, involving large regions of extrastriate cortex and is not simply explicable in terms of the known contrast detection deficit. Current notions that only high-spatial-frequency contrast processing in area V1 is disrupted in amblyopia should be revised.

	Methods

Top Abstract Methods Results Discussion Appendix 1 References

 
Subjects
Table 1 shows the clinical data for the 11 amblyopic subjects (average age, 34 ± 15 years) enrolled in the study. Clinically, amblyopia in humans can be subdivided into pure strabismus without anisometropia, pure anisometropia without strabismus and a mixed form where strabismus and anisometropia coexist. Six of our subjects had strabismic amblyopia, five had mixed strabismic-anisometropic amblyopia. During both the fMRI and psychophysics sessions, subjects wore nonmagnetic spectacles to give them corrected acuity based on refraction. A control group of six normal subjects (average age, 29.8 ± 4 years) was also tested. During the scanning sessions, subjects monocularly viewed a stimulus back-projected into the bore of the scanner and viewed through an angled mirror. The eye not being stimulated was occluded with a black patch that excluded all light from the eye. All studies were performed with the informed consent of the subjects and the approved by the Montreal Neurologic Institute Research Ethics Committee and adhered to the tenets of the Declaration of Helsinki.

View larger version (11K): [in this window] [in a new window]   FIGURE 1. The psychophysical loss. (A) The acuity deficit was typically worse for letters than for gratings. (B) The contrast sensitivity deficit involved high and low spatial frequencies (1 cyc/deg).

View larger version (67K): [in this window] [in a new window]   FIGURE 2. One example of BOLD signal change and t statistic calculation in our amblyopic subjects. (A, B) Time series from one eccentricity (clockwise order) run (fixing and amblyopic eye response time course, respectively), the fundamental component is of reduced amplitude when the cortex is driven by the amblyopic eye. (C, D) shows the time series from one polar angle (clockwise order) run (fixing and amblyopic eye response time course, respectively). All responses are at the same pixel position.

Using an automatic volumetric analysis,^{40 41} we defined separately the VFS map for fixing and fellow amblyopic eye stimulation. We then defined the common boundaries of different visual areas by combining the VFS information. We compared the signal strength within these commonly defined visual areas for dominant and nondominant eye stimulation of normal subjects and for fixing and fellow amblyopic stimulation of amblyopes. Eccentricities between 1.3° and 17° were tested.

Image Acquisition
A 1.5-T (Magnetom; Siemens, Madison, WI) scanner was used to collect both anatomic and functional images. Anatomic images were acquired with a rectangular coil (14.5 x 6.5 in.), head coil (circularly polarized transmit and receive), and T₁-weighted sequence (TR [recovery time] = 22 ms; TE [echo time] = 10 ms; flip angle = 30°), giving 176 sagittal slices of 256 x 256-mm³ image voxels. Functional scans for each subject were collected with a surface coil (circularly polarized, receive only) positioned beneath the subject’s occiput. Each functional imaging session was preceded by a surface coil anatomic scan (identical with the head coil anatomic sequence, except that 80 x 256 x 256 sagittal images of 2-mm slice thickness were acquired) to coregister the data later with the more homogeneous head coil image. Functional scans were multislice T₂*-weighted, gradient-echo, planar images (TR = 3.0 seconds, TE = 51 ms, flip angle = 90°). Image volume consisted of 30 slices orthogonal to the calcarine sulcus. The field of view was 256 x 256 mm, the matrix size was 64 x 64 with a thickness of 4 mm, giving voxel sizes of 4 x 4 x 4 mm. Each experiment consisted of four acquisition runs for each eye (two eccentricity, two polar angle, two clockwise order, and two counterclockwise) each of 128 image volumes acquired at 3-second intervals for either the left and right eye of normal subjects or the fixing and amblyopic eye of amblyopes. Runs were alternated between the eyes in each case while the subject was performing the described task.

Data Analysis
Anatomic Images.
The global T₁-weighted anatomic (a)MRI scans were corrected for intensity nonconformity⁴² and automatically registered⁴³ in a stereotaxic space⁴⁴ using a stereotaxic model of 305 brains.⁴⁵ The surface coil aMRI, acquired in the same session as the functional images, was aligned with the head coil aMRI, thereby allowing an alignment of the functional data with the head coil MRI and subsequently the stereotaxic space. This method was validated in a previous study.⁴¹ The aMRIs were classified into gray matter, white matter, and cerebrospinal fluid (CSF), after which the four (gray matter, white matter, and left and right hemispheres) cortical surfaces for all subjects were simultaneously reconstructed at the inner and outer edges of the cortex.^{46 47 48} All processing steps were completely automatic, and all the data are presented in a stereotaxic space.

