[[alternative]]The effect of individually measured real-ear-to-coupler difference and types of test signals on Desired Sensation Level method hearing aids verification for young children

[[abstract]]本研究之目的為探討使用個別測得真耳耦合器差值以及使用不同類型的測試音對於幼兒適宜感音級選配法助聽器驗證的影響。實驗一之受測對象為31名感音神經性聽損幼兒個案，年齡為0至7歲，以訂製型耳模個別量測真耳耦合器差值，計算其與海綿耳塞及訂製型耳模兩組真耳耦合器差值常模之間的差異，並計算三種不同聽損程度聽力圖，分別以個別量測與使用常模兩種方式所換算DSL 5長期平均語音頻譜放大後目標值的差異。實驗二以33個耳掛型助聽器為樣本，量測助聽器對於四種不同非語音測試音和數位錄製真實語音的輸出聲壓值差異，以及增益值差異。 研究結果顯示RECD的再測差異平均值接近0 dB，RECD的兩耳間差異平均值亦接近0 dB。個別測得真耳耦合器差值與DSL 5內建RECD常模，在高頻4000 Hz有-10至-15 dB的差異，其中個別測得RECD與訂製型耳模RECD常模的整體差異較小。由個別測得RECD及DSL 5內建RECD常模兩者換算所得的LTASS放大後目標值，在高頻4000和6000 Hz會有較大的差異，其中又以重度聽力損失個案的目標值換算差異較大。在測試音方面，使用純音來預估語音的放大後聲壓值，將會高估約10 至20 dB。pink noise在高頻2000和4000 Hz會高估語音的放大後聲壓值，digi ICRA則相反，會低估語音的放大後聲壓值。 在臨床嬰幼兒助聽器選配驗證過程中，量測RECD是一個重要的步驟，應該盡可能個別量測RECD，假使無法順利個別測得RECD，應選用訂製型耳模RECD常模。對於中耳功能正常且無外耳道畸形的個案，其任一耳測得的RECD可應用於對側耳，其驗證誤差會比使用常模要小。驗證助聽器時所使用的測試音是否具有語音頻譜加權，以及是否具有語音變動特性，皆會影響最後驗證的結果，以測試音放大後的聲壓值或者增益值為驗證目標，亦會影響驗證結果的判讀。不同的臨床選配單位在驗證助聽器時，需要將以上因素一併納入考量，以減少驗證結果的差異。[[abstract]]The goal of this study is to estimate the effects of using individually measured real-ear-to-coupler difference and types of test signals on Desired Sensation Level Method hearing aids verification for young children.Thirty-one children with sensori-neural hearing loss served as participants in Experiment I. The age of subjects ranged from 0 to 7 years old. The RECD were individually measured with the customized earmold. The measued RECD were compared to mold RECD norms and foam-tip RECD norms of DSL 5. The LTASS targets of three types of audiogram were generated by DSL 5 fitting algorithm using measured RECD and norms, moreover, the target differences were computed. In Experiment II, the SPL differences, as well as the gain difference, were measured between four non-speech signals and digitally-recorded real speech for thirty-three hearing aids.Results of Experiment I indicated that mean test re-test difference of individually measured RECD was close to 0 dB and the difference between right and left ears was also close to 0 dB. The difference between measured RECD and norms of DSL 5 was in -10 to -15 dB range at 4000 Hz, and the overall difference between measured RECD and mold RECD norms of DSL 5 was smaller. Furthermore, the LTASS targets generated by DSL 5 fitting algorithm using measured RECD and norms had larger difference at 4000 to 6000 Hz, especially for the subjects with severe hearing loss. Results of Experiment II indicated that pure tone had over-estimated the SPL of aided speech approximately 10 to 20 dB, moreover, pink noise had over-estimated the SPL of aided speech at 2000 and 4000 Hz, and inversely, digi ICRA had under-estimated at the same frequency range.In conclusions, during the child hearing aids fitting procedures, measuring RECD is a crucial step and we should measure it as possible as we can. If RECD can’t be obtained individually, it is preferable to use mold RECD norms of DSL 5. For the subjects with normal middle ear function and external ear canal, the RECD measured in one ear could be used to derive real ear SPL of the other ear, the difference would be smaller than using the norms of DSL 5. Whether the test signals have speech weighted spectrum and dynamic characteristic of speech will affect the results during hearing aids verification. Furthermore, verifying the aided test signals based on hearing aids output SPL or gain is considerable to hearing aids setting adjustment. To reduce the variation of verification results among different clinic, the factors mentioned above is considerable when verifying the hearing aids