听神经病患者的单音节识别率与言语声强的函数关系分析

目的分析听神经病（auditory neuropathy，AN）患者在不同测试强度下的单音节识别率，总结AN患者在言语识别方面的特点。方法AN患者10例，共20耳。根据听力图形状将AN患者分为上升型听力组和非上升型听力组，两组分别包括12耳和8耳。分别在患耳0．5、1、2和4kHz纯音平均听阈（4FA）以及4FA上10dB、20dB、30dB共4个强度测试单音节识别率。绘制得分一强度（Performance--Intensity，P—I）函数曲线。结果AN患者的P—I曲线大致可以分为3段。在昕阈附近得分较低，在4FA＋20dB处达到相对较高得分，平均最大言语识别率为42．8％±23．8％；在4FA＋20dB以上得分没有随强度增加显著增长，而是总体得分呈轻微回跌。大多数AN患者P—I曲线呈非单调性，且变异范围较大，最大言语识别率得分的差异可达60％～70％。在各强度下，上升型听力组和非上升型听力组在全部4个强度下得分无统计学差异。结论AN患者的单音节P—I曲线呈现低得分、非单调、变异大的特点。

An analysis of the Performance-Intensity function of patients with auditory neuropathy

Objective To analyze the Performance-Intensity function of patients with auditory neuropathy （AN） in mandarin monosyllable recognition test, and to summarize the characteristics of AN patients in speech recognition. Methods Ten AN patients （20 ears） were studied. All patients were divided into two groups i.e. reverse-slope curves （12 ears） and fiat curves （8 ears） according to their audiogram patterns. All patients were tested with mandarin monosyllable test CD at a series of speech level such as 4 FA, 4 FA＋10 dB, 4 FA＋20 dB, and 4 FA＋30 dB. The monosyllable performance scores of each ear at each level were recorded and Performance-Intensity （P-I） functions of each ear were plotted and analyzed. Results The P-I function of patients with AN can generally be described in terms of three main stages. At low-level stage 1 which was near average pure-tone threshold （4 FA）, the performance was rather poor. Then the score increased with the speech level at stage 2 until a stable but low score was obtained at 4 FA＋20 dB. At stage 3, the score did not increase significantly with speech level, but decreased slightly on the whole instead. Most P-I functions of AN patients did not increase monotonically but varied in a wide range and the maximum difference in speech recognition score can reach 60%-70%. At all 4 speech levels, AN patients with reverse-slope hearing loss performed the same as those with flat hearing loss. Conclusion The P-I functions of AN patients have three major characteristics, namely, low performance score, non-monotonicity and larqe variance.