Furth(1973)曾写道，对于听障儿童来说，“学校的所有课程都变成了语言课，课堂变成了显示听障儿童语言技能有多么贫乏，而不是显示其有多么聪明的场合”。听障儿童的阅读能力相对于健听儿童存在劣势，这严重影响了他们的学业成绩。为了帮助听障儿童更好地适应学业生活，提升其阅读水平刻不容缓。我国对听障儿童阅读能力的实证研究较少并且主要集中在词汇阅读理解领域，考虑到小学是阅读发展的关键时期、阅读能力包含多个方面（维度：理解和流畅性；层次：词汇和句子）以及听障儿童个体差异较大，本研究将首先从发展的角度，探究不同听觉经验条件下，小学听障儿童各维度和各层次阅读能力的发展趋势以及其相对于健听儿童的迟滞程度。

良好的阅读能力需要满足两个条件：掌握一门语言与建立这门语言和文字之间的映射关系。所有的书面语均由口语编码而成，语言和文字之间的映射主要指语音和字形之间的关联，建立音-形映射之后，健听儿童就可以使用他们的口语知识来促进自己的阅读。语音意识（无意识地识别和操纵语音结构）直接影响音-形映射，大量的心理学和语言学研究表明，在字母文字中，语音意识是阅读发展的必要条件。在汉字中，语音意识也是早期阅读的重要组成部分。听障儿童无法获得与健听儿童相同的听觉基础，但在阅读学习中仍然需要建立音-形映射，即使第一语言为手语的听障儿童，似乎也需要借助手指字母建立视觉语音和字形的关联，由此我们可以推断，语音意识在听障儿童的阅读学习中也起着重要作用。以往对于二者关系的研究结果并不一致，考虑到被试的个体差异（听觉经验）、汉字不同于字母文字的语音结构特点（声调、声母和韵母）以及阅读能力的维度和层次，本研究将系统考察不同听觉经验的条件下，汉语听障儿童的各项语音意识对其各项阅读能力的独特作用。

语音加工是对语音的心理表征，包括了语音加工能力和语音加工过程。语音意识是最重要的语音加工能力，阅读障碍人群的研究发现，造成语音意识缺陷的原因可能是语音加工过程中不准确或者模糊的语音表征(Elbro & Jensen, 2005)。听障儿童的听力损伤导致了输入声音的失真或缺失，声音信息的不完整阻碍了将声音结构分解成语音序列的表征过程。由此可以推断，听障儿童的语音意识缺陷也可能同样源于语音表征的不完整。语音加工过程就是进行语音表征的过程，语音加工过程中任何一个环节出现表征困难都会影响语音意识的发展。不管是出于了解加工机制的想法，还是干预的角度，细致研究听障儿童语音加工的各个环节，挖掘其具体的受损过程都十分必要。

本研究将系统考察听障儿童语音加工对于阅读能力的影响。研究一首先全面地对汉语听障儿童的阅读发展水平进行了分析。研究二进一步探讨不同语音意识对于各层次和各维度阅读能力的影响。最后，研究三和研究四关注语音意识背后的语音加工过程，探讨听障儿童语音意识影响阅读的背后机制。

研究一探究了三类儿童（无听觉经验听障儿童、有听觉经验听障儿童和健听儿童）在小学阶段（低年级和高年级）各项阅读能力（词汇理解、句子理解和句子流畅性）的发展趋势。方差分析结果显示，三组被试的词汇理解、句子理解和句子流畅性均随着年级升高呈现出显著提升。听障儿童的阅读能力随着年龄和教育经验的增长而增长。在各年级段上，不管是否有听觉经验，听障儿童的阅读能力整体上均落后于同年级段的健听儿童，并且随着年级升高，差异逐渐增大，表现出马太效应。两组听障儿童之间，词汇理解和词汇流畅性在两个年级段上均没有显著差异。句子理解在两个年级段上均显示有听觉经验儿童的表现更好。句子流畅性从低年级的无差异到高年级的有听觉经验儿童表现更好，也表现出了一定的马太效应。

研究二探究了语音意识（声母意识和声调意识）在三组被试阅读中的影响。将生理年龄、一般认知能力和正字法意识作为控制变量，无听觉经验听障儿童、有听觉经验听障儿童和健听儿童三组被试的语音意识均对阅读有独立预测作用，但其作用存在着解释率和解释机制的不同。语音意识在两组听障儿童阅读能力中起的作用更大，并且其中起主要作用的是声母意识。声母意识主要影响无听觉经验听障儿童的流畅性，而对于有听觉经验听障儿童的理解和流畅性均有作用。语音意识对健听儿童的解释率较少，但声调意识的作用仍然显著。

