-
Views
-
Cite
Cite
Kristi A. Clark, Turid Helland, Karsten Specht, Katherine L. Narr, Franklin R. Manis, Arthur W. Toga, Kenneth Hugdahl, Reply: Cortical differences in preliterate children at familiar risk of dyslexia are similar to those observed in dyslexic readers, Brain, Volume 138, Issue 9, September 2015, Page e379, https://doi.org/10.1093/brain/awv037
Close - Share Icon Share
Extract
Sir,
We thank Indra Kraft and colleagues (2015) for their insightful letter commenting on our recent report in Brain. We agree with the authors in their assessment that, while longitudinal studies are invaluable for disentangling cause and effect in neurodevelopmental disorders, such studies are unfortunately rare because of many practical difficulties. And because of such difficulties, longitudinal studies such as ours often end up with final sample sizes that are relatively low compared to cross-sectional studies. As the authors correctly identified, one of the risks of studies with a smaller sample size is that there is an increased chance of type II errors. One way to ameliorate this is to examine effect sizes in conjunction with statistical significance. In our study, the effect sizes of pre-reading differences were remarkably large (effect sizes >2, calculated as Cohen’s d, reported in the Supplementary material). For example, the average pre-reading cortical thickness of a region in Heschl’s gyrus differed by >0.5 mm between children who developed dyslexia and those who did not (dyslexia = 2.62 mm ± 0.18 mm, control = 3.15 mm ± 0.28 mm; values are mean ± standard deviation), even though the region itself was identified through a whole-brain analysis using a rather liberal threshold of α < 0.05.
