汉语语言中枢的多模态技术研究及其在外科手术定位中的应用

发布时间：2018-06-09 03:12

本文选题：语言 + 汉语　；参考：《复旦大学》2013年博士论文

【摘要】：近年,人类对脑功能的研究越来越重视。2013年初,美国和欧盟都推出了“人脑计划”,希望通过10年绘制出人脑活动图。语言是人脑独有的高级认知功能。人类的一切心理活动都离不开语言。当前,脑病变患者对医疗质量的要求日益提高,已不再是一味地追求生存,而是更加注重有尊严有质量的生活。由于存在个体差异和受病变所致的脑功能移位、重组(reorganization)或塑形(plasticity),神经外科手术借助传统解剖标志来定位脑语言功能区,不仅不准确,而且可误脑语言皮质和语言网络,导致术后病人失语及认知障碍。因此,迫切需要寻找效的技术以实现语言区的精确定位。然而,既往绝大多数语言定位的理论都是基于英语等拼音字母语言的研究结果,迄今,对汉语的语言定位仍缺乏系统的研究。因此,通过探索语言定位的新技术以及汉语语言分布和语言网络既有助于拓宽现有的语言定位手段,又可以从宏观上了解汉语语言的定位和分布特点。本论文基于多模态的技术手段(术中皮质电刺激、任务态功能磁共振、静息态功能磁共振、弥散张量成像、术中磁共振),在国人汉语人群中进行了一系列探索和研究。第一部分通过55例患者的功能预后证实了唤醒麻醉下术中皮质电刺激(金标准)定位语言中枢的可靠性,这是后面两部分研究的立论基础；第二部分提出了3种语言定位新技术和新方法,并经金标准验证,实现了技术创新；第三部分构建了基于功能磁共振和电生理的汉语语言分布图和语言网络,实现了理论创新。第一部分：3T iMRI环境下唤醒麻醉联合术中语言皮质定位技术在语言区脑胶质瘤手术中的应用唤醒手术和术中MRI两种技术已被广泛应用于临床,其应用的可靠性和有效性并分别得到证实。可是,两种复杂技术联合应用的经验和报道却比较有限。我们在论文中报道了本单位在术中MRI环境下进行唤醒手术的经验和方法并且定量的评估了对语言区肿瘤患者的功能预后以及切除程度的影响。共55例(从2010年12月-2012年12月)语言区的脑胶质瘤患者纳入本研究。切除程度的评估通过体积分析法实现,语言的评估采用汉语失语症检查表。手术中MRI扫描采用局部铺巾法,语言皮质定位采用自行研制的唤醒麻醉脑功能刺激设备(专利号：201220303939.1)。结果发现,共有22名患者(40%)在术后1周出现近期语言障碍,只有2名患者(3.6%)出现永久性(术后6月)语言障碍。由于术中MRI的参与,55名患者肿瘤的全切率从34.5%增加到63.6%。其中共有23名(41.8%)患者在术中MRI扫描后获得进一步的切除。中位切除程度从91.6%(范围45.7%-99.0%)增加到100%(范围74.0%-100%),并具有统计学意义(P0.01)。局部铺巾法不仅达到无菌手术操作要求,而且能够满足术中MRI扫描要求和克服唤醒麻醉的气道管理难题。没有发生与铺巾法有关的并发症。脑功能刺激装置显著提高了唤醒手术的工作效率。本部分结果表明1)联合术中MRI以及语言皮质定位两项技术应用于语言区肿瘤手术有助于语言功能的保护的同时提高全切率；2)基于大样本的病例预后随访证实唤醒麻醉下的术中语言皮质定位是可靠的。第二部分:汉语语言中枢定位的技术创新术中语言皮质定位技术是语言定位公认的金标准。然而,该技术具有下列不足,一是无法用于术前的计划,且无法直观的显示与肿瘤的位置关系；二是只局限在骨窗有限的范围内,无法全面的显示整体语言网络；三是需要患者的配合,无法应用于儿童以及语言障碍或不合作患者。因此,迫切需要寻找和探索新的语言定位手段以弥补现有定位手段的不足。本部分基于三种不同的影像学模态(弥散张量成像,任务态fMRI,静息态fMRI),分别针对外科手术目前应用的不足和缺陷,提出了三种语言定位的新技术和新方法。第一种技术,将弓状束纤维示踪技术应用于额叶皮质的定位,并通过金标准对其定位的准确性进行验证,定位的灵敏度是70.0%,特异度是97.6%,如果将弓状束周围1厘米作为标准,灵敏度和特异度分别为90.0%和96.0%,特异度未明显下降的前提下灵敏度得到极大的提高。第二种技术,我们将任务态fMRI与术中MRI结合,首先提出唤醒麻醉下术中功能磁共振技术,克服了术前fMRI的脑移位问题。本论文探索了该技术定位的可行性,并在个体病例上通过电生理结果验证了其可靠性。第三种是方法学的创新,独立成分分析法进行静息态fMRI分析的最大难题就是需要人为的判断各个成分,我们引入了d’(d-prime)的计算方法成功的实现了分别基于组水平和个体水平的语言成分的筛选,同时将该技术应用于肿瘤患者进行语言区定位,初步证实了该方法的可行性。第三部分:基于功能磁共振和电生理的汉语语言皮质分布图和语言网络的建立越来越多的证据表明汉语语言的处理不同英语等语言。由于语言的复杂性和缺乏动物模型,现有的汉语语言的研究基本都局限于功能神经影像的研究。神经外科是唯一可以实现与大脑“面对面”交流的学科,这也就提供了一个独特的机会通过术中语言皮质定位技术构建汉语语言的分布图和拓扑结构。本部分我们首先分别构建了脑胶质瘤患者三种任务模式(图片命名,动词产生和默读诗歌)和正常志愿者的一种任务模式(图片命名)的fMRI的语言分布图；接下来,通过将55例患者术中语言皮质电刺激的阳性位点进行合成,分别得到言语中止、命名性失语和阅读性失语的皮质分布图；最后,以电生理的阳性位点作为种子点,通过功能连接,构建基于金标准的汉语语言静息网络。