Department of Psychology, Lancaster University, UK.
J Assoc Res Otolaryngol. 2011 Aug;12(4):503-17. doi: 10.1007/s10162-011-0266-3. Epub 2011 Apr 12.
Multiple-hour training on a pitch discrimination task dramatically decreases the threshold for detecting a pitch difference between two harmonic complexes. Here, we investigated the specificity of this perceptual learning with respect to the pitch and the resolvability of the trained harmonic complex, as well as its cortical electrophysiological correlates. We trained 24 participants for 12 h on a pitch discrimination task using one of four different harmonic complexes. The complexes differed in pitch and/or spectral resolvability of their components by the cochlea, but were filtered into the same spectral region. Cortical-evoked potentials and a behavioral measure of pitch discrimination were assessed before and after training for all the four complexes. The change in these measures was compared to that of two control groups: one trained on a level discrimination task and one without any training. The behavioral results showed that learning was partly specific to both pitch and resolvability. Training with a resolved-harmonic complex improved pitch discrimination for resolved complexes more than training with an unresolved complex. However, we did not find evidence that training with an unresolved complex leads to specific learning for unresolved complexes. Training affected the P2 component of the cortical-evoked potentials, as well as a later component (250-400 ms). No significant changes were found on the mismatch negativity (MMN) component, although a separate experiment showed that this measure was sensitive to pitch changes equivalent to the pitch discriminability changes induced by training. This result suggests that pitch discrimination training affects processes not measured by the MMN, for example, processes higher in level or parallel to those involved in MMN generation.
在音高辨别任务上进行数小时的训练可以显著降低检测两个谐波复合体之间音高差别的阈值。在这里，我们研究了这种感知学习的特异性，包括音高和训练的谐波复合体的可分辨性，以及其皮质电生理相关性。我们使用四个不同的谐波复合体中的一个，对 24 名参与者进行了 12 小时的音高辨别任务训练。这些复合体在音高和/或其组成部分的频谱可分辨性上有所不同，通过耳蜗，但被过滤到相同的频谱区域。在训练前后，对所有四个复合体进行了皮质诱发电位和音高辨别行为测量。将这些测量的变化与两个对照组进行比较：一个对照组进行了水平辨别任务训练，另一个对照组没有进行任何训练。行为结果表明，学习在一定程度上是特定于音高和可分辨性的。用可分辨的谐波复合体进行训练可以提高对可分辨复合体的音高辨别能力，而用不可分辨的复合体进行训练则不然。然而，我们没有发现证据表明用不可分辨的复合体进行训练会导致对不可分辨复合体的特定学习。训练影响了皮质诱发电位的 P2 成分和稍后的成分（250-400ms）。在失匹配负波（MMN）成分上没有发现显著变化，尽管一项单独的实验表明，该测量方法对与训练引起的音高差辨别变化相当的音高变化敏感。这一结果表明，音高辨别训练影响了 MMN 未测量到的过程，例如，在水平上更高或与 MMN 产生过程平行的过程。
