 |
 |
Dear SPMers, a short question regarding the validity of my PPI-analysis in a single-case study. The patient was measured twice (t1 = during symptoms; t2 = during recovery) using a blocked-design fMRI paradigm with two conditions (facial & non-facial stimuli). For the analyses both sessions were entered in a single design matrix (conditions: facial stimuli t1 / non-facial stimuli t1 / facial stimuli t2 / non-facial stimuli t2). The aim of the study is to analyze: (1) Unspecific differences between t1 & t2 These were assessed with the contrasts 0.5 0.5 -0.5 -0.5 and 0.5 0.5 -0.5 -0.5 (2) Relative differences for facial as compared to non-facial stimuli between t1 & t2 These were assessed with the contrast 0.5 -0.5 -0.5 0.5 and revealed higher activity at t1 in regions A, B, C (3) Differences in dynamic interactions for facial relative to non-facial stimuli between t1 & t2 These were assessed using PPI. For t1 & t2 separate Eigenvariates were extracted from spheres centered at significant regions from (2). Then PPI variable were created for t1 and t2 with the contrasts weights ‘facial = 1’ & ‘non-facial = 2’. Finally PPI GLM analysis was performed with both sessions entered in a single design matrix using the session-specific regressors PPI.ppi / PPI.Y / PPI.P Comparing changes in functional coupling between t1 & t2 reveals large regions with higher connectivity for all three regions (identified in (2)). Can these results be interpreted as between-session changes in functional connectivity? Or do they simply display an artifact due to an error in the PPI analysis (I wonder because the analysis revealed such large & overlapping regions for the three seed regions) Thanks in advance & best regards Ben