Many will perform a map-based analysis on the cortical thickness or volume or area, then use a cluster-wise correction for multiple comparisons. If you want to go the ROI route, then you can test each ROI separately (ie, compute a p-value for the ROI0, then bonferoni correct the p-values for the number of tests (about 400). You could also FDR correct.

doug

Liukarl wrote:

Hello, Freesurfers

After autorecon3, there are cortical measures for the over 30 predefined regions. And for each regions, there are area, volume, and thickness 3 different measures, which means there are approximately 200 measure for both hemispheres.

As a preliminary analysis without specific priori hypothesis, what is the standard way to examine the massive data to find possible group difference for a small sample size about 40 subjects with 20 in each group.

thanks

Karl

搜索本应是彩色的,快来体验新一代搜索引擎-必应,精美图片每天换哦! 立即试 用！ http://cn.bing.com/?form=CRMADS%20

---

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer