Fundamental Research on Cortical Activity Visualization

皮层活动可视化的基础研究

• 批准号: 09044148
• 负责人: KOSUGI Yukio
• 金额: $ 2.62万
• 依托单位: Tokyo Institute of Technology
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B).
• 财政年份: 1997
• 资助国家: 日本
• 起止时间: 1997 至 1999
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-09044148/
• 关键词: cortical activity; neural networks; inverse problems; dipole localization; PET; metabolic imaging; regularization; 脳電位; ネットワークインバージョン; 協調項; SPECT画像; パーキンソン病; 動的正則化

Artificial neural networks can be exploited to solve inverse problems arising from the visualization of neural activities in the brain. In this research, we made use of the network inversion techniques for solving inverse problems with special attention directed towards electroencephalographic dipole localization and the improvement of positron emission tomography. In **r regluarized network inversion technique, for stabilizing the solution, we explicitly include the a priori knowledge by adding penalty terms to the energy function and/or build this knowledge into the architecture of the multi-layered neural networks that are to be used as an inverse problem solver.In the electroencephalogram analysis, the consensus term added to the energy function facilitated S-dipole localization for visually evoked potentials. Effectiveness of our regularization technique is shown in improving the positron emission tomogrgphic images and for generating metabolic images of the brain, under the constraints given by the a priori knowledge inherent to the measurement systems and physiological rules.

人工神经网络可以被用来解决由大脑中神经活动的可视化引起的逆问题。在这项研究中，我们利用网络逆技术来求解逆问题，特别是针对脑电偶极子定位和正电子发射断层成像的改进。在**r正则化网络逆技术中，为了稳定解，我们通过在能量函数中加入惩罚项来显式地包含先验知识和/或将这些先验知识嵌入到多层神经网络的结构中作为逆问题的求解。在脑电分析中，能量函数中加入的共识项促进了视觉诱发电位的S偶极子定位。在测量系统和生理规则所固有的先验知识的约束下，我们的正则化技术在改善正电子发射断层图像和生成大脑的代谢图像方面表现出了有效性。

项目成果

小杉,植本,小川: "ネットワークインバージョンにおける動的正則化" 電子情報通信学会論文誌. J-81-DII no 7. 1639-1646 (1998)

Kosugi、Uemoto、Okawa：“网络反演中的动态正则化”电子、信息和通信工程师学会汇刊 J-81-DII 第 7 期。1639-1646 (1998)。

Y.Kosugi et al.: "Newral Network-Based PET Image Reconstruction" Methods of Information in Medicine. 36. 329-331 (1997)

Y.Kosugi 等人：“基于 Newral 网络的 PET 图像重建”医学信息方法。

Y. Hayashi, Y. Kosugi, B. He: "A Network Inversion Technique for Estimating Equivalent Dipole Description of Visual Evoked Potential"Method of Information in Medicine. (印刷中). (2000)

Y. Hayashi，Y. Kosugi，B. He：“估计视觉诱发电位的等效偶极子描述的网络反演技术”医学信息方法（2000 年出版）。

Kazuhiro Matsui,Yukio Kosugi: "Image Segmentation Using Genetic Method to Select Feature Indices" Proc.ICONIP'98. 352-355 (1998)

Kazuhiro Matsui，Yukio Kosugi：“使用遗传方法选择特征索引的图像分割”Proc.ICONIP98。

Y. Hayashi, Y. Kosugi, B. He: "A Network Inversion Technique for Estimating Equivalent Dipole Description of Visual Evoked Potential"Proc. 3^<rd> Int. Workshop on Biosignal Interpretation. 111. 79-82 (1999)

Y. Hayashi，Y. Kosugi，B. He：“一种用于估计视觉诱发电位等效偶极子描述的网络反演技术”Proc。