Permutation correlation with one non-normal variable


I am trying to assess the relationship between a relatively normally distributed variable (brain volume) and a strongly right-skewed behavioral measures, which cannot be transformed to normality. I assume a linear relationship between the two variables.

My plan is to calculate the pearson correlation, and then permute one of the variables a number of times to get a significance value. Unfortunately I cannot find a lot of information on correlation permutation. Would this be a valid approach?


For anyone else with the same problem, the method I describe appears to be valid (Testing the Significance of a Correlation With Nonnormal Data: Comparison of Pearson, Spearman, Transformation, and Resampling Approaches; Bishara and Hittner, 2012)