Ph.D. thesis (2021)
Abstract
Studying the dynamic of ictal activities as well as the underlying mechanisms generating them are of crucial importance for enhancing our understanding of ictogenesis. Seizures have previously been investigated to understand ictal state transitions; however, these attempts have generated conflicting results. For instance, some findings point towards highly varying connectivity during seizures and relatively stable connectivity in pre-ictal and post-ictal periods [1] while others have found rapid changes in network geometry in pre-seizure epochs and slower geometric changes during ictal activity or no significant differences among the ictal, pre-ictal and post-ictal state transition rates [2,3]. These discrepant observations led us to reexamine the dynamic of state transitions during seizures. The transition from interictal into ictal activities has also been widely investigated to enhance our understanding of the underlying mechanisms of seizure generation. However, it remains largely unclear whether seizures with distinct onset patterns are preceded by varying network interactions and whether different neurophysiological mechanisms underly different seizure onset patterns. To better understand how the epileptic brain as a dynamical system modulates its states, using different functional connectivity measures, we identified patterns of recurrent states and investigated the dynamic of ictal state transitions on different scales: locally and globally. Our findings point towards distinct dynamics for state transitions at these different spatial scales. While the pattern of global state transitions led to the conclusion that the brain changes state less frequently during ictal activity, locally, it underwent a higher rate of state transitions. Further, our results show different patterns of state transitions in the seizure-onset-zone between seizure-free (SF) and non-seizure-free (NSF) patients which could have a practical application in predicting surgical outcome (Chapter 4, Article 1). We further expanded the study of ictal state transitions and investigated the transition from interictal to ictal patterns. More precisely, we explored if the network interactions before the generation of seizures with different onset patterns vary. To this end, we classified seizure onset patterns of 103 intracranial EEG seizures from 20 patients with drug-resistant focal epilepsy. We then surveyed network interactions prior to the generation of distinct seizure onset patterns by calculating directed functional connectivity matrices. Our observations suggest that, at the network scale, different interactions precede the generation of different seizure onset patterns (Chapter 5, Article 2). This result raised an important question of whether different physiological mechanisms underly different seizure onset patterns? By employing a computational framework that takes into account some of the most relevant physiological events (e.g., depolarization block, collapse, and recovery of inhibitory activity) and different scenarios of imbalanced excitatory-inhibitory activity, we addressed this question and explored whether seizures with different onset patterns stem from different underlying mechanisms. Our modeling results revealed that, depending on the excitation level, seizures could potentially be generated due to both enhancement and collapse of inhibition for specific ranges of parameters. While the collapse of inhibition has been widely reported as a main mechanism underlying seizure generation [4–7], some experimental studies have addressed the enhanced inhibitory activity before and at the very onset of ictal activities [8–14]. Our modeling approach could unify these seemingly contradictory mechanisms suggested by the previous experimental studies for the involvement of inhibitory activity in seizure generation [4–11,13,14] and successfully reproduced some of the commonly observed seizure onset patterns. Our findings indicated that different onset patterns can, in fact, arise from different underlying mechanisms (Chapter 6, Article 3). Taken together as a whole, in this thesis we investigated the dynamic and generation of ictal events at different scales and from different perspectives. We explored different functional network connectivity measures to better understand the ictal state transitions. Then, seizure onset patterns were classified and investigated to see if different network interactions precede seizures with different onset patterns. Finally, by constructing a computational model, we investigated the mechanisms that could generate seizures with different onset patterns and suggested pathways for seizure generation that could explain some contradicting seizure generation mechanisms observed in experiments.
Statistics
Total downloads
Downloads per month in the last year
Origin of downloads
