Portfors C V, Roberts P D
School of Biological Sciences, Washington State University, Vancouver, WA 98686, USA.
Oregon Health & Science University, Portland, OR 97239, USA.
Neuroscience. 2014 Jan 31;258:192-203. doi: 10.1016/j.neuroscience.2013.11.012. Epub 2013 Nov 16.
Neurons in the auditory system are spatially organized in their responses to pure tones, and this tonotopy is expected to predict neuronal responses to more complex sounds such as vocalizations. We presented vocalizations with low-, medium- and high-frequency content to determine if selectivity of neurons in the inferior colliculus (IC) of mice respects the tonotopic spatial structure. Tonotopy in the IC predicts that neurons located in dorsal regions should only respond to low-frequency vocalizations and only neurons located in ventral regions should respond to high-frequency vocalizations. We found that responses to vocalizations were independent of location, and many neurons in the dorsal, low-frequency region of IC responded to high-frequency vocalizations. To test whether this was due to dorsal neurons having broad frequency tuning curves, we convolved each neuron's frequency tuning curve with each vocalization, and found that the tuning curves were not good predictors of the actual neural responses to the vocalizations. We then used a nonlinear model of signal transduction in the cochlea that generates distortion products to predict neural responses to the vocalizations. We found that these predictions more closely matched the actual neural responses. Our findings suggest that the cochlea distorts the frequency representation in vocalizations and some neurons use this distorted representation to encode the vocalizations.
听觉系统中的神经元在对纯音的反应上呈空间组织分布,并且这种音频定位预计可预测神经元对诸如发声等更复杂声音的反应。我们呈现了具有低频、中频和高频成分的发声,以确定小鼠下丘(IC)中神经元的选择性是否遵循音频定位的空间结构。下丘中的音频定位预测,位于背侧区域的神经元应仅对低频发声做出反应,而只有位于腹侧区域的神经元应对高频发声做出反应。我们发现,对发声的反应与位置无关,并且下丘背侧低频区域的许多神经元对高频发声有反应。为了测试这是否是由于背侧神经元具有宽泛的频率调谐曲线,我们将每个神经元的频率调谐曲线与每种发声进行卷积,发现调谐曲线并不能很好地预测对发声的实际神经反应。然后,我们使用了一种在耳蜗中产生失真产物的信号转导非线性模型来预测对发声的神经反应。我们发现这些预测与实际神经反应更紧密匹配。我们的研究结果表明,耳蜗会使发声中的频率表征失真,并且一些神经元利用这种失真的表征来编码发声。