Dear John,
>
> I am trying to run FSL Eddy Correct on some DTI data and keep getting the error "Unable to find valid hyperparameters". I cannot seem to fix the issue or find any information related to this error. Any advice would be much appreciated!
I have not heard of this error occurring before. The error message is printed if the “initial guesstimate” (prior to optimisation) of the GP hyperparameters yields a non-positive definite covariance matrix. The only way that this can really happen is if one or more of the variances that eddy uses for the initial guesstimate is zero. This could happen if for example the data has been betted, yielding all values outside the object zero.
Jesper
>
> Best,
>
> John
|