One digit too many: brain rhythms and autonomic responses at the limits of working memory capacity
- 1University of Alabama at Birmingham
Abstract
Physiological measures vary with working memory load, but whether they identify when demands exceed a person's working memory capacity remains unclear. We assessed neural and autonomic responses to exceeding digit span by one digit. During an adaptive digit span task, we recorded EEG, pupillometry, and photoplethysmography from 123 participants. We compared each participant's digit span (approximately 70% correct full-sequence recall) with sequences one digit longer (span+1), operationally defined as overload. Full-sequence accuracy decreased from 74.1% to 51.5%, and subjective mental demand increased. Baseline-normalized pulse wave amplitude (PWA) was lower in span+1 from the first digit through terminal retention period. PWA was lower on correctly recalled than erroneous span+1 trials. Phasic pupil responses increased across serial positions but were smaller in span+1 at late span-matched positions; responses did not differ between conditions at their final digits. Non-baseline-corrected pupil diameter was larger in span+1 before sequence onset and during terminal retention, whereas non-baseline-corrected PWA did not differ. Frontal midline theta increased with serial position and was greater during terminal retention in span+1. Alpha and beta power decreased with serial position, and differed between conditions only at late positions. Terminal beta suppression was greater on correctly recalled span+1 trials. Thus, exceeding calibrated span impaired recall without producing a common physiological breakpoint. The findings suggest that reaching this limit does not immediately lead to resource exhaustion or disengagement. Physiological responses showed shifts in timing consistent with adjustments in preparation for recall of the longer sequence, along with increased effort before and during longer sequences.
Related
- Modulation of fronto-parietal and autonomic activity underpins preserved working memory in middle-aged adults
- Spontaneous Perceptual Reversals reflect in Internal Rhythms, Not Metabolic Shifts or Evoked Responses
- Eye movements reflect capacity-dependent organization of working memory
- Air temperature limits aerobic capacity and scope in active birds
- Distinct electroencephalography alpha-band rhythms map macroscale brain gradients
BibTeX
@misc{kotyusov2026digit,
title = {{One digit too many: brain rhythms and autonomic responses at the limits of working memory capacity}},
author = {A. I. Kotyusov and D. A. Kasanov and A. I. Zaitseva and B. Kotchoubey and Y. G. Pavlov},
howpublished = {bioRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.01.763481},
url = {https://www.biorxiv.org/content/10.64898/2026.11.01.763481v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - One digit too many: brain rhythms and autonomic responses at the limits of working memory capacity
AU - A. I. Kotyusov
AU - D. A. Kasanov
AU - A. I. Zaitseva
AU - B. Kotchoubey
AU - Y. G. Pavlov
PY - 2026
DA - 2026/11/01
PB - bioRxiv
DO - 10.64898/2026.11.01.763481
UR - https://www.biorxiv.org/content/10.64898/2026.11.01.763481v1
AB - Physiological measures vary with working memory load, but whether they identify when demands exceed a person's working memory capacity remains unclear. We assessed neural and autonomic responses to exceeding digit span by one digit. During an adaptive digit span task, we recorded EEG, pupillometry, and photoplethysmography from 123 participants. We compared each participant's digit span (approximately 70% correct full-sequence recall) with sequences one digit longer (span+1), operationally defined as overload. Full-sequence accuracy decreased from 74.1% to 51.5%, and subjective mental demand increased. Baseline-normalized pulse wave amplitude (PWA) was lower in span+1 from the first digit through terminal retention period. PWA was lower on correctly recalled than erroneous span+1 trials. Phasic pupil responses increased across serial positions but were smaller in span+1 at late span-matched positions; responses did not differ between conditions at their final digits. Non-baseline-corrected pupil diameter was larger in span+1 before sequence onset and during terminal retention, whereas non-baseline-corrected PWA did not differ. Frontal midline theta increased with serial position and was greater during terminal retention in span+1. Alpha and beta power decreased with serial position, and differed between conditions only at late positions. Terminal beta suppression was greater on correctly recalled span+1 trials. Thus, exceeding calibrated span impaired recall without producing a common physiological breakpoint. The findings suggest that reaching this limit does not immediately lead to resource exhaustion or disengagement. Physiological responses showed shifts in timing consistent with adjustments in preparation for recall of the longer sequence, along with increased effort before and during longer sequences.
ER -