研究三探究了人工耳蜗植入儿童对词汇层面声调、声母和韵母的失匹配反应。研究采用Oddball范式，声调对比、声母对比和韵母对比的刺激材料分别是：/da2/- /da4/、/ba1/-/ta1/和/ti2/-/tu2/，前面是标准刺激，后面是偏差刺激。对差异波进行组别（人工耳蜗植入儿童和健听儿童）╳位置（额部、额中部和中部）╳波形（P-MMR和MMN）的方差分析。结果显示，两组被试在三组条件下均产生了差异波(MMR)，感知到了语音差异。被试的组别差异主要体现在峰值潜伏期上。声调对比和声母对比条件下，与健听儿童相比，人工耳蜗植入儿童的潜伏期更长，这表明在辨别语音时可能存在时间延迟。在韵母条件下，健听儿童表现出P-MMR潜伏期显著长于MMN，人工耳蜗植入儿童组两个波形亚组之间没有发现潜伏期差异。在波幅绝对值上则没有发现组别差异，仅在声母条件下发现了位置差异，中部的波幅显著小于额部和额中部。

研究四探究了人工耳蜗植入儿童对音位层面声调、声母和韵母的失匹配反应。研究采用Oddball范式，声调对比、声母对比和韵母对比的刺激材料分别是：/T2/- /T4/、/b/-/t/和/i/-/u/，前面是标准刺激，后面是偏差刺激，对差异波进行组别╳位置╳波形的方差分析。结果显示，三组条件下两组被试在所有时间窗和大脑位置上均产生了明显的差异波(MMR)，对比研究三，人工耳蜗植入儿童在音位层面的语音辨别能力要更强一些。与研究三的组别主效应不同，两组被试的MMR在潜伏期和波幅上没有显著差异，其差异主要体现在位置和波形上。综合研究三和研究四的结果，可以发现，在音位层面上，人工耳蜗植入儿童的语音辨别能力达到了健听儿童的水平，而在词汇层面上还存在着弱于健听儿童的部分。语音意识（声调意识、声母意识和韵母意识）主要与词汇层面MMR的潜伏期和波幅显著相关，与音位层面MMR的相关基本不显著。词汇层面的语音表征困难导致了语音意识的困难。

综合四个研究的结果，可以发现，不论听障儿童是否具有听觉经验，他们的各项阅读能力均显著滞后于健听儿童，这其中落后的语音意识起到了重要的作用。而进一步地，语音加工和阅读缺陷根源于听障儿童的生理缺陷。听障儿童的语音输入通道受损，受损的通道输入的声音信息也是受损的，声音在音位和词汇两个层面的语音表征也存在不同程度的损伤，语音表征、特别是词汇层面的表征不完整导致了语音意识的缺陷，而语音意识的缺陷影响了音-形对应规则，最终导致了阅读困难。这一结果很好地解释了为什么大量语音加工能力干预训练效果不理想以及随班就读听障儿童在小学高年级出现“回流”。本研究的结果可以帮助人们更好地开展听障儿童阅读教学干预工作。在听障儿童的康复干预中，首先应努力提高听障儿童的听力水平，增多其听觉经验。其次增加听觉和视觉的语音干预教学，不仅注重语音意识，尤其是声母意识的训练，还应该对语音加工中的各个环节进行动态的、个性化的精准干预。

In his book, Furth (1973) wrote, "all school subjects turn into language lessons and become occasions for showing how little linguistic skill deaf children have rather than occasions for showing how intelligent deaf children can be." As the reading abilities of deaf students lag far behind those of their hearing peers, helping these students develop their reading skills is an urgent matter.

To become readers, two capabilities are required: mastering a language and establishing the mapping relationships between sound and print. As all written language is encoded by spoken language, once students understand the underlying principles of the sound-print mapping, they call upon their spoken language to facilitate reading. Phonological awareness (the passive, unconscious recognition and manipulation of speech structures) is believed as a direct influencing factor, effect on spelling-sound correspondences, in support of this assumption, numerous of research in hearing students has shown that phonological awareness plays a dominant role in the early phases of reading in alphabetic languages and is an important one in Chinese. Deaf students haven’t adequate access to the auditory basis though, they do need to establish the sound-print mapping when learning to read, even for the deaf signers whose first language is sign language. We infer that phonological awareness has a great influence on reading of deaf. In past research, findings on the relations of the two in deaf readers were mixed. This study will systematically investigated the unique effects of phonological awareness on various reading abilities of Chinese deaf children under different auditory experiences.

Phonological processing, including the phonological processing ability and the phonological processing process, is the mental representation of phonology. Phonological awareness is the most important phonological processing process, studies on dyslexia have found that phonological awareness deficit may be caused by the inaccurate or fuzzy representation in the process of phonological processing. The detrimental auditory system of deaf students leads to the distortion or absence of input sound, and the incomplete sound information impairs the representation process of decomposing sound structure into speech sequences. Therefore, deficits in phonological awareness may result from the incomplete representation as well. The phonological processing process is the process of phonological representation, problems in one or more levels in the process of phonological processing would hinder the development of phonological awareness. Whether for the idea of understanding mechanism or the perspective of intervention, each link of phonological processing should be carefully explored to find the deficit part.

