Dear FSL-Team,
we are currently analyzing changes after a training in stroke patients. Therefore, we wanted to include the changes of lateralization for hand movement.
If I got it right, according to literature (Adcock et al., 2003; Jansen et al., 2006; Seghier, 2008) the LI can either be calculated using the voxel extend or the magnitude (e.g. percent signal change).
Both ways comprise pros and cons.
So I thought, probably the "best" way to calculate laterality could be to include information about voxel extend exceeding a specific threshold (voxel thresh_zstat) and the zstat_mean (both derived from my featquery analyses for e.g. M1) , as follows:
LI = (L_voxelextend * L_zstatmean - R_voxelextend * R_zstatmean) /
(L_voxelextend * L_zstatmean + R_voxelextend * R_zstatmean)
I compared the results of this "combined LI" with the LI I got just using the voxelextend or just using the %signalchange and got similar results.
So I wanted to ask whether this approach is permissible or not?
Or do I have to include other scores (e.g. cope_mean or the like)?
Thanks very much for your help, Daniela
|