Network-on-Chip paradigm is an emerging technology that provides communication between the Intellectual Property cores in System-on-Chip (SoC). The drastic development of nanotechnology supports the making of Intellectual property cores with small area, less power dissipation, improvement in latency and achieves high speed. The neural systems that emulate biological sensory capability on reconfigurable hardware are the key requirement in the field of Neuro-Engineering. Nano-neural connectivity technique is highly needed in order to implement reconfigurable Neural Networks (NNs) on-chip. This paper proposes a novel Nano Machine based Hybrid 3D Network-on-Chip for the emulation of a third-generation (3G) Neural Network model called Spiking Neural Network (SNN). This Hybrid Network-on-Chip uses a Nanomachine Transceiver with Carbon nanotube antenna which uses the novel pulse-based technique to transfer the packets between the subnet of NoC to reduce the latency. The experimental result of this work shows the improvement in latency of Nano machine-based NoC compared to the previously reported work for 18 nm TSMC technology.