Dear SPM users,
I am having some trouble trying to understand how the convolution
procedure works with respect to event related MRI.
In Friston et al. (1998) - "Event-Related fMRI: Characterising
Differential Responses", NeuroImage 7, 30-40 -, the observed haemodynamic
time series y(t) is modelled as
ita_0 + sum_i int_0^T ita_i(tau) . u_i(t - tau) dtau equation (1)
where ita_0 is a baseline level, ita_i(tau) is the haemodynamic response
kernel for condition i and u_i(t) is the stimulus presentation sequence
for condition i.
My question is, if u_i(t) is an event-related sequence then presumably it
is 0 everywhere but a finite number of points (i.e. a delta or "stick"
function), in which case why doesn't the integral
int_0^T ita_i(tau).u_i(t - tau) dtau
always equal zero?
Thanks
John
------------------------------------------------------------
John Kornak.
School of Mathematical Sciences and Institute of Hearing Research.
University of Nottingham.
email: [log in to unmask]
http://www.maths.nott.ac.uk/~jk/
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|