Data sources, parameters and characteristics of the 2 models for neuronal network analyses.
| Models | Data source13,24-26 | Parameters35,21,25 | Strenghts17-21 | Weaknesses35 | Utility36 |
|---|---|---|---|---|---|
| Graph Theory | EEG, fMRI, Histology | i. Path length ii. Clustering coefficient iii. Centrality matrix | i. Captures ictal and interictal events as intrinsic to the network. ii. Data sets could be generated and compared with each other for validation. | i. Requires separate models for effective versus functional connectivity. ii. Analysis of effective connectivity often leads to heavy computational load | Good for assessing extent of network changes relating to the interictal state |
| DCM | fMRI, EEG | i. Correlation ii. Covariance iii. Coherence | Incorporates directionality and therefore, effective connectivity as a basic aspect of analysis | i. Assumes triggers of ictal and interictal events to be extraneous to the network under investigation. ii. Heavily constrained by spatial and temporal resolution of data source. | Good for identifying seizure onset zone |
fMRI - Functional magnetic resonance imaging, DCM - dynamic causal modelling, EEG-Electroencephalography