发展性阅读困难是一种常见的学习障碍，具有较高的发生率，不仅会造成学龄儿童学业成绩落后，而且会影响个体的心理和毕生发展。拼音文字阅读困难的研究开展较早，大多数研究者认为拼音文字阅读困难的主要缺陷在于语音加工缺陷。汉字是一种表意文字，不存在形素-音素对应（G-P-C规则），汉语阅读困难的主要原因仍待研究。无论是哪种语言系统，以往研究的结果都存在着较大的差异性，这提示我们阅读困难可能是一个异质性的群体，因此阅读困难群体的异质性问题得到了越来越多的研究者的重视。综观以往研究发现，汉语发展性阅读困难的研究主要是从认知行为角度进行的，对其神经机制和脑结构的研究较少。此外，阅读困难群体异质性的研究多采用行为测验和统计模型相结合的模式，虽然将阅读困难区分为不同的群体，但是不能回答不同的群体应该如何干预的问题。干预-应答模式不仅可以将阅读困难区分为不同群体，还能够将阅读困难的甄别与干预相结合，近年来得到了研究者较多的关注。本研究首先比较了汉语发展性阅读困难儿童与正常儿童的脑结构差异，然后基于汉语阅读和书写的三角模型，通过干预-应答模式将阅读困难群体区分为形-音干预敏感组和形-义干预敏感组，进而从脑结构的角度考察汉语发展性阅读困难群体的异质性问题。VBM分析发现，正常儿童左侧颞中回（BA21）、左侧额上回（BA32）和右侧额上回（BA9）灰质体积大于阅读困难组，阅读困难组在双侧丘脑、左侧缘上回（BA48）、右侧角回（BA39）灰质体积大于正常组，表明阅读困难儿童可能在形音和形义加工脑区存在功能缺陷。正常组与形-音干预敏感组的比较发现，形-音干预敏感组在左侧额上回（BA32）灰质体积小于正常组，在右侧中央后回（BA3）、右侧颞极（BA38）和右侧颞中回（BA20/37）灰质体积大于正常组。相关分析也发现，形-音组右侧颞极灰质体积与识字量存在显著相关。因此我们推论形-音干预敏感组可能存在形音加工区域结构异常，使用右侧语义加工区进行了功能代偿。形-义组右侧额上回灰质体积显著小于正常组，而右侧颞中回（BA 22）灰质体积显著大于正常组，表明形-义组可能在语义加工区结构异常，使用右侧颞中回弥补了左侧语义加工区的激活不足。我们发现形-音干预敏感组、形-义干预敏感组与正常组的灰质体积差异区域并不相同。ROI分析结果显示，在形-音组灰质体积大于正常组的右侧颞极（BA38）和右侧颞中回（BA22）,形-音组与形-义组存在显著差异。因此，我们认为，汉语发展性阅读困难群体存在异质性，这种异质性不仅表现在行为水平，也存在着脑结构基础。

Developmental dyslexia is a common kind of learning disability, which has a high incidence in school-aged children. Research on alphabetic dyslexia started earlier, most researchers agree that the main deficit of alphabetic dyslexia is phonological processing deficit. Chinese character is an ideographic system, GPC rules do not exist , and the main deficit for Chinese dyslexia is still under investigation. No matter what kind of language system, there are a lot of inconsistency in previous studies. It suggests that dyslexic population is a heterogeneous group, therefore researchers began to pay more attention on the heterogeneity of dyslexic group. Previous studies on Chinese developmental dyslexia were mainly conducted from the perspective of cognitive behavior, and the studies about dyslexic abnormal neural mechanisms and atypical brain structures were rare. In addition, most of studies on the heterogeneity of Chinese dyslexia took the combination method with behavior performance tests and statistical models. This method has some subjectivity and limitations. Our research compared brain structural differences between Chinese developmental dyslexia and normal children, and then based on the triangular model of Chinese reading and writing , we employed responsiveness to intervention(RTI) method to divide dyslexic subjects into two groups: Orthography - Phonology training sensitive group and Orthography - Semantic training sensitive group. VBM analysis showed that normal children’s gray matter volume were larger than children with dyslexia in left middle temporal gyrus (BA21), left superior frontal gyrus (BA32) and right frontal gyrus(BA9), but in bilateral thalamus, left supramarginal gyrus (BA48) and right angular gyrus (BA39) the gray matter volume of dyslexic children was larger than that of normal group, suggesting that Chinese dyslexic children may exist Orthography – Phonology deficit and Orthography – Semantic deficit. Significant differences of gray matter volume between normal group and Orthography – Phonology group were revealed in left superior frontal gyrus (BA32), right postcentral (BA3), right temporal pole (BA38 ) and right middle temporal gyrus (BA20/37). Correlation analysis further found that gray matter volume of right temporal pole was significantly related to the vocabulary size in Orthography – Phonology group. We speculate that there may be a phonological deficit in Orthography – Phonology group. Gray matter volume of Orthography – Semantic group in the right superior frontal gyrus (BA9) was smaller than control group, while the gray volume matter in right middle temporal (BA22) was significantly larger than normal readers. Therefore we infer that there is a semantic processing deficit in Orthography – Semantic group, and they may use the right middle temporal gyrus as semantic processing compensation. ROI analysis also found that, gray matter volume of the right temporal pole (BA38) and right middle temporal gyrus (BA22) , in which the Orthography – Phonology group are larger than normal group, are significantly different between the two dyslexic groups. Therefore, we infer that the Chinese developmental dyslexia is heterogeneous, and this heterogeneity is not only reflected at the behavior level, but also encompass the basis of brain structure.