Auditory nerve single-unit population studies have demonstrated that phase locking plays a dominant role in the neural encoding of steady-state speech sounds. Recently, we have reported that the phase-locked activity underlying the human frequency-following response (FFR) could also encode the first two formants of several tonal approximations of steady-state vowels. Since auditory nerve single-unit population studies have also demonstrated that phase locking is used to represent time-varying speech-like sounds, it was reasoned that the phase-locked neural activity underlying the human FFR, likewise, is dynamic enough to represent time-varying sounds. FFRs to a rising and a falling tone were obtained from 8 normal-hearing adults at 95, 85, 75 and 65 dB nHL. Results clearly demonstrated that the human FFR does indeed follow the trajectory of the rising and falling tones. Also, amplitude changes in the FFR supported the view that neural phase locking decreases with increasing frequency. Finally, the relatively smaller FFR amplitude for the falling tone compared to its rising counterpart lends further support to the notion that rising tones produce greater neural synchrony than falling tones. These results indicate that the human FFR may be used to evaluate encoding of time-varying speech sounds like diphthongs and certain consonant-vowel syllables.