Herff Steffen A, Olsen Kirk N, Dean Roger T, Prince Jon
1 The MARCS Institute for Brain, Behaviour and Development, Western Sydney University, Penrith, NSW, Australia.
2 Department of Psychology, Macquarie University, North Ryde, NSW, Australia.
Q J Exp Psychol (Hove). 2018 Jun;71(6):1367-1381. doi: 10.1177/1747021817734978. Epub 2018 Jan 1.
In a continuous recognition paradigm, most stimuli elicit superior recognition performance when the item to be recognized is the most recent stimulus (a recency-in-memory effect). Furthermore, increasing the number of intervening items cumulatively disrupts memory in most domains. Memory for melodies composed in familiar tuning systems also shows superior recognition for the most recent melody, but no disruptive effects from the number of intervening melodies. A possible explanation has been offered in a novel regenerative multiple representations (RMR) conjecture. The RMR assumes that prior knowledge informs perception and perception influences memory representations. It postulates that melodies are perceived, thus also represented, simultaneously as integrated entities and also as their components (such as pitches, pitch intervals, short phrases and rhythm). Multiple representations of the melody components and melody as a whole can restore one another, thus providing resilience against disruptive effects from intervening items. The conjecture predicts that melodies in an unfamiliar tuning system are not perceived as integrated melodies and should (a) disrupt recency-in-memory advantages and (b) facilitate disruptive effects from the number of intervening items. We test these two predictions in three experiments. Experiments 1 and 2 show that no recency-in-memory effects emerge for melodies in an unfamiliar tuning system. In Experiment 3, disruptive effects occurred as the number of intervening items and unfamiliarity of the stimuli increased. Overall, results are coherent with the predictions of the RMR conjecture. Further investigation of the conjecture's predictions may lead to greater understanding of the fundamental relationships between memory, perception and behavior.
在连续识别范式中，当待识别项目为最近呈现的刺激时，大多数刺激会引发更优的识别表现（记忆近因效应）。此外，在大多数领域中，增加中间项目的数量会累积性地干扰记忆。对用熟悉调律系统创作的旋律的记忆，同样对最近的旋律表现出更优的识别，但不受中间旋律数量的干扰影响。一种新颖的再生多重表征（RMR）猜想提供了一种可能的解释。RMR假设先验知识影响感知，而感知又影响记忆表征。它假定旋律在被感知时，因此也在被表征时，同时作为一个整体实体以及其组成部分（如音高、音程、短乐句和节奏）被呈现。旋律组成部分和旋律整体的多重表征可以相互恢复，从而提供抵御中间项目干扰影响的能力。该猜想预测，处于不熟悉调律系统中的旋律不会被感知为完整的旋律，并且应该（a）破坏记忆近因优势，以及（b）加剧中间项目数量带来的干扰影响。我们在三个实验中检验了这两个预测。实验1和实验2表明，对于处于不熟悉调律系统中的旋律，不存在记忆近因效应。在实验3中，随着中间项目数量和刺激的不熟悉程度增加，出现了干扰影响。总体而言，结果与RMR猜想的预测一致。对该猜想预测的进一步研究可能会加深我们对记忆、感知和行为之间基本关系的理解。
