Hi all,
I've been struggling (honestly!) to understand how the high pass filter cut
off period is derived.
I have a block design with 4 trial types in each of 4 sessions, with trial
orders for 1 particular subject as shown below (different Session orders
accross subjects),
Ses1: 1 2 3 4
Ses2: 2 3 4 1
Ses3: 3 4 1 2
Ses4: 4 1 2 3
TR = 2.1, block (epoch) length = 27secs.
I have investigated two different designs. As trial type 2 is essentially
a baseline (fixation) condition, the first design specified 3 trial types
(1, 3, 4). In this case the default cut off period(s) suggested by spm
were [67 67 118 67].
Does spm apply different cut offs to each session or take one of the above
values (the min?) as the cutoff?
I assume the value of 118 reflects the fact that for Session 3 the baseline
condition is the last block of the session. However, I'm afraid i do not
understand how the actual figures were arrived at.
I set up another design explicitly specifying all 4 trial types to see if
that clarified matters. In this case the default cut off period(s)
suggested by spm were [32 32 32 32].
After consulting spm_fmri_ui.m I find that, "The default cutoff period is
twice the maximum time interval between the most frequently occurring event
or epoch (i.e the minium of all maximum intervals over event or epochs)",
which no doubt leaves me better informed although unfortunately none the
wiser.
Another semi-related question, does anyone know of any references that
discuss the optimal block length and soa for block designs? There are a
few papers around that focus on event related designs and also mention
block designs but i haven't found any that specifically examine the
optimization of block designs.
Thanks in advance,
Graham
Graham Byatt
Psychology Dept
University of Western Australia
WA 6907
Australia
ph. 9380 3573 (lab)
or 9386 4776 (home)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|