Complex networks approach to study mesoscopic neural network

• Autorzy: Nasuto S.J.
• Języki publikacji: EN

Abstrakty

EN

This work is a part of a wider research programme investigating the use of animats, hybrid systems consisting of robotic bodies controlled in a closed loop by neuronal cultures. The mesoscopic cortical cultured networks bridge the gap between low level neuronal properties and system-level behaviour. They are dependent on properties of single neurons, synapse types and connectivity patterns. At the same time the activity patterns emerging due to mesoscopic level interactions shape the entire brain dynamics. We studied the spontaneous development of mesoscopic cultures of rat cortical neurons using complex networks approaches. Cultures of cortical neurons were grown on Multi-Electrode Arrays and spontaneous activity was recorded during development (DIV’s 7-35). Functional connectivity was obtained from the culture-wide bursts and typical complex network statistics were estimated. Cultures start with a random pattern of interactions which nevertheless develop small world characteristic as cultures mature, analogously to findings from in vivo studies. Connectivity evolution reveals that the burst networks are not completely random, although they do not show temporal dependencies. The reported results form a benchmark against which the effects of a closed loop on development of network interactions can be assessed. Animats offer an attractive platform linking network activity to behaviour. It allows to investigate the effects of closed loop interactions on neurobiological, dynamical and complex networks properties as well as to elucidate the functional role of the repertoire of complex systems characteristics. Recent work though offers caveats regarding functional connectivity analysis (more broadly, any attempts to decode neural function from popular electrophysiological data analysis methods), prompting a need for more robust complex systems tools capable of unravelling variables playing causal roles in network function. FINANCIAL SUPPORT: This research was supported by an EPSRC grant (EP/D080134/1) “Investigating the Computational Capacity of Cultured Neuronal Networks Using Maching Learning”.

• Wydawca: -
• Czasopismo: Acta Neurobiologiae Experimentalis
• Rocznik: 2017
• Tom: 77
• Numer: Suppl.1
• Opis fizyczny: p.23-24

Twórcy

autor

Nasuto S.J.

• University of Reading, School of Biological Sciences, Biomedical Engineering Section, Whiteknights, UK