An essential feature of the human brain is the ability to extract information from temporally distributed events in the auditory environment. We addressed this temporal encoding ability by modelling how stimulus rate is represented in the auditory cortex. We propose that a cortical representation of stimulus rate can be achieved through the oscillatory properties of nerve cells. Using non-invasive brain measures, we tested the prediction of the model that multiple rebound responses (RRs) occur after the ending of stimulation. MEG recordings revealed successive RRs which originated in the same areas as the N1s elicited by the auditory stimuli at latencies predicted by the model. Our modelling and experimental results therefore provide evidence that the temporal structure of the auditory environment is decomposed in human auditory cortex by cells performing bandpass filtering on periodic input.