普氏蹄蝠下丘谐波内外神经元处理多普勒频移补偿信息的差异

Differences Between Harmonic and Non-harmonic Neurons Processing Doppler-shift Compensation Information in Inferior Colliculus of Hipposideros pratti

为了探讨普氏蹄蝠下丘神经元在处理多普勒频移补偿后回声定位信号中的作用,实验采用双声刺激模式模拟蝙蝠不同飞行状态下产生多普勒频移补偿后的脉冲-回声对,即发声频率改变,回声频率维持恒定的情况下,研究下丘神经元对不同补偿值下的回声反应恢复率.结果发现:根据神经元在某一补偿值下对回声信号反应的恢复率是否超过70%,可将其分为具有选择性(S)和无选择性(NS)的两类神经元.且谐波内S神经元所占比例(68%)远超过非谐波内S神经元(39%).分析神经元的发放模式发现谐波内S神经元中相位型发放模式比例(44.3%)明显高于其他三种类型神经元.另外,虽然S和NS神经元的强度-潜伏期函数类型均以饱和型为主,但谐波内S神经元强度-潜伏期函数的最佳强度(best amplitude,BA)(95.3±14.0)dB SPL低于NS神经元的BA(104.1±10.2)d B SPL(P<0.01),同时也低于非谐波内S神经元的BA(109.7±7.9)dB SPL(P<0.01).以上实验结果表明,在下丘水平,神经元就已对多普勒频移补偿后回声定位信号的处理有了分工,集中在谐波内的S神经元通过提高对某一补偿值下回声信号反应的恢复率实现,对回声信息的精确编码,避免其他杂波干扰信息.同时,谐波内S神经元的发放模式和强度-潜伏期函数特点也满足其在复杂环境中精确声学成像的需求.

In order to investigate the role of inferior colliculus （IC） neurons in Pratt＇s roundleaf bats＇ Doppler-shift compensation （DSC）, we recorded the responses of IC neurons to the echoes by using simulated pulse-echo pairs of bats＇ DSC. According to the recovery rate of neural responses to echo signals at different frequency shifts of DSC, neurons were classified into two types： selective neurons （S neurons） whose recovery rate was more than 70% at some specific frequency shifts of DSC and non-selective neurons （NS neurons） whose recovery rate was less than 70% at any frequency shifts of DSC. Further analysis revealed that the proportion of S neurons of harmonic neurons （68%） was greater than the non-harmonic neurons （39%）, and firing pattems of most harmonic S neurons was phasic （44.3%）. Although the intensity-latency curves were similar in S and NS neurons, the harmonic neuron＇s best amplitude （BA） of the functions in S neurons （95.3 ±14.0 dB SPL） were notably lower than NS neurons （104.1 ± 10.2 dB SPL） （P 〈 0.01）, and there was a very significant difference in BA between the S neurons of harmonic and non-harmonic neurons （95.3 ± 14.0 vs 109.7 ± 7.9 dB SPL, P 〈 0.01）. These results suggested that there is a clear division of labour in IC neurons processing and analyzing echo information during the DSC. S neurons which concentrated in the harmonics can effectively encode the echo information and avoid disturbances of other clutter information by increasing the recovery rate of responses to echo under specific frequency shifts of DSC. And the characteristics of firing patterns and intensity-latency functions of S neurons of harmonic neurons were different than other neurons, which met the need in precise acoustic imaging at complex environments.