| 嘉宾 | 胡正珲（浙江工业大学） |
| 时间 | 2019年10月22日（星期二）晚上 20:00（北京时间） |
| 题目 | Unified Model Selection Approach Based on Minimum Description Length Principle in Granger Causality Analysis |
| 主持 | 谢立（浙江大学） |
报告摘要
Granger causality analysis (GCA) provides a powerful tool for uncovering the patterns of brain connectivity mechanism using neuroimaing techniques. Conventional GCA applies two different mathematical theories in a two-stage scheme: (1) the Bayesian or Akaike information criterion for the predictors associated with history information; (2) F-statistics for evaluating the relative effects of exogenous variables. While specifying endogenous and exogenous effects are essentially the same model selection problem, this could produce different benchmark in the two stages and therefore degrade the performance of GCA. In this course, we present a unified model selection approach based on the minimum description length (MDL) principle for GCA in the context of the general regression model paradigm. Compared with conventional methods, our approach emphasize that a single mathematical theory should be hold throughout the GCA process. First, the MDL principle help to select an appropriate model being suitable to the specific noise level that is present. The MDL principle then serves as a single mathematical framework for specifying endogenous and exogenous effects. Under this framework, all candidate models within the model space might be compared freely in the context of the description length, without the need for an intermediate model. We illustrate its advantages over conventional two-stage GCA approach in a 3-node network and a 5-node network synthetic experiments. The unified model selection approach is capable of identifying the actual connectivity while avoiding the false influences of noise. More importantly, the proposed approach produced more consistent results in a challenge fMRI dataset for causality investigation, mental calculation network under visual and auditory stimulus, respectively. The proposed approach has potential to accommodate other Granger causality representations in other function space. The comparison between different GC representations in different function spaces can also be naturally deal with in the framework.
嘉宾简介
胡正珲，2005年浙江大学物理系凝聚态物理博士毕业，香港科技大学电子与计算机系，美国Rochester Institute of Technology 计算与信息科学系博士后，现为浙江工业大学理学院教授，主要从事数字信号处理与图像分析工作。在相关领域发表SCI论文三十余篇，主持或参与各类省部级科研项目十余项。申请国家专利5 项。曾获香港科技大学“Productivity Reward”、“浙江省钱江人才计划”、“浙江省科技进步二等奖”。
特别感谢本次Webinar主要组织者：
谢立（浙江大学）
活动须知
1. MICS在线学术讲座依托在线直播平台进行，听众请点击直播链接https://live.polyv.cn/watch/400484（注：该链接为10月22日报告链接，该链接每期会和讲者信息一起更新）参加活动，支持安装Windows系统的电脑、MAC电脑、手机等设备；手机客户端也可直接扫描二维码进入直播；
可直接扫描二维码进入10月22日直播
2. 活动开始前5分钟左右，讲者会开启直播，活动时讲者会上传PPT或共享屏幕，听众可以看到Slides，听到讲者的语音，并通过聊天功能与讲者交互；
3. 活动过程中，请不要说无关话语，以免影响活动正常进行；活动过程中，如出现听不到或看不到视频等问题，建议退出再重新进入，一般都能解决问题；建议通过速度较快的网络参加活动，优先采用有线网络连接；
4. 后续每期讲座信息，会通过MICS微信公众号“医学图像计算青年研讨会”（扫描下方二维码关注MICS微信公众号）或MICS QQ群（群号：641894878）进行通知（注：申请加入MICS QQ群时需验证姓名、单位和身份，缺一不可；入群后，请实名，姓名身份单位；身份：学校及科研单位人员T，企业研发I，博士D，硕士M）
“医学图像计算青年研讨会”微信公众号
MICS在线学术讲座的模式和组织方式借鉴了很多VALSE的经验，从VALSE得到了很多的启发，在此对VALSE组委会表示衷心的感谢，也祝愿MICS和VALSE越办越好！
