Department of Biophysics, Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, 6525 EZ Nijmegen, The Netherlands.
J Neurosci. 2010 Jan 6;30(1):194-204. doi: 10.1523/JNEUROSCI.2982-09.2010.
To program a goal-directed orienting response toward a sound source embedded in an acoustic scene, the audiomotor system should detect and select the target against a background. Here, we focus on whether the system can segregate synchronous sounds in the midsagittal plane (elevation), a task requiring the auditory system to dissociate the pinna-induced spectral localization cues. Human listeners made rapid head-orienting responses toward either a single sound source (broadband buzzer or Gaussian noise) or toward two simultaneously presented sounds (buzzer and noise) at a wide variety of locations in the midsagittal plane. In the latter case, listeners had to orient to the buzzer (target) and ignore the noise (nontarget). In the single-sound condition, localization was accurate. However, in the double-sound condition, response endpoints depended on relative sound level and spatial disparity. The loudest sound dominated the responses, regardless of whether it was the target or the nontarget. When the sounds had about equal intensities and their spatial disparity was sufficiently small, endpoint distributions were well described by weighted averaging. However, when spatial disparities exceeded approximately 45 degrees, response endpoint distributions became bimodal. Similar response behavior has been reported for visuomotor experiments, for which averaging and bimodal endpoint distributions are thought to arise from neural interactions within retinotopically organized visuomotor maps. We show, however, that the auditory-evoked responses can be well explained by the idiosyncratic acoustics of the pinnae. Hence basic principles of target representation and selection for audition and vision appear to differ profoundly.
为了针对嵌入在复杂声音场景中的声源进行目标导向的定向反应编程，听觉运动系统应该在背景中检测并选择目标。在这里，我们专注于该系统是否能够分离中矢状面（高度）中的同步声音，这是一项需要听觉系统分离耳廓诱导的频谱定位线索的任务。人类听众在中矢状面的各种位置上，快速地将头转向单个声源（宽带蜂鸣器或高斯噪声）或同时呈现的两个声源（蜂鸣器和噪声）。在后一种情况下，听众必须定向到蜂鸣器（目标）并忽略噪声（非目标）。在单音条件下，定位是准确的。然而，在双音条件下，响应端点取决于相对声级和空间差异。最响亮的声音主导了反应，无论它是目标还是非目标。当声音具有大致相等的强度并且它们的空间差异足够小时，端点分布可以通过加权平均很好地描述。然而，当空间差异超过大约 45 度时，响应端点分布变得双峰。在视动实验中也报告了类似的反应行为，对于这些实验，平均和双峰端点分布被认为是由于在视网膜组织化视动图内的神经相互作用引起的。然而，我们表明，听觉诱发的反应可以很好地用耳廓的特殊声学来解释。因此，听觉和视觉的目标表示和选择的基本原理似乎存在很大差异。
