Greetings PLS experts,

I recently used the PLSgui to perform a longitudinal analysis of structural MRI gray matter images for music training with the ‘Seed Pls’ method. I had 27 participants in my control group, and 29 participants in my training group. For these two groups, each participant was performed 3 times MRI scans (at the beginning of the training, the ends of the 6 months training, and 9 months after the end of training, respectively).The volumes of the left and right anterior(aHC) and posterior hippocampus(pHC), corresponding to the time points respectivly, were defined as the seed values.

The training group and control group  were performed seed PLS alone. According to suggestions from 'PLS Help' , there were two methods to array the datamats to perform this analysis: treating it as one group measured three times with entering these sessions into a single data matrix - essentially a three-condition(eg. gm1, gm2, gm3) study in method1, or treating it as three groups in method2. In general,The permutation rules were
the same between these two methods. However, When I tried both of the two methods, results were a liitle different, details were as follows:

Only the first 3 latent variables in the training group were shown to be significant in both methods. Here is the link of results:
http://s1043.photobucket.com/user/guodafei/media/PLS%20result/method1_LV1_bar_correlations_zpsq3e01ajb.png.html?sort=9&o=0

additional information are as follows,
method1:
LV1(p < 0.000; accounting for 60.67% of the cross-correlation variance)
LV2(p < 0.000; accounting for 27.74% of the cross-correlation variance)
LV3(p < 0.035; accounting for 3.63% of the cross-correlation variance)

method2:
LV1(p < 0.000; accounting for 60.67% of the cross-correlation variance)
LV2(p < 0.000; accounting for 27.74% of the cross-correlation variance)
LV3(p < 0.038; accounting for 3.63% of the cross-correlation variance)
The analysis was performed using 5000 permutations and 5000 bootstraps.

I am afraid that I have a wrong understanding of the permutation rules in 'Regular behav PLS', or should I increase the times for permutations and bootstraps?

Thanks in advance for any advice.

guo o