Abstract

Understanding how neural circuits in the brain are formed and function is a major goal of many neuroscience projects. The brain is composed of nerve cells called neurons, with approximately 16 billion neurons in the human cerebrum and around 86 billion neurons in the entire brain. These neurons form highly complex networks through connections known as synapses. Neurons fire and generate spikes due to the potential difference across ion channels inside and outside the cell. Neuronal signals that generate continuous spike sequences are referred to as being in a burst state. In many studies of neuron activity patterns, the detection and analysis of these burst patterns are central because the correlation of these burst patterns is believed to play an important role in the brain's information processing, such as information transmission processes.While these analysis techniques are effective for investigating current brain activity patterns, they are too complex to observe the influence of the entire network of neuron groups. The applicant believes that new indicators and evaluation methods that can more simply evaluate the activity patterns of neuron groups are necessary.In this study, we propose a method to visualize and quantify the characteristics of neuron group activity patterns using nonlinear time series analysis. We apply nonlinear time series analysis techniques such as attractor embedding, Lyapunov exponents, and fractal dimensions to actual biological signals.