Hi
I faced a similar issue with my code after migrating to MATLAB R2023b. A workaround to circumvent this problem would be to utilize a "cvpartition" object with 5 folds. Below is an example of how you can implement this:
x = randn(250, 50);
y = randn(250, 1);
cv = cvpartition(250, "KFold", 5);
plsregress(x, y, 32, 'CV', cv, 'MCReps', 10);
Let me know if this helps
