Xu L, Furukawa S, Middlebrooks J C
Kresge Hearing Research Institute, Department of Otolaryngology, University of Michigan, Ann Arbor 48109-0506, USA.
Acta Otolaryngol. 2000 Mar;120(2):263-6. doi: 10.1080/000164800750001062.
Frequency transformation by the external ears provides the spectral cues for localization of broadband sounds in the vertical plane. When human subjects listen to spectrally-impoverished narrowband sounds presented in a free field, the perceived locations vary with the centre frequency and are largely independent of the actual source locations. The present study explored the substrate of spatial illusion by examining the responses of cortical neurons to narrowband stimuli. Single-unit responses were recorded in area A2 of anaesthetized cats. Broadband noise bursts were presented at 14 locations in the vertical median plane, from 60 degrees below the front horizon, up and over the head, to 20 degrees below the rear horizon. Narrowband (1/6-oct) noise bursts were presented at + 80 degrees elevation. An artificial neural network was trained to recognize the spike patterns elicited by broadband noise and, thereby, to register the spike patterns with sound-source elevation. When the trained network was presented with neural responses elicited by narrowband noise, the elevation estimated by the neural network varied with the centre frequency of the narrowband stimuli. Consistent with psychophysical results in human, the locations associated with a given centre frequency could be predicted by comparing the stimulus spectrum with the directional transfer functions of the cat's external ear. The results support the hypothesis that full spike patterns (including spike counts and spike timing) of cortical neurons code information about sound location and that the auditory cortical neurons play a pivotal role in localization behaviour.
外耳的频率转换为垂直平面中宽带声音的定位提供了频谱线索。当人类受试者聆听自由场中频谱匮乏的窄带声音时,感知到的位置会随中心频率而变化,并且在很大程度上与实际声源位置无关。本研究通过检查皮层神经元对窄带刺激的反应来探索空间错觉的基础。在麻醉猫的A2区域记录单单位反应。宽带噪声脉冲在垂直中平面的14个位置呈现,从前地平线以下60度,向上越过头顶,到后地平线以下20度。窄带(1/6倍频程)噪声脉冲在仰角+80度处呈现。训练一个人工神经网络来识别由宽带噪声引发的尖峰模式,从而将尖峰模式与声源仰角进行匹配。当用窄带噪声引发的神经反应来测试训练好的网络时,神经网络估计的仰角会随窄带刺激的中心频率而变化。与人类的心理物理学结果一致,通过将刺激频谱与猫外耳的方向传递函数进行比较,可以预测与给定中心频率相关的位置。这些结果支持了以下假设:皮层神经元的完整尖峰模式(包括尖峰计数和尖峰时间)编码有关声音位置的信息,并且听觉皮层神经元在定位行为中起关键作用。