How a particular region of the brain is declared significant?

Reference to [Voxel Based Assessment of gray and white matter] Page No.10 and similarly in bunch of other papers related to VBM a specific region of the brain is analyzed using a cluster of voxels between two groups for significant differences.

Take an example of two groups like 10 subjects and 20 control and for example we are analyzing a specific regions named “L middle temporal gyrus” with 15000 cluster size of voxels. then we would have obviously 15000 t-test and 15000 p-values corresponding to those voxels when comparing this region between two groups.

Now since some P-values would be significant and some not then how would we decide as to when to declare this region significant or will we average all t-values to a single T value and then a single P value is calculated?

Is my above understanding of comparing regions between two groups correct if NOT then please guide me how a particular brain region is analyzed on the basis of P-values and a cluster size?

(https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2844895/pdf/nihms-182453.pdf)

As you can imaging it’s a complex topic and there are different approaches to assessing statistical significance of changes across groups in brain imaging data. This video course gives a nice overview of the topic: https://www.youtube.com/playlist?list=PLB2iAtgpI4YGqeN4wJpwuo8yIWMfKaBD2

1 Like