Regarding "Eagle-eyed visual acuity: an experimental investigation of enhanced perception in autism".

o the Editor: ow-level perceptual abnormalities are increasingly seen to play an important role in some features of autistic spectrum disorder (ASD) (review [1]) by contributing to impairments f social communication through limiting higher-level visual rocessing of faces, for example. Arguably the most interesting indings are that individuals with ASD can sometimes perform etter than matched control subjects when the task involves ttention to detail (e.g., in visual search [2], finding hidden figures 3], or resisting the influence of context within illusions [4]), uggesting that ASD might be associated with enhanced processng of local information (2). A recent article by Ashwin et al. (5) as suggested that enhanced perceptual processing in ASD might nclude superior visual acuity (VA), a measure of individuals’ bility to identify symbols of a set size (5 arc min width) resented foveally at standardized viewing distances. Specifially, Ashwin et al. report mean decimal visual acuities of 2.79 in heir group of observers with ASD and 1.44 in an age-matched ontrol group (VA is typically expressed as a fraction with the umerator referring to the distance at which the subject can just dentify the letter, and the denominator the distance at which an bserver with standard VA could identify the same letter; alhough, by definition, “normal” acuity must be 1.0 [or 20/20 /6 in Snellen notation], when measured with good psychometric rocedures [6], young adults have a median acuity of 1.6 [7]) If rue, Ashwin et al.’s finding would be very important for two easons. First, as far as we are aware, this is the first report of onsistently superior VA in any clinical population (neuropsyhological or otherwise). Second, VA is generally considered imited by the earliest stages of the visual system (i.e., optical roperties of the eye, photoreceptor density) so that this result ould suggest either that: 1) this is not true, acuity is limited by ther (higher level) factors; or 2) there are structural differences n the eyes of observers with ASD. The authors consider both ossibilities by suggesting that either higher number of foveal one cells or higher numbers of dopamine receptors at the retinal r neural level could contribute to their findings. Prompted by the highly counterintuitive nature of both these onclusions and the finding that inspired them, we have invesigated the procedure employed by the authors of this study (one f us—MB—developed the computerized acuity test [Freiburg isual Acuity & Contrast Test (FrACT)] used by Ashwin et al.). We eport that although there are real behavioral differences beween ASD and control groups, technical limitations in the rocedure used to measure acuity call into question the concluion that people with ASD have higher acuity compared with naffected individuals without the context of the experiment. We begin by considering the technical details and limitations f the acuity testing system employed and then describe the pecific chain of events in the experiments reported that led to stimates of VA that—we contend—are gross overestimates of erformance in the ASD group. We conclude by offering a ypothesis as to why, if observers with ASD do not have superior cuity, they might have outperformed control subjects on this ask.