Figure Legends
Fig. 1 (a) Hearing thresholds, (b) WRS in quiet, and (c) SRT in the noise of patients with UMA in unaided and aided conditions, as well as the controls. Group means are presented as mean ± SD. Significant differences are labelled with *(p < 0.05) and **(p < 0.01) and ns, not significant. WRS, word recognition score; SRT, speech reception threshold; SNR, speech-to-noise ratio; HSE, head shadow effect; NH, normal hearing; BCD, bone conduction device; SD, standard deviation.
Fig. 2 Sound-localisation target–response plots of two patients (P6 and P7) and one control (N2) in monaural (unaided and P, left column) and binaural (aided and UP, right column) listening conditions. Stimulus sound levels are indicated by black circle, white square, and cross data points (65-, 70-, and 75-dB SPL, respectively). Best-fit linear regression is indicated by a black line. For participants with an ideal optimal localisation ability, g is 1, whereas MAE and b are 0. MAE, mean absolute error; g, response gain; b, bias; r2, R square.
Fig. 3 (a) The MAE and (b) Gain under monaural listening conditions (unaided and P, Y-axis) are plotted against those under binaural listening conditions (aided and UP, x-axis). Black circle data points indicate eleven UMA children, and white circle data points indicate the controls. The two UMA children and one control depicted in Fig. 2 are marked in this figure (P6, P7, and N2). An MAE near 0 and a gain near 1 demonstrate a close-to-normal sound localisation performance. Data of participants with the same sound localisation performance when listening monaurally and binaurally are displayed on the grey dotted diagonal. A data point below the diagonal in Fig. 3a and above the diagonal in Fig. 3b represents a better sound localisation performance when listening under binaural conditions than under monaural conditions. (c–d) The mean MAE of patients with UMA and the controls, respectively, on the impaired (including the atretic side of UMA and the plugged side of controls) side and the contralateral (the healthy side of UMA and the unplugged side of controls) side. Error bars represent mean ± SD. UMA, unilateral microtia and atresia; MAE, mean absolute error; Gain, response gain; Unaided, the unaided condition of patients with UMA; Aided, the aided condition of patients with UMA; P, the plugged condition of control listeners; UP, the unplugged condition of control listeners; SD, standard deviation.
Fig. 4 (a) MAE outcomes of two subgroups of good performers (n = 5; gain > 0.75) and poor performers (n = 6; gain ≤ 0.75) were separately compared on the atretic and healthy sides. (b) Individual data of response gain of UMA children in unaided conditions are plotted as a function of delta MAE between unaided and aided conditions. The linear regression was conducted to explore the predictive effect of gain on the benefits of sound localisation accuracy by fitting BCDs. P6 and P7 depicted in Fig. 2 are marked in this figure. Delta MAE = MAEaided – MAE unaided. Significant differences are labelled with *(p < 0.05) and **(p < 0.01), and ns, not significant. MAE, mean absolute error.
Supplementary Fig. 1 Test setup and listening conditions. (a) Seven loudspeakers were placed at 30° intervals in a semicircle in a double-walled, soundproof laboratory. (b) The monaural (the left column) and binaural listening (the right column) conditions are designed for UMA patients: unaided and aided conditions and for controls: P and UP conditions. P, the plugged condition; UP, the unplugged condition.