Artikel
From Euler to the Neocortex – The potential of graph analytical approaches to characterise electrophysiological network properties of human cortical brain slice cultures
Von Euler zum Neocortex – Das Potential Graphen-Analytischer Ansätze zur Charakterisierung von Elektrophysiologischen Eigenschaften humaner Cortex-Hirnschnittkulturen
Suche in Medline nach
Autoren
Veröffentlicht: | 25. Mai 2022 |
---|
Gliederung
Text
Objective: Micro-electrode array (MEA) technique offers a possibility to record massive parallel electrophysiological potentials with spatial resolution. Thus, MEA recordings can be leveraged to investigate for network properties of brain slice cultures which could provide novel insights into the electrophysiological pathogenesis of epilepsy or tumour growth. However, the huge extent of the detected data makes analysis challenging and most published articles using MEAs only concentrate on general parameters including firing rate, number of active channels, or interspike-interval distribution. Here, we use graph analytical tools to describe connections of spatially distant areas of human organotypic neocortical slice cultures.
Methods: Spare tissue from resective epilepsy or tumour surgeries were cultured as organotypic cortical brain slice cultures. MEA recordings were obtained using a 256-MEA with either artificial or human cerebrospinal fluid (aCSF/hCSF). Custom Python scripts were used for data analysis. The main libraries for our network analytical approach include Numpy, Scipy, and Networkx. Spike times were extracted using a mean absolute deviation threshold-based approach without spikesorting. For each of the 252-recording channels, we then identified bursting episodes using an adaptive threshold. We then created a graph with each node equaling one MEA channel and each edge referring to shared bursting events. We then calculated the degree centrality, which is a metric for the number of connections a node has within a network that can be used to identify functional hubs within the anatomy.
Results: Channels with a high degree centrality were identified in all neocortical layers. Network properties changed depending on the used medium suggesting adaptive properties and possibly underlying subnetworks. Thus, hCSF led to an increase in degree centrality in Layers 2, 3, and 4. Further, novel nodes could be identified in hCSF that were not present in aCSF.
Conclusion: Using MEA technology to investigate human organotypic brain slice cultures offer a powerful tool for the identification and characterization of functional networks and their alterations in pathological conditions. Graph-theoretical approaches could prove as a useful extension of the analytical armoury, which allows deciphering not only spatially static but also directionally dynamic network properties and thus leveraging the possibility to examine the spread of ictal activity or tumour growth.