Spiking Neural Networks (SNNs) utilize principles inspired by biological neural systems to enable energy-efficient AI through event-driven processing and temporally sparse activity. While models like the leaky integrate-and-fire (LIF) neuron capture basic dynamics, their abstraction overlooks biological complexities such as dendritic nonlinearities and oscillatory mechanisms. This work explores oscillatory neuron behavior, represented by adaptive LIF (adLIF) models with spike frequency adaptation (SFA), and dendritic integration in multi-compartment (MC) neurons. Oscillatory dynamics, enabled by feedback loops or resonator models, enhance temporal processing, while dendritic compartments enable single neurons to approximate multi-layer computations through plateaus initiated via dendritic action potentials (dAPs). Challenges in reconciling biological fidelity with hardware constraints are analyzed across neuromorphic platforms like Loihi, TrueNorth, and BrainScaleS, which trade off asynchronous processing, scalability, and analog/digital precision. The discussion advocates for hybrid oscillatory-dendritic models and hardware-software co-design to unlock SNNs’ potential in tasks requiring temporal integration and sequence detection. By integrating biological insights with neuromorphic engineering, this work outlines pathways to close the gap between abstract neuronal models and biological systems.