Functional Images.
Dynamic motion correction for functional image time series for each run and for different runs were realigned at the same time by using fmr_preprocess (provided in the MINC software package: http://www.bic.mni.mcgill.ca/software/) with three-dimensional Gaussian low-pass filtering of time series data. The first eight scans of each functional run were discarded due to start-up magnetization transients in the data. The t statistic was calculated based on the fundamental frequency effects of the Fourier transformation (see the Appendix; The software, BrainVision, used for the fMRI data analysis is available in the public domain at: http://www.mvr.mcgill.ca/Li/software.htm). The design matrix was formed by combining fundamental frequency effects and the low-frequency drift effects. A general linear model (GLM) method was used to quantify the brain activation. Inference for the effects was developed according to the t statistics. Different runs for each eye were combined in a mixed model.^{49 50} The paired t statistic was used to test the significance of the activation (t₍₁₁₇₎ > 1.98; P < 0.05, two-tailed t-test) between eyes.

Psychophysics
Eye Occlusion and Dominance.
To test monocular function, we occluded either the fixing or fellow amblyopic eye with a black patch designed to exclude all light. This test was used for both the psychophysical testing and for the brain imaging. Under these conditions there is no binocularly mediated suppression of the amblyopic eye because all pattern vision in the good eye has been abolished⁵¹ and thus our estimates of the reduced activation in the brains of amblyopes is a conservative one, as it does not include a binocular suppressive component. Eye dominance was assessed with a standard sighting test.⁵²

Eye-Movement Measurement.
The fixation eye movements for the normal and fellow amblyopic eyes were measured separately for each subject with a video eye tracker (Cambridge Research Systems, Cambridge, UK) which sampled fixations at 50 Hz. Total viewing duration ranged from 3000 to 5000 ms. The subject was asked to fixate on a central fixation mark similar to that used in the scanner.

Grating Acuity.
For each subject, grating acuity for optically corrected subjects was established by using a method of constant stimuli and a two-alternative, forced-choice (2AFC) paradigm with a VSG 3/4 card and a gamma-corrected monitor (Multisync XP 17; NEC). The two intervals were presented, one containing a static vertical grating of 80% contrast and the other, just mean luminance. The subject’s acuity (corresponding to 82% correct identification) was measured with a staircase method over fixed spatial frequency steps from 2 to 52 cyc/deg. The field size had a diameter of 10°. A central fixation mark was provided.

Low-Spatial-Frequency Contrast Sensitivity.
Contrast sensitivity (1 cyc/deg) measures were made with the VSG 3/4 card and gamma-corrected monitor (NEC). Contrast sensitivity was measured with a 2AFC paradigm and a staircase algorithm. Contrast sensitivity thresholds corresponding to 82% correct were determined for each eye of each amblyope. The field size was 10° diameter, and a central fixation mark was provided.

	Results

Top Abstract Methods Results Discussion Appendix 1 References

 
The Psychophysical Picture
The psychophysical loss exhibited by anisometropic and strabismic amblyopes is now well established.^{5 6 7} It involves a letter acuity recognition deficit and a contrast sensitivity deficit. In terms of acuity, the deficit for recognizing letters is usually greater than it is for detecting gratings.⁵³ In terms of contrast sensitivity, higher contrast thresholds (i.e., reduced contrast sensitivity) are found at high spatial frequencies. The contrast threshold abnormality for some amblyopes is limited to high spatial frequencies, thresholds at low spatial frequencies being normal or in some cases better than normal. Others have a small residual, but constant, deficit at low spatial frequencies (1 cyc/deg and lower). Because no cases have been reported in which contrast deficits increase at low spatial frequencies, we have conservatively estimated the contrast threshold deficit in the low-spatial-frequency range by using a 1 cyc/deg stimulus. Above the raised threshold, amblyopes perceive the contrast of stimuli normally with the amblyopic eye.^{10 54 55} Furthermore, amblyopes can discriminate differences in contrast with normal accuracy at low spatial frequencies,⁵⁵ but at high spatial frequencies, there is evidence of a slight reduction in sensitivity.⁵⁵ A similar suprathreshold deficit has been reported in animals made artificially amblyopic and shown to be unrelated to the detection threshold deficit.⁵⁶ Thus, the picture that emerges is one of reduced function for low contrast stimuli of high spatial frequency.