本部分的结果表明1)中央前回腹侧部(BA6区的腹侧部)和额叶岛盖部(BA44)对语言的产生具有至关重要的作用。我们推断中央前回腹侧部以及岛盖部应当是语言产生的核心区域,较传统的broca区要后移；2)左侧额中回(BA9区)在汉语的电生理定位中较英语更易出现语言任务的中止,证实BA9区是汉语语言处理中的一个特殊的节点；3)汉语的语言分布范围较印欧语系更加广泛,个体差异更大；4)单一一种任务的fMRI进行语言定位仍存缺陷,联合多种任务进行定位可提高可靠性；5)基于金标准构建汉语语言静息网络是可行的,可能更接近真实的语言网络。结论本论文围绕转化医学对汉语语言定位开展了一系列的研究,主要结论和价值如下： 1、通过多模态技术构建了汉语语言分布图和网络,发现了汉语语言的特殊性与共性,为汉语语言区手术提供理论指导。 2、通过技术和方法的创新,提出了多种可行且可靠的新技术方案进行语言定位,为语言定位提供技术指导。 3、基于语言定位的复杂性,单一的定位方案并不可靠,可通过多模态技术互相验证。创新点 1、首次报道唤醒手术和术中高场强MRI在语言区肿瘤的联合应用(国内首次) 2、首次通过电生理验证了弓状束纤维示踪定位汉语额叶语言区的有效性(国际首次) 3、首次提出唤醒麻醉下术中功能磁共振(ai-fMRI)的概念并应用于临床,证实了其可行性与可靠性(国际首次) 4、首次建立基于独立成分分析法的语言网络自动筛选方法并应用于个体(国际首次) 5、首次建立具有自主知识产权的汉语语言皮质分布图(国际首次) 6、首次建立基于电生理的汉语语言静息网络(国际首次)
[Abstract]:In recent years, the research of human brain function has been paid more and more attention at the beginning of.2013. Both the United States and the European Union have launched the "human brain program", hoping to draw a human brain activity map through 10 years. Language is a unique high-level cognitive function of the human brain. All the psychological activities of human beings are inseparable from the language. It is no longer the pursuit of survival, but more emphasis on a life of dignity and quality. Because of the existence of individual differences and dislocations caused by the brain function shift, reorganization or plasticity, the Department of Neurosurgery surgery with the aid of traditional anatomical markers to locate the brain function area is not only inaccurate, but also mistaken brain language cortex And language network, resulting in postoperative patients aphasia and cognitive impairment. Therefore, there is an urgent need to find effective techniques to achieve the precise positioning of the language area. However, most of the previous theories of language localization are based on the results of the phonetic alphabet language such as English and so on. So far, the language localization of Chinese is still lacking systematic research. The exploration of the new technology of language positioning and the distribution of Chinese language and language network can not only help to broaden the existing language positioning means, but also understand the localization and distribution characteristics of the Chinese language from the macro level.
