Neuromorphic silicon neuron circuits are hybrid analog/digital VLSI circuits that emulate the electrophysiological behavior of real neurons and conductances. These circuits are used to implement a wide range of neuromorphic systems, including real-time, low-power systems for brain-inspired computational solutions. The paper describes various circuit designs and techniques used to implement silicon neurons, covering different computational models, from biophysically realistic Hodgkin-Huxley models to generalized adaptive integrate-and-fire models. It compares different design methodologies and presents experimental results from fabricated VLSI chips. The paper discusses the functional components of silicon neurons, including synapse blocks that receive and integrate spikes, and soma blocks that perform spatio-temporal integration. It also covers dendritic and axon circuit blocks that model signal propagation along neuronal fibers. The paper presents various design styles, including current-mode, sub-threshold, and voltage-mode designs, and discusses their advantages and applications. The paper also describes specific circuits used in silicon neuron implementations, such as the Tau-Cell, DPI, and Axon-Hillock circuits. These circuits are used to generate spike events, control spiking thresholds, and implement spike-frequency adaptation. The paper also discusses the use of thermodynamic models and phenomenological models to simulate neuron behavior. The paper presents several examples of silicon neuron implementations, including the Thalamic relay neuron, a sub-threshold Hodgkin-Huxley based neuron, and compact integrate-and-fire circuits for event-based systems. These implementations demonstrate the versatility of neuromorphic silicon neuron circuits in various applications, including neuromorphic vision sensors and dynamic vision sensors. The paper highlights the importance of power efficiency, matching, and scalability in the design of these circuits.Neuromorphic silicon neuron circuits are hybrid analog/digital VLSI circuits that emulate the electrophysiological behavior of real neurons and conductances. These circuits are used to implement a wide range of neuromorphic systems, including real-time, low-power systems for brain-inspired computational solutions. The paper describes various circuit designs and techniques used to implement silicon neurons, covering different computational models, from biophysically realistic Hodgkin-Huxley models to generalized adaptive integrate-and-fire models. It compares different design methodologies and presents experimental results from fabricated VLSI chips. The paper discusses the functional components of silicon neurons, including synapse blocks that receive and integrate spikes, and soma blocks that perform spatio-temporal integration. It also covers dendritic and axon circuit blocks that model signal propagation along neuronal fibers. The paper presents various design styles, including current-mode, sub-threshold, and voltage-mode designs, and discusses their advantages and applications. The paper also describes specific circuits used in silicon neuron implementations, such as the Tau-Cell, DPI, and Axon-Hillock circuits. These circuits are used to generate spike events, control spiking thresholds, and implement spike-frequency adaptation. The paper also discusses the use of thermodynamic models and phenomenological models to simulate neuron behavior. The paper presents several examples of silicon neuron implementations, including the Thalamic relay neuron, a sub-threshold Hodgkin-Huxley based neuron, and compact integrate-and-fire circuits for event-based systems. These implementations demonstrate the versatility of neuromorphic silicon neuron circuits in various applications, including neuromorphic vision sensors and dynamic vision sensors. The paper highlights the importance of power efficiency, matching, and scalability in the design of these circuits.