 |
 |
Dear Dr Weis, > >Is it true that when using short SOAs I have to choose 'Interaction >among trials (Volterra )' - YES, because otherwise SPM does not acount >for short presentation rates at all ? No this is not true. As I understand it the argument about using short SOAs comes from modelling which was done assuming linearity (with no interactions between trials), and assuming that trials are randomly ordered. SPM doesn't need to 'account for' short presentation rates, it simply needs there to be a significant difference in the ability of different columns in your design matrix to model BOLD responses to different types of trial, and this theoretical work seems to indicate that packing in a large number of trials improves that statistical power of these comparisons. >How short can the SOA reasonably be ? When the chief non-linearity, ie. saturation at extremely short SOAs is (approximately) accommodated, the models suggest that the SOA should be no shorter than a second. In practice, SOAs of about 2 seconds seem reasonable. However, it should be emphasized that this theoretical thinking is only just beginning to be tested empirically. The proof of the pudding is in the eating, regardless of how attractive the recipe book seems to be. Best wishes, Richard Perry. from: Dr Richard Perry, Clinical Research Fellow, Wellcome Department of Cognitive Neurology, Darwin Building, University College London, Gower Street, London WC1E 6BT. Tel: 0171 504 2187; e mail: [log in to unmask] Pager: 04325 253 566. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%