In Figure 1 the psychophysical deficits are seen for our group of amblyopes. This figure shows a comparison of the acuity deficits (in A, grating and letter acuity; ratio of fixing eye/amblyopic eye) and the contrast sensitivity deficits (in B, grating acuity and contrast sensitivity at 1 cyc/deg). Grating acuity involved detection, and letter acuity involved recognition (linear acuity, each line containing six letters, five of which had to be correctly identified). In each case, we saw the expected result. The acuity deficit was greater for letters, and the contrast sensitivity deficits involve high and low spatial frequencies.

The fMRI Picture
Figure 2 shows typical data (stars and dotted lines) for the fixing and amblyopic eyes of one of our amblyopic subjects (LM) for one voxel for the eccentricity (Figs. 2A 2B) and polar angle (Figs. 2C 2D) stimuli. The solid continuous curve represents the fitted response for calculation of the t statistic. The amblyopic response has a reduced t statistic but the difference is not large. We used the t statistic, because we wanted a measure that reflected response variance as well as amplitude.⁴⁹ For example, the Fourier amplitudes derived from the data displayed in Figures 2B and 2C are similar, but the higher t statistic for the data in Figure 2C stems from the better correlation of the signal with the sinusoidal variation of the stimulus. Although the use of the t statistic to quantify the fidelity of the response is superior to using just an amplitude measure,⁴⁹ we subsequently verified that our conclusions did not critically depend on our use of the t statistic. Similar results were obtained when using Fourier magnitude analysis.

Figure 3 shows an example of the regional brain activation by the mapping stimuli, for the fixing eye (Figs. 3A 3C) compared with the fellow amblyopic eye (Figs. 3B 3D) of one of the subjects (LM). A comparison of Figures 3A and 3B shows stronger activation of the fixing eye by the eccentricity stimulus where as Figures 3C and 3D illustrate the stronger activation of the fixing eye by the polar angle stimulus. Figure 4 shows a comparison of the regional reductions in activation for the combined runs (two polar angle and two eccentricity) for four randomly selected amblyopes (fixing versus amblyopic eye activation) for the retinotopically mapped cortical areas (solid black lines). In general, we found good correspondence between the VFS maps for fixing and fellow amblyopic eye stimulation, such that the commonly defined boundaries of different visual areas, used for our later analysis, were similar for fixing and fellow amblyopic eye stimulation. The reductions involved both striate and extrastriate cortical areas, and both hemispheres were affected. Only amblyope OA showed a strong hemispheric asymmetry and a better response with the amblyopic eye in V1-V2).

View larger version (115K): [in this window] [in a new window]   FIGURE 3. An example (subject LM) of the brain activation region for fixing (A, C) and amblyopic (B, D) eye stimulation in response to expanding, eccentricity (A, B) and clockwise, polar angle stimuli (C, D). Note the similar patterns of activation when each eye is stimulated, as well as the reduced activation when the amblyopic eye was stimulated. Color bar, the t statistic; threshold is t > 4.79.

View larger version (63K): [in this window] [in a new window]   FIGURE 4. For illustrative purposes only, an oblique view of the unfolded cortex for four randomly selected amblyopes with the reduction in activation ((combined runs using good eye) – (combined runs using amblyopic eye); quantified in terms of the t statistic) for amblyopic eye activation indicated. Reduced activation occurs not only in V1 but also extends to other retinotopic extrastriate visual areas (i.e., V2, V3, Vp, V3a, and V4). The hemispheric asymmetry apparent for one subject (OA) was not a general finding (see Fig. 12 ).

View larger version (47K): [in this window] [in a new window]   FIGURE 5. Group analysis. The mean and SEM of the t statistic from different runs for each VOI analysis for each subject group. There were 6 normal subjects and 11 amblyopes.

View larger version (25K): [in this window] [in a new window]   FIGURE 6. Brain activation (t statistic) from dominant versus nondominant eye stimulation in normal subjects; the slope of the dashed line is 1; the slope of the solid line is estimated by using robust regression method to the data. Slope is defined as the regression slope in y=ax+b.

