Print

Print
Dear Karny,

Let me explain with some arbitrary values. Depending on your threshold, let's say a voxel is significant if t value > 4. For your schizophrenia patients, maybe all voxels got t value < 4 and thus no 'activation'. Among the controls, maybe some voxels got t value > 4 and thus 'activated'.

In one tailed two sample t test [schizophrenia patients - controls], some voxels can be significant if the difference in t value is > 4, e.g. [3 - (-2)] = 5.
In this case, this voxel i has t value of 3 in the former group and t value of -2 in the latter group. The resultant becomes > 4.

Best,
Andy

On Mon, Jul 27, 2015 at 4:32 AM, Karny <[log in to unmask]> wrote:
Hi
In one sample t test of schizophrenia patients for a specific contrast we see activations in a wide range of areas but there's no activation in the DLPFC. However the controls activated the DLPFC as revealed by a one sample t test for the same contrast.
But in a two sample t test for the same contrast the analysis revealed a higher activation of the DLPFC of schizophrenia patients compared to controls.
Is there a rational explanation for these results?
Thanks...
Karny