The Role of Long-Term Memory in Mental Transformations of Pitch

ABSTRACT

Most people can recognize and perform a musical piece under a variety of transformations such as altering the key or varying the tempo. However, we also know that other mental transformations of music can be difficult to generate and to recognize. Two factors that might affect this mental flexibility are the familiarity of the piece and musical ability of the listener, in this case singing accuracy. The current experiment addressed the accuracy and flexibility of representations of novel and traditional melodies among accurate, moderate, and inaccurate singers. Participants sang or recognized melodies in either their original form or as a transformation: a transposition, a shift of serial position, and a reversal of the melody. Participants showed an advantage for traditional melodies, but only when singing or recognizing tunes in their original form. Participants were similarly disrupted by mental transformations of traditional and novel tunes in both production and recognition tasks. Interestingly, we found that the only advantage for traditional melodies when singing repetitions of the melody occurred among the moderate singers, but all three groups showed an advantage for traditional melodies when recognizing exact repetitions.

KEYWORDS:

• Mental Transformation
• familiarity
• memory
• music
• singing
• auditory Imagery

Acknowledgments

We would like to thank the following undergraduate research assistants from the University at Buffalo, SUNY for their assistance with data collection for the current study: Amy O’Leary, Melissa Mercado, Nicholas Nolan, and Cailin Shupbach.

Supplementary Material

Supplemental data for this article can be accessed here.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. This sample was drawn from a larger original sample (N = 134). Forty-six participants from the larger sample (34%) sang the wrong number of notes on the majority of trials in the production task, and were dropped from the analyzed sample. The frequency of dropped trials did not vary by group, transformation condition, or level of familiarity, see Supplementary Results. Five more participants (<4%) were dropped due to extreme pitch discrimination thresholds or missing data resulting from experimenter error. The final sample (n = 83) was selected because it provided the least ambiguous scoring of production data.

2. One traditional melody (“Jingle Bells”) comprised eighth notes and one traditional melody (“This Old Man”) comprised of quarter notes and one-half note. The rhythm reflected in these two melodies were not considered complex (i.e. syncopated) rhythms. All other melodies in the experiment were composed of isochronous rhythms comprising quarter notes. We observed a similar pattern of results for the study when the production and recognition data was analyzed without the traditional/novel melody pairs for “Jingle Bells” and “This Old Man.”

3. As shown in Table 1, we did find group differences in assessment measures in addition to differences in singing accuracy. However, follow-up analyses in which participants were grouped based on their performance on cognitive and self-report assessment measures did not replicate the effects of the Group factor found in the current study, see Supplementary Results.

4. This participant was retained in analyses of production and recognition performance because his production and recognition data did not show evidence that the manipulation of melody familiarity affected him differently than the rest of the sample.

5. In addition to the replicated group × condition interaction in the current study, we directly addressed whether the current study replicated the results of Greenspon et al. (2017) by re-analyzing the current data using the criteria for singing accuracy defined in our earlier paper. As shown in the Supplementary Results, the present data replicated those results.

Additional information

Funding

This research was supported by NSF Grant BCS-1256964.