טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
Ph.D Thesis
Ph.D StudentHanna Keren
SubjectNeural Network Synchrony: Dynamics and Control
DepartmentDepartment of Medicine
Supervisor Full Professor Marom Shimon
Full Thesis textFull thesis text - English Version


Abstract

Synchronization of neural networks is one of the hallmarks of brain activity and is considered significant in health and disease. As such, in recent years tremendous effort is invested in means to characterize, control and stabilize neural synchronized activity at the whole brain level. The conditions for emergence of different synchronization modes as well as the tradeoffs and constraints involved in its stabilization, however, remain elusive due to obvious complications entailed by studying whole brain dynamics. This thesis begins by presenting an effective experimental framework to control neural synchronized response features (probability and latency) of cortical networks in-vitro, over many hours. Such approach is offered as a model for studying controllability of neural networks in the wider context and identifying possible functional impacts entailed by enforcing such control. Specifically, we show that enforcement of stable high activity rates by means of closed-loop control may enhance alteration of underlying global input–output relations and activity dependent dispersion of neuronal pair-wise correlations across the network.


Synchronization modes across long distances are then studied and manipulated by altering the characteristic length scale of the network, λ, using pharmacological disinhibition. In order to access long-range synchrony in these networks, an experimental model that enables monitoring of spiking activities over centimeter scale distances is constructed. We show that over such distances the length scale, λ, which is the minimal path that activity should travel through before meeting its point of origin ready for reactivation, dictates the synchrony mode. The value of λ is giving rise to either irregularly occurring network-wide events when λ is larger than the effective propagation dimension of the network, or to a continuous self-sustained reentry propagation, when λ decreases towards network dimension. The latter mode of reentrant propagation is a regular dynamical mode, marked by precise spatiotemporal patterns (`synfire chains') that may last many minutes. Termination of a reentry phase is preceded by a decrease of propagation speed to a halt. Stimulation decreases both propagation speed and λ values, which modifies the synchrony mode respectively. The parametrization provided contributes to the understanding of the origin and termination of different modes of neural synchrony and their long-range spatial patterns, while hopefully catering to manipulation of the phenomena in pathological conditions.

Overall, these results demonstrate a characterization of network short and long-range synchrony and its possible alteration by external inputs, to achieve stabilization and manipulation of neural activity. Moreover, possible functional impacts of such interventions are manifested, along with the critical role of addressing these questions in the relevant temporal and spatial scales.