Towards Spike-Based Speech Processing: A Biologically Plausible Approach to Simple Acoustic Classification