Biology Reference
In-Depth Information
14.5. Results
In this section, we present the computational results from a patient with epilepsy.
Figures 14.7, 14.8 and 14.9 show the connectivity among brain regions clustered
by the proposed graph-theoretic approach. Figures 14.7 and 14.8 manifest the
same pattern that the
L
(
T
)
D
regions consistently exhibit high connectivity prior
to the seizure onset over a 2-hour time horizon. With a proper threshold for con-
structing the connectivity graph, the uniqueness maximum clique is generated dur-
ing interictal state of the subject. Moreover, the maximum clique size increased
and covered almost all the brain regions during the epileptic seizure. It is known
during the epileptic seizure, the abnormal discharge from epileptic brain regions
may spread to the other brain regions and thus results the increase of the brain
connectivity. Thus the maximum clique size reaching the maximum possible size
(i.e., equal to the network size) was observed during an epileptic seizure. Through
visual inspection on raw EEG recordings, the
L
(
T
)
D
region is the area where
the seizure was initiated. The EEG recordings were reviewed by two separated
board certificated physicians to identify the location. In Fig. 14.9, the brain re-
gions clustered by our approach were both from left and right orbitofrontal (i.e.,
R
(
O
)
F
and
L
(
O
)
F
) regions whereas the right hippocampus appeared to have
Fig. 14.7. A plot shows the brain regions clustered by graph-theoretic approach in the constructed
brain connectivity graph. The highlighted areas represent the selected brain regions in the maximum
clique. The
L
(
T
)
D
areas tend to have strong connectivity prior to the actual seizure onset.
Search WWH ::
Custom Search