Development and evaluation of the paediatric Malay matrix sentence test for Malay native speakers aged 6 to 12 years old [EMBARGOED]

Abstract

This thesis discusses the development of a novel speech in-noise test designed specifically for Malay-speaking children that could enrich the existing audiologic test battery for the paediatric population in Malaysia. The aim of the study is to develop and evaluate the Matrix Sentence Test in Malay (PaedMalayMST) for children aged six to 12 years old. The development of the test involved the construction of testing materials, normalisation of the speech audio stimuli and evaluation of the developed test. The testing materials were derived from a constructed Paediatric Malay corpus, incorporating both Malay children’s written and spoken corpora. A six-by-three matrix was utilised to generate pseudo-sentences composed of numbers, objects, and adjectives (colours). A total of 54 pictures were created in the form of black and white line drawings and colours. These pictures were designed for a closed-set response format which consisted of three different versions of the 18 words in the word-based matrix. The audio speech stimuli were recorded, edited, and paired with two background noises: test-specific noise (TSN) and six-talker babble noise (BN). The normalisation study of the recorded speech materials in TSN and BN were conducted on 20 normal hearing adults. Twenty children aged between six to 12 years old were recruited to normalise the picture materials. Loudness adjustments were made within a ± 3 dB limit, leading to the generation of nine test lists of 20 sentences. A picture-based matrix of 18 validated pictures were chosen based on familiarity and ease of identification by the target age group. The evaluation of the PaedMalayMST focused on assessing the test lists equivalency, training effects, and establishing normative values. A total of 34 adults and 29 children were evaluated monaurally in a closed-set response format in both TSN and BN. All test lists exhibited equal intelligibility with a mean speech reception threshold (SRT) and slope value of -13.8 ± 0.8 dB SNR and 13.1 ± 2.9 %/dB in TSN and -14.4 ± 1.4 dB SNR and 11.9 ± 2.4 %/dB for BN. A list of 20 sentences were recommended for training in both TSN and six-talker BN. The established monaural closed-set test format normative values for TSN are -11.64 ± 1.1 dB SNR (six- to 10-year-old), -12.76 ± 0.9 dB SNR (11- to 12-year-old), and -13.67 ± 0.7 dB SNR (adults). Meanwhile, the monaural closed-set test format normative values for BN are -10.29 ± 0.9 dB SNR (six- to 10-year-old), -12.82 ± 1.4 dB SNR (11- to 12-year-old), and -14.69 ± 1.4 dB SNR (adults). In conclusion, the PaedMalayMST demonstrated reliability and accuracy for assessing speech perception in normal hearing children aged six to 12 in a closed-set response format, making it a valuable tool for assessing speech perception in a closed-set response format.