This paper is based on multimodal techniques (cortical electrical stimulation, task state functional magnetic resonance, resting state function magnetic resonance, diffusion tensor imaging, and intraoperative magnetic resonance). A series of studies and studies were carried out in Chinese people. The first part confirmed the cortical electrical stimulation (gold) during the wake anaesthesia in 55 patients. Standard) positioning the reliability of the language center, which is the foundation of the last two parts of the study; the second part puts forward 3 new techniques and new methods for language positioning, and realizes technical innovation through the verification of gold standard. The third part constructs a Chinese language distribution map and a language network based on functional magnetic resonance and electrophysiology. The theory has realized the theory. Innovation.
Part I: application of awaking anesthesia combined with intraoperative localization of language cortex in 3T iMRI brain glioma surgery
The two techniques of wake-up operation and intraoperative MRI have been widely used in clinical practice, and the reliability and effectiveness of their applications are confirmed. However, the experience and reports of the combined application of the two complex techniques are limited. In this paper we report the experience and methods of the unit in the operation of the wake-up operation under the operation of the MRI and quantitative A total of 55 patients with glioma in the language area (December 2010 -2012 December) were included in the study. The assessment of the degree of resection was achieved by volume analysis, and the language evaluation adopted the Chinese aphasia check list. In the operation, the MRI scan used the local napkin method. Cortical localization was performed by a self-developed wake-up anesthetic brain stimulation device (Patent No. 201220303939.1).
The results showed that a total of 22 patients (40%) had recent language disorders at 1 weeks after the operation, only 2 patients (3.6%) had permanent (postoperative June) language barriers. Due to the involvement of MRI, 55 patients increased the total tumor rate from 34.5% to 63.6%., of which 23 (41.8%) patients received further resection after MRI scanning. The degree increased from 91.6% (range 45.7%-99.0%) to 100% (range 74.0%-100%) and had statistical significance (P0.01). The local napkin method not only met the requirements of aseptic operation, but also met the requirements of MRI scan in the operation and overcome the airway management problem of wake up anesthesia. There was no complication associated with the napkin method. Brain function stimulation device The efficiency of the wake-up operation was significantly improved.
The results of this part show that 1) combined operation of MRI and two techniques of language cortex localization in language area tumor surgery can help to protect the language function while improving the total cut rate while 2) based on the prognosis of large sample cases, it is proved that the localization of the language cortex in the operation under the wake up anesthesia is reliable.
The second part: technological innovation of location of Chinese Language Center.
The technique of language cortical positioning is the golden standard of language positioning. However, the technique has the following shortcomings, one is that it can not be used before the operation and can not display the relationship with the tumor directly. Two it is limited to the limited scope of the bone window and can not show the overall language network in an all-round way; and the three is the need for the patient's coordination, It is impossible to apply to children, language disorders or non cooperative patients. Therefore, it is urgent to find and explore new language positioning methods to make up for the shortcomings of existing positioning methods.
