Hopefully this is clarified here:
On 13/04/17 14:06, Ivy Tso wrote:
> Hi SPM experts,
> I ran a DCM analysis by first specifying a full bilinear model (with 4
> nodes, visual input into one node, psychological modulation on several
> intrinsic connections), then using post hoc Bayesian model selection to
> find the optimal (reduced) model. When I examined the parameter
> estimates, I found that for some (quite a few, actually) subjects, one
> or more self-connection parameters are positive. My understanding is
> that self-connections should always be negative, as they are a
> self-decaying process.
> I can't seem to find any option to constrain the values of
> self-connection parameters to < 0 in SPM either.
> Could someone advise if this result seems strange and speculate why this
> was the case and how to fix this?
> Ivy F. Tso, PhD, LP.
> Assistant Professor of Psychiatry
> Adjunct Assistant Professor of Psychology
> University of Michigan
> 2553 Rachel Upjohn Building,
> 4250 Plymouth Road, Ann Arbor, MI 48109
> Tel: (734)232-0373
> Fax: (734)936-7868
> Email: [log in to unmask] <mailto:[log in to unmask]>
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG