A Biologically Inspired Supervised Learning Rule for Audio Classification with Spiking Neural Networks