Neural plasticity in adults with amblyopia (Vernier acuityylearningyorientationyhyperacuityypractice)

Amblyopia is a neuronal abnormality of vision that is often considered irreversible in adults. We found strong and significant improvement of Vernier acuity in human adults with naturally occurring amblyopia following practice. Learning was strongest at the trained orientation and did not transfer to an untrained task (detection), but it did transfer partially to the untrained eye (primarily at the trained orientation). We conclude that this perceptual learning ref lects alterations in early neural processes that are localized beyond the site of convergence of the two eyes. Our results suggest a significant degree of plasticity in the visual system of adults with amblyopia. Amblyopia is a developmental disorder that occurs during a period of neural plasticity early in life (1). The consequences of amblyopia include a reduced complement of cortical neurons that can be driven through the amblyopic eye and reduced visual acuity (2, 3). Generally, treatment of amblyopia is only undertaken in infants and young children; however, recent studies suggest that repetitive practice can improve performance on a variety of visual tasks in adult humans with normal visual capacities. The improvement in performance that follows practice may be quite specific to the learned orientation (4–8), retinal location (5, 7–11), spatial frequency (5, 8), and direction of motion (12). Here we report that adult amblyopes also demonstrate substantial and significant perceptual learning of Vernier acuity, and that this learning reflects alterations in early neural processes, perhaps due to sharpening of neural responses (13), that are taskand orientation-specific, but are localized beyond the site of convergence of the two eyes. To determine the limits of cortical plasticity in humans adults with naturally occurring amblyopia, we studied Vernier acuity. Amblyopes have marked deficits in Vernier acuity that are highly correlated with their loss of Snellen acuity (14), and improvement in Vernier acuity in normal vision has been suggested to be linked to sharpening of neural responses (13) in the visual cortex. Our experiment consisted of three phases: (i) pretraining measures of Vernier and detection thresholds in each eye for several line orientations, (ii) a training phase where each observer repetitively trained on the Vernier task at a specific orientation until they had completed 4000–5000 trials, and (iii) posttraining measures (identical to the pretraining measures). Because we were interested in perceptual learning, as opposed to simply learning the psychophysical technique, or learning a strategy for making psychophysical observations with an amblyopic eye, five of the six observers had previous experience in making Vernier and detection judgments using our signal-detection methods (14). RH, RJ, and BJ had several years of experience (and hundreds of thousands of trials); KW and FG had less previous experience. Only EW had no prior experience. One potential problem with testing highly experienced observers is that they may have already improved to their limit (15) through many trials with feedback. Therefore, we trained our previously experienced observers with oblique lines, since all their previous experiments had been with horizontal or vertical targets. Normal observers, with extensive experience with horizontal and vertical targets, improve with oblique targets (16).