 |
 |
Dear SPMlers, I am currently testing several models with DCM, with parametric modulations (coming from an RL model) in my b and c matrix. As it turns out, the winning model goes well with my hypothesis, however I have some statistical problems. That is, when I am doing a t-test (against 0) on the different factors of my b matrix, nothing turns out significant, although BMS tells me that the model works better than models without these modulators. Additionally my MAP estimates are very small. If I understand this right, I am not "allowed" to say that in my winning model, the connections are modulated by any inputs? or where is the error in my reasoning/in my statistical approach? thanks,