Memory for melody : investigating the link between experience, perception, and memory formation

Abstract

Music exhibits remarkable properties in the context of memory. For example, memory for melodies is long lasting, persistent, and spared from some forms of dementia and severe brain injuries. Over the course of 10 experiments, we here attempt to shed light on the question of what makes music 'special' by specifically investigating interference in memory for melody. Most domains show a decrease in recognition performance between the first and second presentation of an object as the number of intervening items increases. We tested this cumulative disruptive interference effect in the context of memory for melodies, with results showing that memory for melodies is not much affected by the number of intervening items. Specifically, the probability of correctly recognising a melody was statistically identical and above chance between 1 and up to nearly 200 intervening melodies. To explain these findings, we provide a new Regenerative-Multiple-Representations (RMR) conjecture. The conjecture describes a crucial link between prior experience, perception, and subsequent formation of memories. Using the theoretical framework of this conjecture, we further explored memory for melodies in a series of experiments. In the process, we revealed how to disrupt and shape memory for melodies' resilience to cumulative interference using melodies in an unfamiliar tuning system and with pitch-only and rhythm-only sequences. In a final analysis, we used memory as a window into perception in more general terms, and analysed data from all experiments combined to measure the degree of similarity in listeners' perception of music. The findings of this dissertation contribute to our understanding of fundamental memory phenomena, while providing practical implementations and elucidating further the mechanisms that explain why music may indeed be 'special'.

Date of Award	2017
Original language	English