View larger version (29K): [in this window] [in a new window]   FIGURE 7. Brain activation (t statistic) for fixing versus amblyopic eye in amblyopic subjects; Data from subjects whose initials are within dashed squares are significant (T > 1.960; P < 0.05, two tailed t-test). Dashed line: represents equal activation; solid line: the robust fit to the amblyopic data as a whole. Slope is the regression slope in y=ax+b.

Because the large intersubject variability evident in Figures 6 and 7 for the normal and amblyopic populations limits the sensitivity of the group comparisons (i.e., either in terms of the means in Figure 5 or the slopes in Figure 7 ), we assessed the significance (volume of interest [VOI] paired t-test; fixing versus amblyopic eye, P < 0.05) of the reductions in cortical activation for each amblyopic subject separately using the fellow fixing eye as the control. The advantage of such a comparison is that each subject can act as his or her own control, with a subsequent reduction in variability. The disadvantage is that the fellow fixing eye’s activation may be slightly reduced below that of the dominant eye of a normal observer (suggested by the results in Fig. 5C but not statistically significant in the group data) and as a consequence, any activation difference found between the amblyopic and normal fixing eye would underestimate the extent of the amblyopic dysfunction. In Figure 7 , for each visual area, we have indicated the subjects (by enclosing initials in a dashed box) whose reduced cortical activation when driven by their amblyopic eye was statistically significant.

To estimate the extent to which any reduction in cortical activation that is evident in the amblyopic population (i.e., Fig. 7 ) correlates with the contrast sensitivity deficit, we compared, on a subject-by-subject basis, the difference in t statistics between the fellow fixing and amblyopic eyes with both the ratio of grating acuities (Figs. 8A 8B) and contrast sensitivities (Figs. 8C 8D) . We made this comparison for visual areas V1 (Figs. 8A 8C) and V2 (Figs. 8B 8D) because the cells in these areas are thought, on the basis of their spatial properties, to make the main contribution to behavioral contrast sensitivity.^{57 58} We found no significant correlation (Fig. 8 , bottom right of each panel) for either area V1 or V2 between either visibility measure and the reduced activation associated with responses from the amblyopic eye, although the correlation between the V1 loss and the contrast sensitivity (r = 0.60) loss bordered on significance. A similar correlation analysis (data not shown) between fMRI and psychophysics for areas V3, Vp, and V3a also yielded insignificant correlations.

View larger version (24K): [in this window] [in a new window]   FIGURE 8. Grating acuity and contrast sensitivity ratios for each amblyope versus the reduced cortical activation; y-axis is the F statistic for the fixing eye minus the t statistic for the amblyopic eye; x-axis is the ratio of contrast sensitivity or grating. The respective correlation coefficients are shown.

View larger version (23K): [in this window] [in a new window]   FIGURE 9. The correlations between the reductions in striate and extrastriate cortex are shown by plotting the t statistic differences for V1 against those for other visual areas. The respective correlation coefficients are shown.

View larger version (32K): [in this window] [in a new window]   FIGURE 10. The difference in percentage of BOLD signal between the fixing and fellow amblyopic eyes (C, D) of our amblyopic group for centrally located (A) and peripherally located (B) stimuli. (E, F) Group correlations are shown for the reduced V1 activation for centrally located stimuli and the contrast sensitivity and grating acuity deficits for the same subjects.

View larger version (30K): [in this window] [in a new window]   FIGURE 11. The central versus peripheral reduction (%BOLD signal change in fixing eye – %BOLD signal change in amblyopic eye) in functional activation for individual amblyopic subjects. Dashed line: equal reduction for central and peripheral stimulation, whereas responses above the dashed line indicate larger differences between fixing and amblyopic eye activation for central stimulation.

View larger version (26K): [in this window] [in a new window]   FIGURE 12. Hemisphere difference comparison (same stimuli, same eye). Dashed line: the equality line (i.e., right hemisphere response = left hemisphere response), t represents the derived t statistic, and df is degree of freedom.

