Classical and Bayesian Inference in fMRI - Advance Image Processing in Magnetic Resonance Imaging - page 548

FIGURE 17.8 (continued) significant bilinear effects in the absence of a significant
intrinsic coupling. Number in brackets represent the posterior probability, expressed as a
percentage, that the effect size is larger than 0.17. This cutoff corresponds to a time constant
of 4 sec or less — in DCM stronger effects have faster time constants. (b) Fitted responses
based upon the conditional estimates and the adjusted data. The insert shows the location
of the regions centered on the primary visual cortex V1: 6, −84, −6 mm; motion-sensitive
area V5: 45, −81, 5 mm. SPC: 18, −57, 66 mm; IFG: 54, 18, 30 mm. The volumes from
which the first eigenvariates were calculated corresponded to 8 mm radius spheres centered
on these locations. Subjects were studied with fMRI under identical stimulus conditions
(visual motion subtended by radially moving dots) while manipulating the attentional
component of the task (detection of velocity changes). The data were acquired from normal
subjects at 2-T using a Magnetom VISION (Siemens, Erlangen) whole body MRI system,
equipped with a head volume coil. Here, we analyze data from the first subject. Contiguous
multislice T2*-weighted fMRI images were obtained with a gradient echo-planar sequence
(TE = = = 64 × 64 × 32, voxel size 3 × 3 × 3 mm).

Each subject had 4 consecutive 100-scan sessions comprising a series of 10-scan blocks
(D F A F N F A F N S) under 5 different conditions. The first condition (D) was a dummy
condition to allow for magnetic saturation effects. F (fixation) corresponds to a low-level
baseline where the subjects viewed a fixation point at the center of a screen. In condition
A (attention), subjects viewed 250 dots moving radially from the center at 4.7° per sec
and were asked to detect changes in radial velocity. In condition N (no attention), the
subjects were asked simply to view the moving dots. In condition S (stationary), subjects
viewed stationary dots. The order of A and N was swapped for the last two sessions. In
all conditions, subjects fixated the center of the screen. In a prescanning session, the
subjects were given five trials with five speed changes (reducing to 1%). During scanning
there were no speed changes. No overt response was required in any condition.

40 msec, TR 3.22 sec, matrix size

connectivity and attention was allowed to modulate the backward connections
from IFG and SPC .

The results of the DCM are shown in Figure 17.8a . Of primary interest here is

the modulatory effect of attention that is expressed in terms of the bilinear coupling
parameters for this third input. As hoped, we can be highly confident that attention
modulates the backward connections from IFG to SPC and from SPC to V5 .
Indeed, the influences of IFG on SPC are negligible in the absence of attention
(dotted connection in Figure 17.8a). It is important to note that the only way that
attentional manipulation could affect brain responses was through this bilinear
effect. Attention-related responses are seen throughout the system (attention epochs
are marked with arrows in the plot of IFG responses in Figure 17.8b). This atten-
tional modulation is accounted for, sufficiently, by changing just two connections.
This change is, presumably, instantiated by the instructional set at the beginning of
each epoch. The second point this analysis illustrates is how the functional segre-
gation is modeled in DCM. Here one can regard V1 as a “segregating” motion from
other visual information and distributing it to the motion-sensitive area V5 . This
segregation is modeled as a bilinear “enabling” of V1 to V5 connections when, and
only when, motion is present. Note that in the absence of motion the intrinsic V1
to V5 connection was trivially small (in fact, the MAP estimate was 0.04). The

key advantage of entering motion through a bilinear effect, as opposed to a direct

−