Dynamic Cerebral Contrast-enhanced Perfusion MRI - Biomedical Signal Analysis: Contemporary Methods and Applications - page 307

1 2 3 4

rCBF: 0.05
rCBV: 1.00
MTT : 21.30

rCBF: 0.04
rCBV: 0.81
MTT : 19.83

rCBF: 0.01
rCBV: 0.20
MTT : 21.44

rCBF: 0.01
rCBV: 0.14
MTT : 22.01

5 6 7 8 rCBF: 0.01

rCBV: 0.19
MTT : 19.59

rCBF: 0.01
rCBV: 0.11
MTT : 19.95

rCBF: 0.00
rCBV: 0.06
MTT : 21.87

rCBF: 0.00
rCBV: 0.06
MTT : 20.20

9 10 11 12

rCBF: 0.02
rCBV: 0.43
MTT : 23.15

rCBF: 0.02
rCBV: 0.35
MTT : 21.26

rCBF: 0.01
rCBV: 0.23
MTT : 20.14

rCBF: 0.03
rCBV: 0.51
MTT : 19.74

13 14 15 16 rCBF: 0.02

rCBV: 0.34
MTT : 19.69

rCBF: 0.03
rCBV: 0.64
MTT : 20.73

rCBF: 0.01
rCBV: 0.11
MTT : 20.43

rCBF: 0.04
rCBV: 0.82
MTT : 23.04

Figure 11.2
Cluster-specific concentration-time curves for the ”neural gas” network of a
dynamic perfusion MRI study in a subject with a stroke in the right basal ganglia.
Cluster numbers correspond to figure 11.1. MTT values are indicated as multiples
of the scan interval (1.5 sec), rCBV values are normalized with regard to the
maximal value (cluster #1). rCBF values are computed from MTT and rCBV by
equation (11.3). The X-axis represents the scan number, and the Y-axis is arbitrary.

was performed by a “neural gas” network employing 16 CVs (i.e. a
maximal number of 16 separate clusters at the end of the hierarchical
VQ procedure, as shown in figure 11.1).
Figure 11.2 shows the prototypical cluster-specific CTCs belonging
to the pixel clusters of figure 11.1. These can be computed from equation
(11.2), where the pixel-specific PTC x ( τ ) is replaced by the cluster-
specific CV.