Abstract
Linear and non-linear TEOAE protocols were compared in terms of nine parameters in order to define the protocol producing recordings with the highest signal quality (lowest noise and highest signal-to-noise ratio). The pilot project acquired data using ILO-92 apparatus from 220 neonates (397 ears) at the second/third day after birth in three European laboratories. A Gabor spectrogram time-frequency representation of the recordings showed considerable frequency dispersion in TEOAE latencies >4.0 ms. The data, analysed with a Wilcoxon test, indicated that a linear TEOAE protocol: (i) generates recordings of a lower noise and a higher signal-to-noise (S/N) ratio in the 2.0, 3.0, 4.0 kHz TEOAE bands; (ii) the increase in the S/N ratio can result in a decrement of the required number of TEOAE sweeps; (iii) the higher values of S/N can be used in the estimation of more robust pass-fail criteria, minimizing the percentage of false positives and negatives.