Dear SPMers,
I have got a similar problem to that mentioned in the attached email
(also refer to http://www.mailbase.ac.uk/lists/spm/2000-05/0108.html).
We are trying to specify a block fMRI model in the batch mode and would
like to use 'fixed response (Half-sine)' as basis function. According to
the documentation, we should set bf_ep(1).ep_type=3. Thus, we specify
the condition as follows.
cond_names = {'active', 'rest'};
conditions(1) = struct( ...
'volterra', 'no', ...
'names', {cond_names}, ...
'types', {{'epochs', 'epochs'}}, ...
'onsets', {{ons1,ons2}}, ...
'durations', {{dur1,dur2}}, ...
'bf_ep', [1 2], ...
'variable_dur', 'no' ...
);
bf_ep(1) = struct( ...
'ep_type', 3, ...
'fct_nb', 1, ...
'conv', 'yes', ...
'deriv', 'yes', ...
'length', 8 ...
);
However the 'basis functions' on the display window shows 'mixed'
instead of 'fixed response (Half-sine)', which we can get by using GUI.
Another difference is that 'conditions per session' on the window
becomes 5 instead of 2 in manual mode.
I am wondering if it is possible to remove these two differences.
Any help is greatly appreciated.
Sincerely,
Li
============================================
model specification in batch mode
Subject: model specification in batch mode
From: Barry Giesbrecht ([log in to unmask])
Date: 15 May 2000 - 14:48 BST
sorted by: [ date ] [ thread ] [ subject ] [ author ]
Next message: Karl Friston: "Re: Conjunctions"
Previous message: andreas joachim bartsch: "design setup"
Dear SPMers,
I have been trying to specify an event-related fMRI model in batch mode
and I am
having difficulty with two aspects. Any help on these issues is greatly
appreciated.
First and foremost, when I specify a basis set, I want to use "hrf (with
time
derivative)", which according to the batch mode documentation should be:
bf_ev(i).ev_type = 2
in my code, i specify it as follows (code includes condition
definition):
%-------------------------------------------
conditions(1) = struct( ...
'names', {{'qll', 'qbl', 'tll', 'tbl', 'other'}}, ...
'onsets', {{qllr2c, qblr2c,tllr2c, tblr2c, other_r2c}}, ...
'types', {{'events','events','events','events','events'}}, ...
'bf_ev', [1 1 1 1 1], ... %basis functions for event design
'volterra', 0, ...
'variable_dur', 0);
%-------------------------------------------
% Hemodynamic Response Function
bf_ev(1) = struct( ...
'ev_type', 2);
%-----------------------------------
When the Design graphics window is displayed, the descriptor for basis
functions
is 'mixed', not 'hrf (with time derivative)' as it is when i specify the
model
through the gui. Similarly, when I look in the entry for the basis
function
descriptor in Sess{1}.BFstr it too says 'mixed' and not 'hrf (with time
derivative)'. Am I specifying the basis function incorrectly to get 'hrf
(with
time derivative)', if so, could someone please show me the error in my
ways. If
not, is the 'mixed' label specific to batch mode and am I really getting
a 'hrf
(with time derivative)' if I specify in the manner indicated above?
The second issue regards warnings that come up when I go to specify &
estimate
the model in batch mode...the warnings are as follows:
*****************************
SPM99: spm_fmri_spm_ui (v2.31) 09:42:43 - 15/05/2000
========================================================================
Warning: couldn't access the sub ref ...
> In p:\matlab\spm99hi\spm_input.m at line 191
In p:\matlab\spm99hi\spm_fmri_spm_ui.m (sf_abort) at line 664
In p:\matlab\spm99hi\spm_fmri_spm_ui.m at line 266
In p:\matlab\spm99hi\spm_bch.m at line 156
lastvar =
types: 4
RT: 2
nsess: 1
nscans: 199
replicated: 0
conditions: 1
conditions_nb: 5
stochastics_flag: 0
parametrics_type: {'none'}
regressors_nb: 0
files: {[199x63 char]}
global_effects: 'scale'
HF_fil: 'specify'
LF_fil: 'hrf'
int_corr: 'none'
trial_fcon: 1
now_later: 0
time_sampl: {[]}
interp: ''
remain: {[1x199 double]}
s_typ =
.
subs =
stop_writing
SPM99: spm_fmri_design (v2.27) 09:42:46 - 15/05/2000
----------------------------------------------------------------
Saving fMRI design : ...SPM_fMRIDesMtx.mat saved
Warning: couldn't access the sub ref ...
********
What do the warnings 'couldn't access the sub ref ...' mean? And do they
impact
the validity of the model estimation?
Again, any help is greatly appreciated.
Sincerely,
Barry
**************************
Barry Giesbrecht, PhD
Center for Cognitive Neuroscience
Levine Science Research Center Rm B203
Box 90999
Duke University
Durham, NC, 27708, USA
E-mail: [log in to unmask]
Tel: +1 919 681-1029
Fax: +1 919 681-0815
Center's Web Page: http://www.mind.duke.edu
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|