Isojima G, Suzuki T, Ishigami H
Department of Otorhinolaryngology, Nagoya Teishin Hospital, Aichi, Japan.
Auris Nasus Larynx. 1995;22(1):16-23. doi: 10.1016/s0385-8146(12)80177-5.
Frequency modulation (FM) sounds were used as simultaneous maskers to estimate the psychophysical tuning curve (PTC). Maskers (800 ms duration) were FM sounds sweeping upward (1.2-->2.8 kHz) in Experiment I and downward (2.8-->1.2 kHz) in Experiment II. The signal was a pure tone of 2 kHz and the signal levels were 15-, 20-, 30-, and 35-dB SL. Masked PTCs were obtained as a function of masker frequency. The results were as follows: (i) When an upward-sweeping FM sound is used for high signal level, the slope of the PTC is shallower on the low frequency side and steeper on the high frequency side, and the 10-dB bandwidth of PTC coincides with that of the restricted PTC. (ii) When a downward-sweeping FM sound is used for the high signal level, the above relation for the PTC slope is reversed, and the 10-dB bandwidth for PTC coincides with that of unrestricted PTC. (iii) PTCs determined by masking in the presence of FM sounds apparently restrict the subject to listening to off-frequency, and it is shown that the bandwidth of PTC properly reflects the critical band. (iv) The properties of these PTCs can be easily explained by a neural network model of lateral inhibition, so we can accept a hypothesis that the hierarchical processing mechanism is incorporated within the inferior colliculus for parallel signal processing. (v) We conclude that the frequency resolution and the critical band filtering are, therefore, realized in a psychophysically relevant way in the auditory midbrain level.
调频（FM）声音被用作同步掩蔽声来估计心理物理调谐曲线（PTC）。在实验I中，掩蔽声（持续时间800毫秒）是向上扫频（1.2→2.8千赫）的FM声音，在实验II中是向下扫频（2.8→1.2千赫）的FM声音。信号是2千赫的纯音，信号强度为15、20、30和35分贝感觉级（dB SL）。根据掩蔽声频率获得掩蔽后的PTC。结果如下：（i）当高信号强度使用向上扫频的FM声音时，PTC在低频侧的斜率较浅，在高频侧较陡，且PTC的10分贝带宽与受限PTC的带宽一致。（ii）当高信号强度使用向下扫频的FM声音时，PTC斜率的上述关系相反，且PTC的10分贝带宽与非受限PTC的带宽一致。（iii）在FM声音存在下通过掩蔽确定的PTC显然限制受试者聆听偏离频率的声音，并且表明PTC的带宽恰当地反映了临界带宽。（iv）这些PTC的特性可以通过侧向抑制的神经网络模型轻松解释，因此我们可以接受这样一个假设，即在下丘中纳入了分层处理机制用于并行信号处理。（v）我们得出结论，因此，频率分辨率和临界带滤波在听觉中脑水平以与心理物理学相关的方式得以实现。
