Study On The Algorithms Of ECoG Analysis Based Functional Mapping Of Brain

Resecting the lesions to maximum as well as protecting the normal brain tissues surrounding the lesions plays an important role in neurosurgery. At present, the gold-standard method of functional mapping in neurosurgery is ECS(electrocortical stimulation) which does not meet the requirements of surgery in speed and safety. Therefore study on the algorithms of intraoperative functional mapping under accurate, rapid, safe had its clinical application background and great significance. In this paper, based on the analysis of feature, the single functional and multi-functional cortex mapping have been studied step by step; the processing algorithms of ECoG(electrocorticography) have been discussed in the meantime.Firstly, it studied on motor cortex mapping with mu rhythm. According to the way that research specific functional areas by the specific rhythm, the psd analysis had been used to analyze the ECoG. Then it chose mu rhythm as the feature of ECoG on motor cortex and studied the motor functional mapping. Findings in this study indicate that the ERD(event-related desynchronization) phenomenon of mu rhythm can be used to research the mapping of motor cortex area.Secondly, it studied on language cortex mapping with high gamma. The log power of ECoG which recorded under the task state had been compared with the log power of ECoG which recorded under the rest state. Then it designed a feature to study the naming functional mapping in language cortex. The results show that the log power of high gamma band also can be used to research the mapping of naming functional in language cortex area.Finally, it studied on multi-functional areas mapping with multi-rhythms characteristics. Aiming to solve the limitation of functional areas positioning which is based on special rhythm, it proposed a new multi-functional areas mapping method which is based on multi-rhythms characteristics. In this chapter, it chose multi-rhythms characteristics as feature, and designed an algorithm which is based on wavelet analysis and support vector machine, then validated by ECoG samples which recorded from motor cortex, feeling cortex, broca cortex, naming cortex and no functional cortex area. Results show that the method successfully identified ECoG samples. The accuracy rate of algorithm is 93.12% with about 5 minutes processing time. It proved that that method is efficient and feasible in cerebral cortex functional mapping. And such method enriched the clinical application technology of intraoperative functional mapping of human brain.