Abstract:Novel computing hardwares are necessary to keep up with today's increasing demand for data storage and processing power. In this research project, we turn to the brain for inspiration to develop novel computing substrates that are self-learning, scalable, energy-efficient, and fault-tolerant. The overarching aim of this work is to develop computational models that are able to reproduce target behaviors observed in in vitro neuronal networks. These models will be ultimately be used to aid in the realization of these behaviors in a more engineerable substrate: an array of nanomagnets. The target behaviors will be identified by analyzing electrophysiological recordings of the neuronal networks. Preliminary analysis has been performed to identify when a network is in a critical state based on the size distribution of network-wide avalanches of activity, and the results of this analysis are reported here. This classification of critical versus non-critical networks is valuable in identifying networks that can be expected to perform well on computational tasks, as criticality is widely considered to be the state in which a system is best suited for computation. This type of analysis is expected to enable the identification of networks that are well-suited for computation and the classification of networks as perturbed or healthy.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to develop computational models capable of reproducing the target behaviors of in vitro neuronal networks. These models will ultimately be used to implement these behaviors on substrates that are more amenable to engineering, such as arrays of nanomagnets. Specifically, the research focuses on how to evaluate the computational capabilities of neuronal networks by analyzing their electrophysiological recordings, especially on identifying whether the network is in a critical state, because the critical state is considered to be the state in which the system is most suitable for computation. The research background lies in the fact that current computing technologies are facing the problem of ever - increasing data processing and storage requirements, and the traditional von Neumann architecture has difficulty in应对 this challenge. Therefore, new types of physical computing architectures need to be developed. These architectures should be self - learning, scalable, energy - efficient, and fault - tolerant. The brain, as a self - organizing system, demonstrates powerful computational capabilities with extremely low energy consumption, which provides inspiration for the design of new computing hardware. The paper presents an analysis method based on the size distribution of neuronal avalanches to evaluate whether in vitro neuronal networks are in a critical state. Preliminary analysis results show that this method can effectively distinguish between critical and non - critical state networks, which is of great value for determining which networks are suitable for computational tasks. In addition, this research is also expected to provide insights into the functions of neuronal networks under healthy and perturbed conditions, thereby providing a reference for understanding how different diseases affect neuronal connections and communications. In summary, this research aims to develop new computational models and hardware by analyzing the behaviors of in vitro neuronal networks to meet the future demands for high - performance computing and to provide tools for further understanding the working mechanisms of the nervous system.

Evaluation of the criticality of in vitro neuronal networks: Toward an assessment of computational capacity