	Discussion

Top Abstract Methods Results Discussion Appendix 1 References

Currently, it is believed that the neural loss in V1 underlies the contrast sensitivity deficit in animals with amblyopia because of the strong correlation between the behavioral loss measured in terms of contrast sensitivity and grating acuity and comparable single-cell measures on the same animals.¹⁵ However, the current results and those of several previous studies of brain imaging^{26 31 33} in humans suggest no strong correlation between the reduced cortical activation and these two key measures of the psychophysical visibility loss: contrast sensitivity and grating acuity. This is either because human amblyopia is unlike its animal model or because the measures of cortical function that can be used in humans are fundamentally different from the single-cell measures used in animal studies. It is possible that the behavioral measures reflect the activity of just a few relevant cortical cells and therefore are reflected in the physiology, whereas fMRI and MEG represent more mass activity and as a consequence correlate less well. Furthermore, fMRI is an indirect measure of neural activity in V1, being more to do with its synaptic input and intracortical processing than the spiking activity of output cells,²⁴ and it may provide a quite different reflection of V1 processing. MEG (e.g., Anderson et al.³¹ ), provides a more direct measure of the magnetic fields associated with neural activity but is susceptible to cancellation effects from multiple sources as a consequence of cortical geometry. Another issue involves the type of stimuli used for functional studies. These may not be optimal for making the link with the psychophysics or with single-cell studies. Ideally, one should use well-localized (i.e., foveal) spatiotemporal narrowband stimuli of low contrast to try to isolate the responses of just the subset of the V1 neurons that the animal models suggest are the most vulnerable.^{15 23 64 65 66} The approach taken by Muckli et al.,²⁶ in which they compared grating acuity as assessed psychophysically in an fMRI event-related design, speaks directly to this issue. They found not only a dissociation between the psychophysics and fMRI, using what would appear to be appropriate, narrowband stimuli, but they also report either no deficit or enhanced amblyopic activation for areas V1 and V2. Having said that, the type of stimuli remains an important issue that future studies should try to resolve, because the reduced activation documented in this and previous studies (Sireteanu R et al. IOVS 1998;39:ARVO Abstract 4186),^{25 28 29 30 31 32 33 34 35} using high-contrast narrow and broadband stimuli, suggest that there are significant V1 deficits in functional activation for suprathreshold, low-spatial-frequency stimuli.

The Extrastriate Contribution
One of the purposes of this study was to assess the degree of extrastriate involvement, as a number of previous studies (Sireteanu R et al. IOVS 1998;39:ARVO Abstract 4186)^{25 26 27 33 36} have suggested that the extrastriate cortex is either solely affected or also affected, and yet we know so little about extrastriate function in animals made artificially amblyopic. We show that the dysfunction may extend to many of the extrastriate areas that can be topographically mapped (i.e., V2, V3 Vp, and V3a exhibited significant reductions across our amblyopic group). The reduced activation in striate and extrastriate areas mainly involves the central field representation—true of the amblyopic sample as a whole and in most, but not all, individual amblyopes (Fig. 7) . The lack of a strong correlation between the reduced cortical activation when driven by the amblyopic eye and the psychophysical deficit appears to hold for V2 as well as V1 and for central as well as whole-field stimulation. The reduced activation found in extrastriate areas showed a strong positive correlation with that found in V1 and this correlation was particularly strong in areas V2, V3, and Vp and weakest in area V4.

The Fellow Fixing Eye’s Response
There is psychophysical evidence that the fellow fixing eye of amblyopes is not completely normal. This evidence involves form as well as global motion.^{67 68 69 70 71 72} Our functional data are suggestive of reduced function, although the group activation levels were not significantly different between the dominant eyes of normal subjects and the fixing eyes of amblyopes. Such a reduction in function could be due to a reduction in the number of binocularly activated neurons in these cortical areas. Although much is known about the loss of binocularity of V1 neurons as a result of a strabismus,¹¹ there are also reports of decreased numbers of binocular extrastriate cells.^{18 73} Even though we suspect that the fellow fixing eye is reduced in its function compared with the dominant eye of normal subjects, we have still made all the comparisons between the response driven by the amblyopic and fellow fixing eyes to help combat the sometimes large intersubject differences in functional activation. In so doing, we are almost certainly underestimating the reduction in functional cortical activation produced by stimulation of the amblyopic eye.

We provide the following answers to the three posited questions: (1) Is V1 affected? The answer is yes, it is affected in most amblyopes. We found one case (OA) in which V1 and V2 were spared. (2) How extensive is the loss of cortical function, and does it simply follow as a consequence of an impaired V1 input? The answer is, there is extensive extrastriate loss, most of which correlates with the V1 loss, suggesting a primary site in V1, at least for the stimuli we used. (3) Does the brain imaging deficit, be it striate or extrastriate, correlate with the well-documented psychophysical loss? The answer is no, suggesting additional processing deficits for well detectable stimuli of low spatial frequency.