In the current study, we systematically investigated the relations between phonological processing and reading. First, study 1 was conducted to comprehensively analyzing Chinese deaf students’ reading development. Further, the independent contributions of phonological awareness to reading was examined in study 2. And finally using the ERP technique, study 3 and study 4 focussed on the phonological processing behind phonological awareness, to explore the mechanism behind the influence of phonological awareness on reading comprehension.

Study 1 aimed to investigate the developmental tendency of reading (word comprehension, sentence comprehension and sentence fluency) among three groups of students (deaf students without auditory experience, deaf students with auditory experience, and hearing students) in their primary school period (lower and higher grades). The results of variance analysis indicated that word comprehension, sentence comprehension and sentence fluency significantly improved gradually with grade among all three groups. Deaf students’ reading increased with age and educational education. In all grade levels, regardless of whether they have auditory experience or not, the reading abilities of deaf children lag behind that of hearing children in the same grade, and the gap was larger in higher grade, showing Matthew effect. Among two deaf students, word comprehension and word fluency showed no significant difference in either grades, while deaf students with auditory experience performed better in both grades in sentence comprehension, sentence fluency ranged from no difference in lower grades to better performance in higher grade.

Study 2 aimed to investigate the relative importance of phonological awareness (consonant awareness and tone awareness) in predicting four reading abilities. After controlling for chronological age, general cognitive ability and orthographical awareness, hierarchical regression analysis showed that the phonological awareness among three groups all had independent contributions to reading, however, the explanatory rate and mechanism of these effects were different. Phonological awareness, especially consonant awareness, accounted more variance in two deaf students. The unique effect of consonant awareness was significant on the reading fluency in deaf students without auditory experience, whereas, its predictive role was played on reading fluency and reading comprehension in deaf students with auditory experience. To a lesser degree, however, hearing students do rely on tone awareness to reading.

Study 3 aimed to examine in phonemic level, the mismatch responses to tone, consonant, and vowel in children with cochlear implant. The oddball paradigm was employed, and the stimulus materials of tone condition, consonant condition, and vowel condition were /da2/ vs /da4/, /ba1/ vs /ta1/ and /ti2/ vs /tu2/ respectively, with the former stimuli were standards, and the latter ones were deviants. And three way ANOVAs with group (children with cochlear implant and children with normal hearing) as a between-subject factor and with site (frontal, frontocentral, and central) and waveform (P-MMR and MMN) as within-subject factors were conducted for each stimulus type.

Our results indicated that MMR were observed in all three conditions, thus both cochlear implant children and hearing children could discriminate the phonological differences. Group difference of MMR was mainly found in peak latency. Under the tone contrast and consonant contrast condition, children with cochlear implant exhibited longer MMR latencies compared with hearing students, indicating a time delay in discriminating speech. Under the vowel contrast condition, the latencies of P-MMR sub-group were found longer than MMN sub-group in cochlear implant children, however, no differences was found between the two waveform sub-groups in hearing children. Only site differences were observed in MMR amplitude of consonant condition, the elicited MMR amplitude was larger than the other two sites.

Study 4 aimed to examine in lexical level, the mismatch responses to tone, consonant, and vowel in children with cochlear implant. The oddball paradigm was employed, and the stimulus materials of tone condition, consonant condition, and vowel condition were /T2/ vs /T4/, /b/ vs /t/ and /i/ vs /u/ respectively, with the former stimuli were standards, and the latter ones were deviants. And three way ANOVAs with group, site and waveform were conducted for each stimulus type. Our results indicated that MMR were observed in all time windows of three conditions. In contrast to study 3, cochlear implant children showed better phonological discrimination at the phonemic level. Different from the group main effect in Study 3, there was no significant difference in the latency and amplitude of the MMR between the two groups, and the differences were mainly reflected in the site and waveform. Combining the results of Study 3 and Study 4, we found that the speech discrimination ability of children with cochlear implant reached the level of children with normal hearing at the phonemic level, while there were weaker parts than children with normal hearing at the lexical level. Phonological awareness was mainly significantly correlated with MMR at the lexical level, therefore, difficulties in the process of phonological processing at the lexical level lead to difficulties in phonological awareness. The results provide a reference for precision intervention in the future.

In conclusion, regardless of the auditory experience, deaf students’ reading lag far behind those of hearing students of the same grade level, in which the backward phonological awareness plays an important role. Furthermore, phonological processing and reading deficits are rooted in the physical defects. Because of the damaged auditory channel, the inputted sound massages are impaired, the phonological processing at the two process levels also has different degree of impairment, thus leading to the phonological awareness deficits that affect their spelling-sound correspondences, resulting in their reading failure.

These results provide a fine explanation for the unsatisfactory outcomes of many phonology interventions and the refluxed problem of deaf students in regular class. In the future, phonology intervention for deaf children should not only focus on the training of phonological awareness, but also be more detailed to each level of phonological processing.