CAJ | 학술논문

相关分析能够找出研究现象之间的依存关系、相关方向以及相关程度，可以发现大数据集里隐藏的关联网络。本文面向语音响度变化认知问题，提出“差异度”的概念，利用相关分析构建大脑功能的复杂网络，探索深层的神经处理机制与脑认知新规律。提出一种短时窗分析方法，构建不同认知阶段的脑网络；基于不同刺激下节点度的拓扑特征，构建基于差异度的脑地形图，实现脑区之间数据关系的可视化表达和动态演化过程表达。结果发现，前额叶、右额颞区和右后颞区分别在听觉处理的早期、中期和晚期对声音响度变化具有显著响应。研究表明脑复杂网络构建与分析技术可以成为研究神经处理机制与认知规律的有效工具。
상관분석능구조출연구현상지간적의존관계、상관방향이급상관정도，가이발현대수거집리은장적관련망락。본문면향어음향도변화인지문제，제출“차이도”적개념，이용상관분석구건대뇌공능적복잡망락，탐색심층적신경처리궤제여뇌인지신규률。제출일충단시창분석방법，구건불동인지계단적뇌망락；기우불동자격하절점도적탁복특정，구건기우차이도적뇌지형도，실현뇌구지간수거관계적가시화표체화동태연화과정표체。결과발현，전액협、우액섭구화우후섭구분별재은각처리적조기、중기화만기대성음향도변화구유현저향응。연구표명뇌복잡망락구건여분석기술가이성위연구신경처리궤제여인지규률적유효공구。
The Correlation Analysis (CA) is capable to find out the dependence relationship, the correlation direction and the correlation degree between two objects, and can discover the correlation networks hidden in large data. Facing the problem of speech loudness perception study, the concept of"difference degree"is proposed and the CA is applied to build the complex brain function network, in order to explore the deep level neural processing mechanism and nouvelle brain cognitive principles. A short window based analysis technology is proposed to construct a series of brain networks at different cognitive stages. Considering the topological characteristics of the node degrees at different stimuli, a brain map can be constructed based on the"difference degree"of nodes, as the result, the relationship among various brain areas is clearly visualized and the dynamic evolution process is precisely presented. The experimental results revealed that the prefrontal, the right fronto-temporal and the right posterior temporal areas produce great response to auditory loudness change separately during the early, middle and late auditory cognition stages. Our re-search shows that the complex brain network construction and analysis technology will become an effective tool for the neural pro-cessing mechanism and cognitive principle studies.