China 
Africa 
Explore chapters and articles related to this topic
High-Performance and Customizable Bioinformatic and Biomedical Very-Large-Scale-Integration Architectures
Published in Tomasz Wojcicki, Krzysztof Iniewski, VLSI: Circuits for Emerging Applications, 2017
Yao Xin, Benben Liu, Ray C.C. Cheung, Chao Wang
The study of neural coding refers to searching and characterizing the correlations between neural action potentials or spikes and external-world sensory stimuli, motor actions, or mental states. The coding involves encoding and decoding corresponding to two opposite mapping directions [9]. The encoding is to predict the spike response of neurons to various types of stimuli. Conversely, the decoding problem concerns the estimation of the stimulus that gave rise to certain neuronal pattern. The neural coding algorithms need to characterize the models of neural spiking activity, which is a stochastic point process. Therefore, statistical theory, probability methods, and stochastic point processes are always applied in this field.
Variable-delay feedback control for stabilisation of highly nonlinear hybrid stochastic neural networks with time-varying delays
Published in International Journal of Control, 2023
Ailong Wu, Han Yu, Zhigang Zeng
Artificial neural network (ANN) has developed extremely rapidly in the past two decades. The successful application of ANN in many research fields has attracted extensive attention (e.g. Jin & Liu, 2008; Lin & Chen, 2009; Peng et al., 2008). ANN is essentially a sort of physical circuit model abstracted from biological neural network based on the understanding of human brain network. ANN encountered in white noise circumstance is inevitable. Actually, any brain nervous system is in the actual white noise environment, whose synaptic conduction is a noise process caused by neurotransmitters and other random fluctuations. In addition, the limited transmission speed of neuroelectric signals and the delayed release of neurotransmitters in synapses indicate that the neurodynamics process in space–time is a time-delay operation. Therefore, the real neural network should be a random time-delay dynamic system (Liao & Mao, 1996a, 1996b). In particular, the cluster discharge in neuronal discharge activity shows the mutual transformation between resting state and repeated discharge state, which is a typical multi-scale fast and slow dynamic phenomenon. This kind of complex oscillation is difficult to meet the linear growth mode, which is closer to a polynomial growth mode. Thus, we need to deeply explore the dynamic characteristics of neurodynamic solutions (including synchronisation characteristics, structural characteristics, control characteristics, etc.) and the effects of various brain electrical excitation and internal parameters on neural discharge activities, such as high nonlinearity, randomness and time-delay process. Hence, constructing cross-level real brain network dynamic models, such as highly nonlinear time-delay hybrid stochastic neural network (SNN) model whose coefficients grow polynomially, and then clarifying the structural laws and dynamic characteristics, has important guiding significance for exploring the neural coding principles and methods based on neurodynamics.