In vitro neuronal cultures and network analysis to build a proof-of-concept
biological computing AI device

Fuster Palá, Auba

In vitro neuronal cultures and network analysis to build a proof-of-concept biological computing AI device

Citation

Fuster Palá, A. In vitro neuronal cultures and network analysis to build a proof-of-concept biological computing AI device. 2022. handle: http://hdl.handle.net/10230/54577

This citation was generated automatically.

Abstract

The limitations of current computation and artificial intelligence approaches have led to the emergence of biological computation, which has been studied in the last years as an alternative to traditional computation methods. The NeuChiP European project aims to develop a proof-of-concept for a biological computing AI device. The first step towards biological computation is generating in vitro neuronal cultures with the appropriate characteristics for using them as new AI systems. This thesis aimed to study different culture setups using both primary cultures and human induced pluripotent stem cells(iPSC) derived cortical neurons. Also, the cultures were recorded using calcium imaging and the results were analyzed from a network point of view to study both the activity and the connectivity of the cultures. The results suggested that it is possible to configure the culture functional connectivity by using physical constraints, chemical agents interfering in synaptic transmission, as well as mechanical agents like a random cut. Those have a direct effect on the activity and connectivity of the network and can be used to modulate it. However, after some time, the culture was able to adapt to the external changes thanks to the inherent plasticity of biological neuronal networks, which highlights the necessity of finding an effective and durable stimulation method. As expected, the reproducibility of the results obtained with primary culture using human iPSC-derived cortical neurons was long and painstaking due to the novelty of the protocol. To conclude, this thesis will be a starting point for future studies in the biological computing research field