Inspired by the biological nervous system, unsupervised spiking neural networks (SNNs) with the spike-timing-dependent plasticity (STDP) learning rule have been considered as the next-generation artificial neural networks (ANNs). However, to construct a functional SNN with high pattern recognition accuracy and low power consumption, hardware elements that present synaptic behavior still need to be developed. In this work, we studied Gd_0.3Ca_0.7MnO₃ (GCMO)-based memristive devices comprised of an asymmetrical electrode configuration, Al/GCMO/Au. We verified its switching properties, focusing on single pulse switching and its usability as artificial synapse by means of the STDP learning rule. The dynamic range is well controlled by the pulse amplitude and width, and the conductance change shows a clear dependence on the interval between the pulses. Moreover, pattern recognition accuracy (>87%) is obtained in biologically plausible unsupervised SNN simulations when the device characteristics are utilized as the synaptic weight in the network. The results shed some light on the complexity of the operation of the devices for utilization in unsupervised SNNs, that is, the evolution of the ANNs for which the first proof-of-concept is currently being reported. Additionally, the bioplausibility of the simulated network opens the door to considering biohybrid systems and their enormous application possibilities.