	Appendix 1

Top Abstract Methods Results Discussion Appendix 1 References

 
General Linear Model
The time series of BOLD signal change was normalized

where (i) is the normalized BOLD-fMRI time series, S(i) is the raw fMRI time series, i is the sampled time point (image frame), is the mean value of the series, and is the SD of the time series.

Based on general linear model(GLM)^{49 50}

(A1)

where is the regression parameters, is random error, and X_r is the n x k design matrix. We formed ₁(i) as:

where f is the fundamental frequency that was obtained by fast Fourier transformation of the response time series; iff is inverse Fourier transformation (the frequency of the stimuli was 6 in our study); real is the real part of the number. This is based on Euler’s equation: e^j = cos() + sin()j, where j = is the imaginary part of the inverse Fourier transformation. To eliminate the low frequency drift, we added polynomial drift ₂(i), , _k(i) in the design matrix (we set the order of the polynomial equal to 1; k=3, that is we add the ; ₂(i)=1 and slope drift ₃(i)=i in the design matrix).

The noise was modeled according to the first-order autoregressive model (AR(1))

so that the model becomes

(A2)

The least-squares method was adopted to solve the ^{equation (A2)} . We wanted to test H: Aß=c, A is pxq matrix, statistic for this is: A–c; is the estimated ß; H will be rejected if A is sufficiently different from c. The residual sum of squares (RSS) is calculated:

where || || is the norm, and

²=RSS/; where =n–rank(X)

(A3)

the estimated SD is

The t statistic is calculated as

In this study, because we were interested in fundamental frequency effects, we set c=0 and contrast A = (100) in ^{equation A3} .

The mixed-effects model was used to combine runs of each eye:

where j=1, , , is the total number of runs; Z is the design matrix (average the effects using covariates Z_j=1); _j is normally distributed with 0 mean and variance S_i² + _random² independently for j, _random² is estimated using the expectation maximum algorithm. The threshold was estimated using Tstat_threshold; P = 0.05, df = 117, t = 4.79^{49 50} ).

	Acknowledgements

 
The authors thank Magda Michna, Bruce Hansen, Elizabeth, and all the subjects who participated in the study; Jason Lerch and Alan Evans (McConnell Brain Imaging Centre) for surface reconstruction; and Keith Worsley for suggestions regarding the statistical analysis of the data.

	Footnotes

 
Supported by Canadian Institutes of Health Research Grant MOP53346 (RFH).

Submitted for publication August 29, 2006; revised October 7 and 17, 2006; accepted January 17, 2007.

Disclosure: X. Li, None; S.O. Dumoulin, None; B. Mansouri, None; R.F. Hess, 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.

Corresponding author: Robert F. Hess, 687 Pine Avenue W. (H4-14), McGill Vision Res, Department of Ophthalmology, McGill University, Montreal, Quebec, Canada, H3A 1A1; robert.hess{at}mcgill.ca.

	References

Top Abstract Methods Results Discussion Appendix 1 References

 

