Phase stability of auditory steady state responses in newborn infants

This study examined the phases of auditory steady state responses (ASSRs) evoked by exponentially amplitude-modulated (AM) tones in 44 newborn infants (within 3 days of birth) and in 15 older infants (within 3 to 15 wks of birth). Our hypothesis was that the phases of the ASSRs would show orderly changes with modulation rate/carrier frequency and that this stability could be used with phase-biasing statistical techniques to augment response detection.

Multiple ASSRs were recorded to four modulated tonal carriers with intensities of 50 dB SPL, which were combined and presented simultaneously. The carriers of 0.5, 1, 2, and 4 kHz were modulated at rates between 78 and 95 Hz. Recordings lasted 12.3 mins. Data were analyzed offline with particular attention to phase and its possible exploitation in response detection using a phase-weighted t test (PWT). Population normative phase values were compared with self-normative values. The latter uses phase estimates from ASSRs that are detected at an earlier time to estimate expected phases of ASSRs, which have not yet been detected. This was implemented as an interstimulus phase-weighted t test (iPWT). A secondary analysis compared using fixed test durations where data were evaluated once at the end of the recording with variable test durations where data were evaluated after every sweep.

Average phases were not statistically different between the newborn and older infants. The mean ASSR phases across both infant groups were 10°, 36°, 83°, and 110° in the left ear and 78°, 97°, 135°, and 138° in the right ear for the four modulated carriers, respectively. Of a total of 172 detected ASSRs across the four carriers, 63% (109/172), 84% (144/172), and 99% (170/172) of the phase values fell within ±30°, ±45°, and ±90° of the population mean values, respectively. Self-normative phase values were slightly closer to actual measured phases, than population normative values. Compared with the F test, with a fixed duration, the iPWT technique did slightly better (71.7% versus 77.1% detected). Compared with the F test, with variable test duration, test time was reduced using the iPWT technique for normal and weighted averaging by 4 and 2.9 sweeps (66 and 48 secs), respectively, while false-positive rates were maintained. Compared with tests that relied on the F-ratio and a fixed time of 12.3 mins, using variable test times and the iPWT approach resulted in a halving of test time, while slightly improving comparable ASSR detection rates (66.7% versus 72.5%). An inter-ear average phase difference of 52° was found, which was not accounted for by modulation rates used for left/right ears. Converting phase to latency yielded similar results to prior studies.

The phase responses of ASSRs evoked by AM tones are stable in newborn and young infants. When using the multiple auditory steady state response (MASTER) technique, it is possible to employ phase-biasing methods to reduce test time and increase detection rates. Using self-normative intrastimulus phase difference values provides better estimated phases than average population phases for purposes of response detection.