help
86 Subscribers 
help > Too high test stat
Showing 1-3 of 3 posts
Jul 14, 2014 08:07 AM | Youngmin Huh - Seoul National University Hospital
Too high test stat
Hi, all
I tried NBS with DTI tractography data.
Every values in connectivity matrix (93*93) was normalized into 0 to 1.
I investigated group difference.
I found some edges from the results had extremely high test stat like 21518481347167528.00.
Is this fine?
YM Huh
I tried NBS with DTI tractography data.
Every values in connectivity matrix (93*93) was normalized into 0 to 1.
I investigated group difference.
I found some edges from the results had extremely high test stat like 21518481347167528.00.
Is this fine?
YM Huh
Jul 15, 2014 12:07 AM | Andrew Zalesky
RE: Too high test stat
No - this does not seem fine. I
suggest investigating this further. I would start by looking at the
histogram of the connectivity values for the edges with very high
test stats.
Andrew
Originally posted by Youngmin Huh:
Andrew
Originally posted by Youngmin Huh:
Hi, all
I tried NBS with DTI tractography data.
Every values in connectivity matrix (93*93) was normalized into 0 to 1.
I investigated group difference.
I found some edges from the results had extremely high test stat like 21518481347167528.00.
Is this fine?
YM Huh
I tried NBS with DTI tractography data.
Every values in connectivity matrix (93*93) was normalized into 0 to 1.
I investigated group difference.
I found some edges from the results had extremely high test stat like 21518481347167528.00.
Is this fine?
YM Huh
Jul 15, 2014 02:07 AM | Youngmin Huh - Seoul National University Hospital
RE: Too high test stat
Thanks for your reply!
I should look at connectivity matrix values at first.
Thanks.
YM Huh
I should look at connectivity matrix values at first.
Thanks.
YM Huh