Subvocalization during Preparatory and Non-preparatory Auditory Imagery

ABSTRACT

Given previous results showing that auditory imagery is associated with subvocal muscle movements related to pitch control, the present study addressed whether subvocalization of pitch is differentially involved during imagery that precedes the execution of an imagined action as compared to non-preparatory imagery. We examined subvocal activity using surface electromyography (sEMG) during auditory imagery that preceded sung reproduction of a pitch sequence (preparatory) or recognition of a pitch sequence (non-preparatory). On different trials, participants either imagined the sequence as presented, or imagined a mental transformation of that sequence. Behavioral results replicated previous findings of poorer reproduction and recognition of transformed sequences compared to sequences in their original form. Physiological results indicated that subvocal activity was significantly above baseline for all conditions, greater than activity observed for the bicep control site, and greater for longer sequences, but did not reliably scale with transformation type. Furthermore, greater subvocal activity during preparatory imagery was associated with greater subvocal activity during non-preparatory imagery for muscles involved in pitch control and articulation. Muscle activity involved in pitch control was similarly recruited for both preparatory and non-preparatory auditory imagery. In contrast, muscle activity involved in vocal articulation was most strongly recruited during motor preparation. Our findings suggest that pitch imagery recruits subvocal muscle activity regardless of whether the imagined action is intended to be effected.

KEYWORDS:

• Subvocalization
• singing
• auditory imagery
• short-term memory

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Ethics Statement

The research presented was approved by the University at Buffalo, SUNY Institution Review Board.

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/25742442.2022.2163582.

Notes

1. This sample size exceeds the size needed to achieve 80% power for detecting the transformation effect in our behavioral data based on a power analysis performed using G*power 3.1 (Faul et al., 2007) using the effect size for the transformation effect reported in Greenspon et al. (2017).

2. Six participants were excluded from the five-sensor ANOVA, three participants were excluded from analyses using the right sternohyoid sensor, two participants were excluded from analyses using the upper left lip sensor, and all remaining within-sensor analyses excluded one participant.

3. This step resulted in the exclusion of between 2.5% and 6.5% of trials in each task (i.e., production and recognition) for each of the five sensors (Bicep, Left and Right Sternohyoid, Left and Right Lip) due to extreme sEMG values (>3 SD from the mean).

4. A 2 (three- and four-note) × 4 (no transformation, transposition, serial order shift, and reversal) ANOVA on relative pitch accuracy replicated the transformation effect found for the pitch accuracy measure. No other effects were significant for analyses on relative pitch accuracy, including the main effect of length.

Additional information

Funding

This work was supported by the NSF BCS1848930.