Animal vocalizations and human speech are typically characterized by a complex spectrotemporal structure, composed of multiple harmonics, and patterned as temporally organized sequences. However, auditory research often employed simple artificial acoustic stimuli or their combinations. Here we addressed the question of whether the neuronal responses to natural echolocation call sequences can be predicted by manipulated sequences of incomplete constituents at the midbrain inferior colliculus (IC). We characterized the extracellular single-unit activity of IC neurons in the great roundleaf bat, Hipposideros armiger (both sexes), using natural call sequences, various manipulated sequences of incomplete vocalizations, and pure tones. We report that approximately two-thirds of IC neurons exhibited a harmonic interaction. Neurons with high harmonic interactions exhibited greater selectivity to natural call sequences, and the degree of harmonic interaction was robust to the natural amplitude variations between call harmonics. For 81% of the IC neurons, the responses to the natural echolocation call sequence could not be predicted by altered sequences of missing call components. Surprisingly, nearly 70% of the neurons that showed a harmonic interaction were characterized by a single excitatory response peak as revealed by pure tones. Our results suggest that prevalent harmonic processing has already emerged in the auditory midbrain IC in the echolocating bat.

动物发声和人类语音的典型特征是复杂的频谱时间结构，由多个谐波组成，并被模式化为时间上组织的序列。然而，听觉研究通常采用简单的人工声学刺激或其组合。在这里，我们解决了神经元对自然回声定位调用序列的反应是否可以通过中脑下丘 （IC） 处不完整成分的操纵序列来预测的问题。我们使用自然叫声序列、各种操纵的不完整发声序列和纯音调表征了大圆叶蝙蝠 Hipposideros armiger （两性） 中 IC 神经元的细胞外单单位活性。我们报告说，大约 2/3 的 IC 神经元表现出谐波相互作用。具有高谐波相互作用的神经元对自然调用序列表现出更大的选择性，并且谐波相互作用的程度对调用谐波之间的自然振幅变化具有鲁棒性。对于 81% 的 IC 神经元，对自然回声定位调用序列的反应无法通过缺失调用成分的改变序列来预测。令人惊讶的是，近 70% 显示谐波相互作用的神经元的特征是纯音所揭示的单个兴奋性反应峰。我们的结果表明，在回声定位蝙蝠的听觉中脑 IC 中已经出现了普遍的谐波处理。