Loading...

Recognizing Center of Siezur with Clustering Algorithm

Akhshi, Amin | 2016

1308 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 49318 (04)
  4. University: Sharif University of Technology
  5. Department: Physics
  6. Advisor(s): Rahimitabar, Mohammad Reza
  7. Abstract:
  8. Complex systems are composed of a large number of subsystems behaving in a collective manner. In such systems, which are usually far from equilibrium, collective behavior arises due to self-organization and results in the formation of temporal, spatial, spatio-temporall structures. Examples of complex systems are turbulent flow, stock markets, dynamics of a brain, etc. In study of the complex systems, we always encounter with handling and analysing of a Big-Data set. There are several approaches to overcome this problem, among which the most powerful method is the clustering analysis. Clustering algorithm is based on the classifying of dynamics of complex system using some similarity measure. In this thesis, we analyze extended EEG data sets of 25 patients with epileptic brain. At first, time series of EEG is divided to equal time windows, then using the corresponding mutual information matrix in each window, clustering analysis have been carry out for estimated matrices. In similar way, we define the "mean phase coherence matrices" in each window and apply clustering analysis again. The mean phase coherence is able to specify the degree of synchronization of different parts of brain. We find qualitatively that the clustering analysis of mutual information and mean phase coherence matrices provide same results for the states of epileptic brain. We are able to classify several brain states, which dynamics of brain jumps from one state to another in each window, so that in some states brain stays for a short period. We determine the state dynamics of epileptic brain and also show that the method enables us to detect the structure responsible for seizure generation (epileptic focus)
  9. Keywords:
  10. Epilepsy ; Mutual Information ; Clustering ; Mean Phase Coherency ; Electroencphalogram ; Synchronisation

 Digital Object List

 Bookmark

...see more