Neural Mechanisms Supporting Robust Discrimination of Spectrally and Temporally Degraded Speech