1. Cleland BG, Mitchell DEM, Gillard-Crewther S, Crewther DP. Visual resolution of ganglion cells in monocularly deprived cats. Brain Res. 1980;192:261–266.[CrossRef][ISI][Medline][Order article via Infotrieve]
2. Cleland BG, Crewther DP, Crewther SG, Mitchell DEM. Normality of spatial resolution of retinal ganglion cells in cats with strabismic amblyopia. J Physiol (Lond). 1982;326:235–249.[Abstract/Free Full Text]
3. Hess RF, Baker CL. Assessment of retinal function in severely amblyopic individuals. Vision Res. 1984;24:1367–1376.[CrossRef][ISI][Medline][Order article via Infotrieve]
4. Hess RF, Baker CL, Verhoeve JN, Tulunay Keesey U, France TD. The pattern evoked electroretinogram: Its variability in normal subjects and its relationship to amblyopia. Invest Ophthalmol Vis Sci. 1985;26:1610–1623.[Abstract/Free Full Text]
5. Gstalder RJ, Green DG. Laser interferometric acuity in amblyopia. J Pediatr Ophthalmol. 1971;8:251–256.
6. Levi M, Harwerth RS. Spatio-temporal interactions in anisometropic and strabismic amblyopia. Invest Ophthalmol Vis Sci. 1977;16:90–95.[Abstract/Free Full Text]
7. Hess RF, Howell ER. The threshold contrast sensitivity function in strabismic amblyopia: evidence for a two type classification. Vision Res. 1977;17:1049–1055.[CrossRef][ISI][Medline][Order article via Infotrieve]
8. Hess RF, Campbell FW, Zimmern R. Differences in the neural basis of human amblyopias: effect of mean luminance. Vision Res. 1980;20:295–305.[CrossRef][ISI][Medline][Order article via Infotrieve]
9. Hess RF, Pointer JS. Differences in the neural basis if human amblyopias: the distribution of the anomaly across the visual field. Vision Res. 1985;25:1577–1594.[CrossRef][ISI][Medline][Order article via Infotrieve]
10. Hess RF, Bradley A. Contrast coding in amblyopia is only minimally impaired above threshold. Nature. 1980;287:463–464.[CrossRef][Medline][Order article via Infotrieve]
11. Hubel DH, Wiesel TN. Binocular interaction in striate cortex of kittens reared with artificial squint. J Neurophysiol. 1965;28:1041–1059.[Free Full Text]
12. Sengpiel F, Blakemore C. The neural basis of suppression and amblyopia in strabismus. Eye. 1996;10:250–258.
13. Smith EL, 3rd, Chino YM, Ni J, Cheng H, Crawford ML, Harwerth RS. Residual binocular interactions in the striate cortex of monkeys reared with abnormal binocular vision. J Neurophysiol. 1997;78:1353–1362.[Abstract/Free Full Text]
14. Roelfsema PR, Koenig P, Engel AK, Sireteanu R, Singer W. Reduced synchronization in the visual cortex of cats with strabismic amblyopia. Eur J Neurosci. 1994;6:1645–1655.[CrossRef][ISI][Medline][Order article via Infotrieve]
15. Kiorpes L, Kiper DC, O’Keefe LP, Cavanaugh JR, Movshon JA. Neuronal correlates of amblyopia in the visual cortex of macaque monkeys with experimental strabismus and anisometropia. J Neurosci. 1998;18:6411–6424.[Abstract/Free Full Text]
16. Kiorpes L, McKee SP. Neural mechanisms underlying amblyopia. Curr Opin Neurobiol. 1999;9:480–486.[CrossRef][ISI][Medline][Order article via Infotrieve]
17. Kalil RE, Spear PD, Langsetmo A. Response properties of striate cortex neurons in cats raised with divergent or convergent strabismus. J Neurophysiol. 1984;52:514–537.[Abstract/Free Full Text]
18. Schroder JH, Fries P, Roelfsema PR, Singer W, Engel AK. Ocular dominance in extrastriate cortex of strabismic amblyopic cats. Vision Res. 2002;42:29–39.[CrossRef][ISI][Medline][Order article via Infotrieve]
19. Singer W, von Grunau M, Rauschecker J. Functional amblyopia in kittens with unilateral exotropia. I. Electrophysiological assessment. Exp Brain Res. 1980;40:294–304.[ISI][Medline][Order article via Infotrieve]
20. Chino YM, Shansky MS, Jankowski WL, Banser FA. Effects of rearing kittens with convergent strabismus on development of receptive-field properties in striate cortex neurons. J Neurophysiol. 1983;50:265–286.[Abstract/Free Full Text]
21. Baker F, Grigg P, von Noorden G. Effects of visual deprivation and strabismus on the response of neurons in the visual cortex of the monkey, including studies of striate and prestriate cortex in the normal animal. Brain Res. 1974;66:185–208.
22. Crawford MLJ, von Noorden GK. The effects of short-term experimental strabismus on the visual system in Macacca mulatta. Invest Ophthalmol Vis Sci. 1979;18:496–505.[Abstract/Free Full Text]
23. Movshon JA, Eggers HM, Gizzi MS, Hendrickson AE, Kiorpes L, Boothe RG. Effects of early unilateral blur on the macaque’s visual system. III. Physiological observations. J Neurosci. 1987;7:1340–1351.[Abstract]
24. Logothetis NK, Pauls J, Augath M, Trinath T, Oeltermann A. Neurophysiological investigation of the basis of the fMRI signal. Nature. 2001;412:150–157.[CrossRef][Medline][Order article via Infotrieve]
25. Imamura K, Richter H, Lennerstrand G, et al. Reduced activity in the extra-striate visual cortex of individuals with strabismic amblyopia. Neurosci Lett. 1997;225:173–176.[CrossRef][ISI][Medline]