Clustering Electroencephalogram Recordings to Study Mesial Temporal Lobe Epilepsy - Clustering Challenges in Biological Network - page 276

stronger connectivity than left hippocampus. The raw EEG recordings revealed
that the epileptic seizure originated from R ( O ) F regions. The above results sug-
gest that the brain regions selected in a maximum clique prior to seizure onset, are
indeed the regions where the seizure is initiated.

14.6. Conclusion and Discussion

In conclusion, we have presented a new approach for modeling and analyzing the
brain connectivity. This approach successfully localized the epileptic focus (foci),
the strongest connectivity were found in the regions where epileptic seizures are
initiated. Previous studies also reported that the EEG recording acquired from
epileptic regions contained higher nonlinear information than other brain regions
[17]. The clustering result can be very useful for the current clinical environment,
since about 20% of epilepsy patients will undergo the epilepsy surgery seeking
better seizure control or the possibility of becoming seizure free. The outcome
of the epileptic surgery depends largely on the accuracy of the focus localization.
From the results, the proposed approach can serve as a tool for focus localization
in pre-surgical evaluation stage. To further confirm our findings, a larger number
of patients with different type of epileptic seizures is required. Moreover, our
result also agrees with the fact that the connectivity pattern plays an important
role for the brain function. Although these observations could lead us toward the
understanding of the information process in temporal lobe epilepsy, however, the
questions such as how the alteration of connectivity occurs and how it develops
prior to a seizure still remain open. The understanding of the mechanisms of
seizure development is the key to develop a reliable epilepsy seizure detector and
to control or prevent an impending seizure well before its actual onset.

Acknowledgment

This material is based upon work supported by the National Science Foundation
under Grant No. 0546574. The authors thank Dr. Chris Sackellares and Dr. Deng-
Shan Shiau for their fruitful knowledge about epilepsy.

References

[1] J. Arnhold, P. Grassberger, K. Lehnertz, and C. E. Elger. A robust method for de-

tecting interdependences: Application to intracranially recorded EEG. Physica D ,
134(4):419-430, 1999.