When a spatiotemporal broadband stimulus of suprathreshold contrast was used, our amblyopic subjects, as a group, did not show enhanced or even normal activity for amblyopic eye stimulation in area V1, contradicting a recent report.
26 The responses of amblyope OA
(Fig. 4C)show a localized enhancement of activation in V1/V2d, but this was not a common finding. Generally, we found small but consistent
reductions in activation in area V1 when driven by the amblyopic eye compared with that of the fellow fixing eye. Small but consistent differences also occur between the V1 response when driven by the fellow fixing eye of amblyopes compared with the dominant eye of normal subjects but not, as far as we could see, between the eyes of normal observers. As a group, these differences become significant in areas V1, V2, V3, Vp, and V3a. This suggest that V1 responds abnormally when driven by either the fellow fixing eye or the amblyopic eye, although the reduced activation is greater for the latter. Some amblyopes within our sample exhibited greater reductions than others in their V1 response, but we found no correlation with the contrast detection loss across our amblyopic group. It should be pointed out that our approach was one that would tend to give a conservative estimate of the magnitude of reduced functional activation in amblyopia: we completely covered the nonstimulated eye, thereby excluding any suppressive influences; and to reduce the intersubject variability, we compared amblyopic responses to those of the fellow fixing eye in the same subjects, notwithstanding the possibility that the fellow fixing eye might itself be slightly reduced in its activation. We choose the two key psychophysical measures that have been shown to capture the variability in the visibility loss in a large sample of human amblyopes,
59 —namely, the grating acuity and contrast sensitivity. Neither measure strongly correlated with the reduced functional activation in V1 or any other extrastriate area. This finding is at odds with conclusions in two previous studies
30 34 that the reduced functional activation in amblyopia to spatiotemporal broadband stimuli, similar to that used in the current study, correlates with the psychophysical loss. However, it is consistent with two previous studies
26 33 in which spatiotemporal narrowband stimuli were used and showed no group correlation with either of these two key psychophysical measures. We also show that the reduced activation selectively involves central field stimulation, a finding recently reported by Conner and Mendola
36 and Muckli et al.,
26 who report enhanced amblyopic peripheral responses. Our finding of a larger activation deficit for central vision may be because of visual field locus per se or because the deficit selectively involves the processing of higher spatial frequencies that are represented only in the more central parts of the field and selectively stimulated by our radial checkerboard stimulus. What is surprising if this latter explanation is correct is that there is not a better correlation between the functional central field deficit and psychophysics (i.e., grating acuity). Amblyopes exhibit larger saccadic fixation eye movements than do normal subjects, but we do not believe that these played a significant role in the reduced activations that we report. First, it has been shown that eye movements less than 3° have little effect on fMRI response,
60 61 and none of our amblyopes had fixation eye movements exceeding 1.8°. Second, it is known that the magnitude of the eye movement abnormality in amblyopia correlates highly with the visual acuity
62 63 yet our reduced activations did not show any significant correlation with visual acuity for either full field or central stimulation. All our subjects had strabismus, but five also had anisometropia. Although we did not observe any difference in the responses of subjects with or without anisometropia, a much greater sample would have been needed to resolve the question.