High-density single-neuron recordings show diverse tuning for acoustic and phonetic features across layers in human auditory speech cortex.