Hi All,
I'm trying to study the effect of a treatment (cognitive rehab) on FA. Hypothesis was that FA will increase as a result of treatment. I used paired design with -e option and design.grp in randomise, as recommended. To my surprise, i got completely opposite results, with FA greater in many areas (1-p =0.95) before treatment. I've checked for outlier subjects, registration errors but i dont see anything obvious and i'm assuming the TBSS code is correct :) There was no significant change in MD and L1 but there was greater radial diffusivity ( mean of L2,L3) post-treatment in many areas after thresholding. This result is surprising because patients improved on the Cognitive tests post-treatment. Ive come up empty thus far searching through literature. Studies have shown increased FA consistently after similar cognitive treatments. Even if treatment time (2weeks) wasnt long enough , it still wouldnt explain decrease in FA, would it? Any ideas explaining this finding? Thanks
-Vish
|