 |
 |
Dear SPM-Experts I’ve run a context conditioning experiment and I compared trials in which no electrical shock was associated with a context-stimulus (CS-) and trials in which a shock was associated with a different context in prior trials (CS+ unpaired) using a t-test (CS+ unpaired > CS-). The results didn’t reveal amygdala und hippocampus activations (even though these structures are known to be important for context conditioning) most probably because those structures are much more active in the initial stage of the experiment and significantly reduce their activity in later trials. There are studies which addressed this problem by incorporating additional regressors that represented time x condition interaction. These regressors were created by multiplying each condition (CS+ paired, CS+ unpaired, CS-) with a zero mean exponential function with a time constant of one fourth of the session length (Büchel et al., 1999) or by doing the same but using a linear decaying function (Marschner et al., 2008). Now my questions are: 1. How do I exactly build those regressors in SPM 8? 2. Should I use an exponential or linear decaying function? Is there a good (theoretical) reason to prefer one over the other? I would appreciate any kind of help on this! Best, Christian