We present an integrated spiking neural network (SNN) of the Basal Ganglia (BG) and Thalamo-cortico-thalamic (TCT) loop implemented on the neuromorphic hardware. Our objective is to demonstrate neuromorphic learning and decision-making facilitated by the neuromodulator Dopamine (DA). Referred to as the BG-TCT SNN, it consists of three modules viz. BG, thalamic and cortical modules that are based on existing implementations, albeit integrated here for the first time. DA- modulated learning of plastic parameters in selected pathways and heuristic adjustments of static parameters in the network is done iteratively. The output from all populations of the BG-TCT SNN oscillates in their respective base states, conforming to existing research, when there is background noise inputs to the thalamic and cortical modules and absence of visual stimuli. During testing with synthetic visual cues comprising of periodic spike train inputs at frequencies \(f\in [5, 20]\) Hz, all parameters are held static and DA-modulation is disabled. We observe that the base-state network can make decisions distinctively in response to visual cues with \(f\in [5, 15]\) Hz when performance accuracy is 90%. The performances in all observed categories drops for \(f > 15\) Hz. Future work will enhance the BG-TCT SNN for learning high frequencies and complex visual scenes. Energy consumption for computing the network on a local board is \(\approx \) 2–3 Watt-hour, an order of magnitude improvement over remote execution on the server.