 |
 |
Hello Susie, Approach 1 indeed disregards the higher power of the parametric statistics, and correlation within the patient group between SPMs and performance variables seems more advantageous. In any case, it is advisable to check correlation plots, as also in SPM outliers can easily produce spurious effects. Ad 2) Comparing the monovariate effects of each cognitive variable visually, that is interpreting the patterns of significant clusters/voxels between two different cognitive measures, can give you a first hint, if different cortex areas are supporting these different domains (well rather the other way round in neurodegeneration: if decline of grey matter volumes go along with decline in performance). It should be noted that results are more locally specific if total volumes, that may also be affected by the process, are included as nuisance variable - some people however prefer total intracranial volume as covariate of no interest, as in diffuse atrophic processes entering total GM may indeed wipe out any regional effect. If both variables influence the performance, you could extract regional volumes you found by SPM, combine them with total GM and estimate the contributions in an 'offline' analysis. Ad 3) You can directly compare slopes between the two regressors in a t-test that opposes e. g. performance score 1 to performance score 2 (-1 1 or 1 -1), all adjusted for the other nuisance covariates. A problem arising here is may be collinearity, so ideally, the cognitive score regressors are orthogonal on each other or at least not strongly correlated. It may be worth starting with the least correlated scores, or orthogonalize one score on the other. If just entered as nuisance variable, "global" cognitive scores may be entered to make cognitive-brain-relationships more specific for the task of interest (I have seen this in papers on Alzheimer disease, but could not give you the precise source at the moment). Another method to analyse combined effects is conjunction analysis, however, I have scarce experience herein. Any other comments are appreciated, best wishes, Philipp Max Planck Institute of Psychiatry NMR Research Group Kraepelinsr. 2-10 80804 Munich Mail: [log in to unmask] Phone: 0049-89-30622-413