Brain, Vol. 126, No. 9, E2,
September 2003
© 2003 Guarantors of Brain
doi: 10.1093/brain/awg217
Letters to the Editor |
Visual magnocellular deficits in dyslexia
Department of Physiology, University of Oxford, Oxford, UK
Correspondence to: Chris Chase E-mail: chris.chase{at}claremontmckenna.edu
There is now much evidence that the visual magnocellular system in many dyslexics is slightly less sensitive than in controls. We were surprised therefore by the conclusion of Amitay et al. (2002
) that the dyslexics in their study did not suffer from a magnocellular visual deficit, when actually their results seem to provide quite strong evidence that their dyslexics did have such a deficit. Their opposite conclusion seems to have resulted from their strong assumptions about how M impairments should be manifested in a dyslexic population and from the way in which they scaled dyslexic performance.
They assume that a magnocellular deficit should produce a totally consistent pattern of results across a variety of different tasks designed to assess magnocellular functioning, and that any variability implies that no magnocellular deficit exists. This is unwarranted because a mild magnocellular deficit could be obscured by measurement ‘noise’ which may lead to inconsistent performance. We can estimate from their table 2 the power of their study to detect differences (Amitay et al., 2002). Our calculations show that the chances of false negatives were fairly high for many of their tasks. To achieve reasonable power (0.80) with 30 subjects and alpha set to 0.05, effect sizes would have to be at least 0.65 (one-tailed t-test). In fact, effect sizes as large as this were found on some tasks but not on others, so performance variability could have been the result of weak power and poor task specificity, rather than normal magnocellular functioning.
The authors also assumed that a magnocellular deficit only affects brief stimuli. This is incorrect. The magnocellular visual pathway has a high temporal contrast sensitivity (Merigan and Munsell, 1993), but this does not mean that it only processes rapid visual sequences, such as motion. The magnocellular channel is selectively sensitive to visual motion because the other major retinogeniculate pathway (parvo) is insensitive to temporal frequencies above 10 Hz. But the magnocellular pathway can also process low speeds and long target durations. The durations (130 ms versus 900 ms) and speeds (3.5 deg/s versus 10.6 deg/s) used in their coherent motion task were likely to be only weakly selective for magnocellular processing. Furthermore, there is no evidence we know of to show that their spatial frequency discrimination task would in fact have silenced the magnocellular pathway. While we have suggested that a visual magnocellular deficit may be part of a general developmental neuropathology affecting sensory processing in the more rapid magnocellular pathways of the visual, auditory and sensorimotor systems (Stein and Walsh, 1997), we do not conclude that only the perception of brief stimuli would be affected. Actually the first evidence for an magnocellular impairment in dyslexics came from studies of contrast sensitivity using static displays (Lovegrove et al., 1980).
How many dyslexics in their sample might have had magnocellular deficits? Figure 1 (from Amitay et al., 2002) shows that there were only six dyslexics who had magnocellular Z-scores that were well below the sample mean, but the method the authors used for normalizing the magnocellular Z-scores may have reduced the apparent size of any impairments. They chose to calculate the population standard deviation from both the dyslexics’ as well as the controls’ scores. It would have been more appropriate to use only the control data; otherwise, the normal mean is biased by including a group known to perform below average on the tasks under study. In fact, the authors did use only control data when calculating subject Z-scores for their reading tasks; we do not understand why they did not do likewise on the other measures. The resulting bias is quite evident in fig. 1; this shows that as many as five controls (17%) performed below average on the magnocellular Z-score (Amitay et al., 2002).
Thus, the difference between dyslexic and control groups was only 
1 sd using their Z-score scale constructed from the pooled standard deviation of both groups [see fig. 6 (Amitay et al., 2002)]. But if the standard deviation of only the control group had been used to scale the magnocellular Z-score, the difference would have been much greater. In fact the majority of the dyslexic sample would then fall well below the control average, suggesting a magnocellular deficit.
Finally, the dyslexics in this study were remarkably impaired. We have never seen a sample of 30 dyslexic young adults (some in university) who were on average six standard deviations below the mean of controls on reading measures, and 11 standard deviations below controls in orthographic judgment, while at the same time achieving above average performance on Similarities and Block Design subtests from the WAIS-III. The unusual severity of this sample’s reading impairment raises questions about the sample, selection procedures and the accuracy of the psychometric measures.
To summarize, when the dyslexics’ magnocellular performance is measured on a control-normalized scale, the vast majority turn out to have impairments. In addition, statistical power was weak for many of the tasks used in their study, hence the variations in subjects’ performance could be due to measurement ‘noise’ rather than absence of an magnocellular deficit.
References
Amitay S, Ben-Yehudah G, Banai K, Ahissar M. Disabled readers suffer from visual and auditory impairments but not from a specific magnocellular deficit. Brain 2002; 125: 2272–85.
Lovegrove WJ, Bowling A, Badcock D, Blackwood M. Specific reading disability: differences in contrast sensitivity as a function of spatial frequency. Science 1980; 210: 439–40.
Merigan WH, Maunsell JR. How parallel are the primate visual pathways? Annu Rev Neurosci 1993; 16: 369–402.[Web of Science][Medline]
Stein J, Walsh V. To see but not to read: the magnocellular theory of dyslexia. Trends Neurosci 1997; 20: 147–52.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||