Brain-behavior relations in reading and dyslexia: Implications of Chinese results

It is ironic that our recent report on the neural correlatesof Chinese dyslexia (Siok, Perfetti, Jin, & Tan, 2004) raisesa concern by Ziegler (2006) that this report could under-mine an agreed-upon conclusion favoring a phonologicaldeficit as the cause of reading disability. Our past researchover 20 years in English (Perfetti, 1985) and Chinese (Per-fetti, Liu, & Tan, 2005; Pollatsek, Tan, & Rayner, 2000;Tan & Perfetti, 1998) has argued that phonological pro-cesses are intrinsic to word reading and universal acrosswriting systems, and we certainly have no reason to mini-mize the importance of these processes for understandingreading disability. We address Ziegler’s concern by makingtwo observations. (1) The Siok et al. evidence supporting arole for the left middle frontal gyrus (LMFG) instead ofthe left posterior temporoparietal regions found in alpha-betic research does not undermine the consensus on theuniversality of phonology. Instead this evidence shows thatthe implementation of phonology depends on language andwriting systems, a conclusion also reached through behav-ioral evidence (Perfetti et al., 2005). (2) The detailed neuralbasis of dyslexia remains an open question, rather thanaccepted wisdom.On the first point, Siok et al. did not conclude that pho-nology plays no role in Chinese reading disability and Zie-gler’s critique recognizes this. Indeed, the evidence in favorof a phonological basis for dyslexia in alphabetic writinghas accumulated impressively over 20 years, from a timeat which the prevailing views emphasized visual problemsas the cause of dyslexia. This evidence comes from neuro-biological (Brunswick, McCrory, Price, Frith, & Frith,1999; Eden et al., 2004; Price & Mechelli, 2005; Richards,Aylward, & Berninger, 2006; Shaywitz et al., 1998; Templeet al., 2003), cognitive (Bruck, 1992; Schatschneider,Fletcher, Francis, Carlson, & Foorman, 2004; Stanovich& Siegel, 1994; Vellutino, Fletcher, Snowling, & Scanlon,2004), and genetic perspectives (Taipale et al., 2003). Themore biologically specific form of the phonological hypoth-esis is that dyslexia is universally associated with functionaldisruption of the left temporoparietal brain regions(Paulesu et al., 2001). It is this hypothesis that is falsifiedby Siok et al. (2004).To summarize briefly, using homophone and lexicaldecision tasks, we found that reading difficulty in Chi-nese is associated not only with a weakness in mappingof a character’s orthography to pronunciation (as mea-sured by the homophone decision task), but also witha substandard connection between orthography andmeaning (as indexed by the lexical decision paradigm).Importantly, these two deficits were characterized by dys-function of the left middle frontal gyrus. The left tempo-roparietal brain regions associated with disabled readingin English and other alphabets were not involved in thereading of Chinese children. Thus, either Chinese readingdisability does not have a phonological component orthe phonological component in Chinese does not involvethe left temporoparietal regions. Because the Siok et al.behavioral data showed weakness in phonological pro-cessing, the conclusion is that this weakness is associatedwith some other neural substrate. The left middle frontalgyrus, which was under activated in Chinese dyslexics, isa likely candidate.However, this is not the end of the story, but the begin-ning. The research goal becomes to understand the func-tion of the LMFG in reading. It is not obvious that theLMFG, whatever its more general role in cognitive func-