Dear SPM list The generic advice in designing a DCM study seems to be to ensure that it is a 2x2 factorial design with a driving sensory factor and a contextual factor (e.g. attention). Is there a statistical/mechanistic basis for this advice in terms of the DCM analysis or is this advice just good practice for design efficiency and ensuring there is an interesting effect to explore. I'd be interested to understand more. Thanks for any help you might provide. Steven