Age-related changes in pupil dynamics and task modulation across the healthy lifespan

Abstract

The pupil is modulated by luminance, arousal, bottom-up sensory, and top-down cognitive signals, and has increasingly been used to assess these aspects of brain functioning in health and disease. However, changes in pupil dynamics across the lifespan have not been extensively examined, hindering our ability to fully utilize the pupil in probing these underlying neural processes in development and aging in healthy and clinical cohorts. Here, we examined pupil responses during the interleaved pro-/anti-saccade task (IPAST) in healthy participants across the lifespan (n = 567, 5-93 years of age). Based on the extracted measurements of pupil dynamics, we demonstrated age-related changes in pupil measures and task modulation. Moreover, we characterized the underlying factors and age-related effects in components of pupil responses that may be attributed to developmental and aging changes in the associated brain regions. Finally, correlations between factors of pupil dynamics and saccade behaviors revealed evidence of shared neural processes in the pupil and saccade control circuitries. Together, these results demonstrate changes in pupil dynamics as a result of development and aging, providing a baseline with which altered pupil responses due to neurological deficits at different ages can be studied.

Keywords: aging; anti-saccade; development; pupillary response; saccade preparation.

Figure 1

IPAST experiment paradigm and measurements of the pupil response. (A) Experiment paradigm for the interleaved pro- and anti-saccade task (IPAST). Each trial began with the appearance of a colored central fixation cue on a dark background. The color of FP provided the task instruction for the trial (Blue: PRO; pro-saccade. Red: ANTI; anti-saccade). After fixation, the FP was removed from the screen briefly, after which a white peripheral stimulus appeared horizontally to the left or right to the FP. Participants were instructed to either make a saccade to the stimulus location (PRO) or towards the opposite direction (ANTI). Note that the displayed arrows here indicating correct eye movement directions are for illustrative purposes only. (B) Measurements of the pupil response were calculated to capture the pupil dynamics, including baseline pupil size, pupil response onset latency, constriction size, peak constriction time, peak constriction velocity, dilation size, dilation velocity, and peak dilation velocity.

Figure 2

Mean pupil response across the lifespan. Mean pupil response before (A,B) and after baseline correction (C,D) for PRO and ANTI conditions during IPAST fixation period across the lifespan.

Figure 3

Pupil measurements across the lifespan. Measurements of pupil dynamics across the lifespan for PRO (blue) and ANTI (red) conditions: (A) baseline pupil size, (B) pupil response onset latency, (C) constriction size, (D) peak constriction time, (E) peak constriction velocity, (F) dilation size at target onset, (G) dilation velocity at target onset, and (H) peak dilation velocity. Colored dots represent individual subject data points, and colored curves represent smoothing spline fits of each task condition for all participants.

Figure 4

Task modulation (Anti-effect = ANTI – PRO) of the healthy lifespan for dilation size (A) and dilation velocity (B). Grey dots represent Anti-effect calculated for individual subject, and black curves represent smoothing spline fits plotted for all participants.

Figure 5

Factors of the measurements of pupil dynamics. (A) Correlation matrix and factor structure of the measurements of pupil dynamics. The color represents Spearman’s correlation coefficients between pairs of pupil measures; grey squares indicate correlation coefficients of 1. Dashed square outlines represent groups of pupil measures that loaded together in the factor analysis as displayed on the right. (B–D) Pupil factor scores for the three pupil factors across the lifespan. Grey dots represent individual subject factor scores, black curves represent GAM fits for all participants, gray ribbons are the 95% confidence intervals of the GAM fits, and the bottom tiles indicate periods of significant change.

Figure 6

Correlation matrix of factor scores of pupil dynamics and saccade behaviors. The colored squares represent Spearman’s correlation coefficients between significantly correlated pairs of factor scores after Bonferroni correction.