I am developing a list of frequently asked questions (and answers)
regarding COBRA.
To add questions and/or answers or correct
existing ones, please contact me, Serge Dumoulin.
List of frequently asked questions:
- Is the retinotopic mapping stimulus program available?
- Does COBRA assume the functional data are acquired
on exactly the paradigm which is described in your paper? - Can COBRA handle different stimulus cycles for
polar-angle and eccentricity stimulation? - Is the vector derived from the anatomical MRI “truly” normal to the cortex?
- Does COBRA take into account the delay of the
hemodynamic response function? - Where can I download EMMA?
Is the retinotopic mapping stimulus program available?
The mapping stimuli we used originally was developed by Rick Hoge, but the software is now “retired”.
I am currently using a matlab program
using psychtoolbox. You can download my old (Mac OS 9.2.2, matlab 5.2.1) version here, or you can download a Mac OS X (matlab 7.0.4) version here.
Does COBRA assume that the functional data are acquired
on exactly the paradigm which is described in your paper?
No. COBRA only assumes a phase-encoded paradigm (see also: Engel et
al. 1994, 1997; Sereno et al. 1995; DeYoe et al. 1996). Even though
COBRA does not depend on the actual parameters of the stimulus design, it does
need to know one crucial parameter: number of stimulus cycles per scan
(rotating or expanding).
This parameter is the 5th required argument (v0.3), i.e. this is the number
of cycles of wedges and circles during the functional scan, independent
of the exact dimensions of the wedges and circles. If you have more than
one wedge or circle this parameter should be adjusted appropriately.
Basicly, this is the frequency (cycles/scan) that contains the
fundamental component of the fMRI response.
Note that this is the number of cycles after excluding the initial
time-frames (due to startup transients in the fMRI data). The number of
initial time-frames to be ignored can be defined by the 6th cobra argument.
Can COBRA handle different stimulus cycles for
polar-angle and eccentricity stimulation?
Yes, this feature is provided in COBRA version 0.2 and up.
Is the vector derived from the anatomical MRI “truly” normal to the cortex?
Yes, in T1-weighted images because the intensity of white>>gray>>CSF
(and in classified images where this can be controlled). Possible
exceptions within the brain would be ventricles, where the white>>CSF
boundary is also interpreted as a surface normal. This may cause a local
VFS reversal when the fundus of the Calcarine sulcus meets the
occipital horn of the lateral ventricles. Further exceptions outside the
brain near the occipital lobe would be skin-skull or skin-air boundaries.
Does Cobra take into account the delay of the
hemodynamic response function?
Yes, the standard way to cancel the phase lag is to average forward and
backward runs. This feature is provided in release 0.3 and up.
That said, however, the VFS computation does not depend on the
absolute phase, it only cares about the relative phases. Thus a
global phase-shift due to the hemodynamic delay does not matter
(see the simulations, in Dumoulin et al.,
2003). However, local differences in hemodynamic-delays around
the borders of the visual areas would shift the estimated border
location slightly.
Where can I download EMMA?
You can find EMMA on
bic-ftp-site, and on
Keith Worsley’s site.
You can download the Mac OS X version
here.