The MAndarin spoken word—Picture IDentification test in noise—Adaptive (MAPID-A) measures subtle speech-recognition-in-noise changes and spatial release from masking in very young children

BackgroundSpatial release of masking (SRM) is a measure of an individual's ability to perform speech-noise segregation, which is usually quantified by the extent of improvement of the individual's speech recognition performance when the noise is switched from a spatially co-located position (e.g., speech and noise both presented from the front) to a spatially separated position (e.g., speech presented from the front and noise presented from the right side) with reference to the target speech. SRM is a combined measure of head shadow and binaural unmasking benefits. SRM has only been investigated in young children at group level but not at individual participant level in the international literature due to the lack of reliable speech recognition test materials able to detect subtle statistically significant within-participant changes in speech-recognition-in-noise thresholds.MethodThe performance to signal-to-noise ratio (P-SNR) functions of twenty-four disyllabic words were obtained from 40 native Mandarin-speaking children aged 3.6-6.2 years with reported normal speech, language and hearing. The test items' difficulty levels were homogenized by adjusting the speech intensity level of each item so that the adjusted signal-to-noise ratio for 50% correct score (SNR-50%) point of each item would overlap at the mean SNR-50% point of all test items. In the MAPID-A, the homogenized test items were randomly presented in an adaptive testing procedure at a fixed noise intensity level, but the speech intensity level of the upcoming test item varied in 2-dB SNR steps depending on the recognition result of the previous test item. The SNR reversal point is marked by a change from a decrease to an increase in the SNR or vice versa. Two successive SNR reversal points marked the boundaries of an excursion. The mid-points from 12 excursions (in dB SNR) were averaged to produce the adaptive SNR-50% measure (aSNR-50%).ResultsThe aSNR-50% results were obtained from another 12 children aged 4.8-5.3 years with reported normal speech, language and hearing. The average 99% confidence interval (CI) of all participants' mean aSNR-50% values was ±1.61 dB SNR; therefore, 3.22 dB SNR was the average critical difference required to confirm a significant difference in the scores obtained from the same participant between two test conditions. Statistically significant within-participant SRM was identified in 95% of the participants; in other words, aSNR-50% obtained from the spatially separated condition outperformed aSNR-50% obtained from the spatially co-located condition. The adaptive testing procedure was highly reliable, with an within-participant test-retest reliability of 90.6%. and significantly limited testing time to an average of 4.2 min. This research study has fulfilled its aim on detecting subtle within-participant SRM in very young children starting from 4 years of age with a reliable statistical procedure. MAPID-A offers a reliable and efficient clinical tool to investigate speech-recognition-in-noise and SRM performances in young Mandarin-speaking children.ConclusionsThe narrow CIs, high test-retest reliability, and short testing time has proven that the MAPID-A is a promising sensitive, reliable and time-efficient clinical tool to detect subtle within-participant speech-recognition-in-noise changes in children as young as 4-5 years. The MAPID-A offers a clinical tool to behaviorally track young children's development in speech-recognition-in-noise and SRM, and to potentially review the development of the auditory neural pathway and the cerebral dominance for speech-recognition-in-noise in young children.