Auditory nerve single-unit population studies have demonstrated that phase-locking plays a dominant role in the neural encoding of the spectrum of speech sounds. Given this, it was reasoned that the phase-locked neural activity underlying the scalp-recorded human frequency-following response (FFR) might preserve information about certain acoustic features of speech sounds. It was recently reported (Ananthanarayan, A.K., 1999. J. Audiol. Neurootol. 4, 95-103) that the FFR spectrum to simple two-tone approximations of several English back vowels does indeed contain peaks corresponding to the first and second formant frequencies. In this investigation FFRs to the more complex steady-state synthetic English back vowels (/u/, /)/, and /a/) were evaluated. FFRs were obtained from 10 normal-hearing human adults at 85, 75, 65, and 55 dB normal-hearing level (nHL). Spectrum analyses of the FFRs revealed distinct peaks at harmonics adjacent to the first and the second formants across all levels suggesting that phase-locked activity among two distinct populations of neurons is indeed preserved in the FFR. For each vowel the spectral peaks at first formant harmonics dominated the spectrum at high stimulus levels suggesting formant capture. The observation of less robust peaks for harmonics between the formants may very well suggest selective suppression to enhance spectral peaks at the formant frequencies. These results suggest that the scalp-recorded FFR may provide for a non-invasive analytic window to evaluate neural encoding of speech sounds in the brainstem of normal-hearing individuals and how this encoding may be degraded subsequent to cochlear hearing impairment.