数学科学学院学术报告[2026]096号

（高水平大学建设系列报告1355号）

报告题目：Analysis of UK Biobank data through distributed regression with fused structure identifies the occipital pole and temporal lobe for cognition

报告人：李挺 副教授 （南方科技大学）

报告时间：2026年10月20日14:00 - 15:00

报告地点：粤海校区汇研楼633会议室

报告摘要：Assessing the relationship between whole brain imaging and neurological phenotypes is crucial to understanding human behaviors. However, the large size and complexity of whole brain data present unique challenges. In an effort to understand how brain regions react to cognitive demands through the analysis of large-scale UK Biobank whole brain data of over 50,000 individuals, we introduce distributed tensor regression via fused structure. This method effectively reduces the computational load through distributed computing, while preserving the intrinsic piece wise smoothness structure in the analysis of neuroimaging data. A novel objective function and communication strategy between distributed machines are constructed to ensure that local smoothness in voxel coefficients is maintained in distributed computing. Our data analysis consistently identifies the role of the areas in the occipital pole and temporal lobe in three touch screen cognitive tasks -- time to complete pair-matching round, duration to complete numeric path and numeric memory. While similar findings have been reported from fMRI studies, it is the first time such results have been identified using MRI with a large sample size.

报告人简介：李挺，南方科技大学统计与数据科学系研究员，副教授。主持国家自然科学基金面上项目，青年项目，获得国家级高层次人才项目（青年）。本科毕业于浙江大学，博士毕业于香港科技大学数学系，于耶鲁大学生物统计系从事博士后研究。曾在香港理工大学担任助理教授。长期从事复杂数据的统计研究以及大模型与生成模型，主要包括网络结构数据研究，复杂表型的基因相关性分析以及人脑图像数据分析。已在Annals of Statistics、JASA、JMLR、Annals of Applied Statistics、Genome Research、Human Brain Mapping、ICML 等统计、生物统计核心期刊和机器学习顶会上发表论文20篇。

邀请人：胡湘红

数学科学学院

2026年10月19日