There is only ~ 0.66 correlation between tract length matrix outputs (for the same input data) of the pipeline's various probtrackx options: GPU, GPU batch, CPU, CPU batch. This suggests that the various options produce notably different tract length matrices. A higher correlation of > 0.8 would be preferable.
There is only ~ 0.66 correlation between tract length matrix outputs (for the same input data) of the pipeline's various probtrackx options: GPU, GPU batch, CPU, CPU batch. This suggests that the various options produce notably different tract length matrices. A higher correlation of > 0.8 would be preferable.