This part is based on three different imaging modality (diffuse tensor imaging, mission state fMRI, resting state fMRI). In view of the shortcomings and shortcomings of the current application of surgical operations, three new techniques and new methods for the localization of language are proposed. The first technique is applied to the localization of the frontal cortex by the bow beam tracing technique and through the gold standard. The sensitivity of the location is verified by 70%, the sensitivity is 70%, the specificity is 97.6%. If the 1 cm around the bow bundle is the standard, the sensitivity and specificity are 90% and 96% respectively, the sensitivity is greatly improved under the condition that the specificity does not decrease obviously. Second techniques, we combine the task state fMRI with the intraoperative MRI, first of all. The functional magnetic resonance (fMRI) technique under wake up anesthesia was proposed to overcome the problem of brain displacement of fMRI before operation. The feasibility of this technique was explored in this paper, and its reliability was verified by electrophysiological results in individual cases. The third is the innovation of methodology and the biggest problem of the rest state fMRI analysis by independent component analysis is the need. In order to judge the individual components artificially, we introduced the d '(d-prime) method to successfully screen the language components based on the group level and the individual level, and applied the technique to the tumor patients to locate the language area, and initially confirmed the feasibility of the method.
The third part: the distribution of Chinese language cortex and the establishment of language network based on fMRI and electrophysiology.
More and more evidence shows that Chinese language is handled in different languages such as English and other languages. Due to the complexity of language and the lack of animal models, the study of the existing Chinese language is limited to the research of functional neuroimaging. The Department of neurosurgery is the only subject that can meet the "face-to-face" flow of the brain. This provides a unique feature. In this part, we first constructed the language distribution maps of the three tasks of the brain glioma patients (picture naming, verb generation and silent reading) and a task model of normal volunteers (the picture named) fMRI. The positive loci of the electrical stimulation of the language cortex of 55 patients were synthesized and the cortical distribution maps of speech discontinuation, named aphasia and reading aphasia were obtained respectively. Finally, the positive loci of electrophysiology were used as seed points, and the resting network of Chinese language based on gold standard was constructed by functional connection.
The results of this part show that 1) the anterior central ventral part of the central ventral region (the ventral part of BA6 region) and the frontal lobe Island (BA44) play a vital role in the production of language. We infer that the anterior central ventral side of the central region and the island cover should be the core region of the language, and the traditional Broca region should be moved back; 2) the left middle frontal gyrus (BA9 region) is in Chinese electrophysiology. It is easier to show the suspension of language tasks than English in physical location. It is proved that BA9 area is a special node in Chinese language processing; 3) Chinese language is more widely distributed than Indo European, and the individual difference is greater; 4) a single task of the language location still has defects, combining a variety of tasks can improve reliability. Sex; 5) constructing a resting network of Chinese language based on the golden standard is feasible and may be closer to the real language network.
conclusion
This thesis has carried out a series of studies on the location of Chinese language in translational medicine. The main conclusions and values are as follows:
1, the Chinese language distribution map and network were constructed through multi-modal technology, and the particularity and commonality of Chinese language were found, which provided theoretical guidance for Chinese language area surgery.
2, through the innovation of technology and method, we put forward various feasible and reliable new technology solutions for language localization, and provide technical guidance for language localization.
3, based on the complexity of language localization, a single localization scheme is not reliable and can be verified by multi-modal technology.
innovation point
1, we first reported the combined application of wake-up surgery and intraoperative high field MRI in the language area tumors (the first time in China).
2, for the first time, the validity of arcuate fiber tracer localization of Chinese frontal lobes was verified by electrophysiology.
3, we first proposed the concept of functional magnetic resonance imaging (ai-fMRI) during awake anesthesia and applied it in clinical practice, which confirmed its feasibility and reliability (the first time in the world).
4, it is the first time to set up an automatic language screening method based on independent component analysis (ICA) and apply it to individuals (the first time in the world).
5, the first time to establish a Chinese language cortical distribution map with independent intellectual property rights (the first time in the world).
6, the first Chinese language resting network based on electrophysiology has been established for the first time.
【学位授予单位】：复旦大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：R651.1

【共引文献】