Short latency afferent inhibition: comparison between threshold-tracking and conventional amplitude recording methods