Quantifying Emotional Arousal through Pupillary Response: A Novel Approach for Isolating the Luminosity Effect and Predicting Affective States

TL;DR

The study tackles the confounding influence of dynamic screen luminosity on pupil-based arousal inference during naturalistic video viewing. It introduces a three-stage pipeline: a personalized, non-linear LEPM to estimate and remove luminosity-driven pupil changes, an ADM to relate the luminosity-corrected residual to self-reported arousal, and a GBR-SAPM to further improve prediction using gradient-boosted features. Results show that luminosity correction dramatically improves arousal prediction (e.g., from $r=0.26$ to $r=0.65$ in simple setups, and $r=0.765$ with GBR-SAPM), while independent judges yield worse ground-truth alignment than self-reports. The work demonstrates a scalable, practical method for isolating psychosensory pupil responses in standard lab conditions, enabling robust continuous arousal assessment from pupillometry for psychology, health, and consumer research. It also highlights that arousal–light interactions are non-linear and multiplicative, justifying the move beyond additive correction toward data-driven, subject-specific calibration.

Abstract

Pupil dilation is recognized as an objective indicator of emotional arousal, but confounding factors such as the luminosity of stimuli and the surrounding environment have greatly limited its practical usefulness. This study presents a new approach to isolate and remove the effect of luminosity on pupil dilation. We validated this approach by showing 32 video clips with different content and emotional intensity to 47 participants, who reported their level of emotional arousal after each video. We developed a model capable of predicting the effect of luminosity on pupil size as a function of screen brightness, which adapts to individual physiological differences and different types of monitors through a brief pre-experimental calibration. We thus estimated the pupil size due exclusively to luminosity and subtracted it from the total recorded pupil size, obtaining the component due exclusively to arousal. From the latter, we predicted the arousal of each participant for each video using two models. We first used a simple linear regression model. When we used the luminosity-corrected pupil size, we obtained a correlation between predicted and self-reported arousal of r = 0.65 +/- 0.12, and R2 of 0.43 +/- 0.12. The uncorrected pupil size, instead, showed virtually no predictive power (r = 0.26 +/- 0.15, R2 = 0.09 +/- 0.089). We then used an Extreme Gradient Boosting model, obtaining even better results in the case of luminosity correction (r = 0.765 +/- 0.047, R2 = 0.556 +/- 0.085). Our results highlight that separating emotional and luminosity components from pupillary responses is crucial for accurately predicting arousal.

Figures (13)

• Figure 1: Pupil size as a function of luminosity for red, green, blue, and gray (scattered line for experimental data and dotted line for the fitted curve).
• Figure 2: Testing the combined approach.
• Figure 3: Experiment flow of LEPM validation.
• Figure 4: Experimental setup.
• Figure 5: Aggregate ground truth responses across all participants, categorized by stimuli arousal and valence: H = High Arousal, L = Low Arousal, P = Positive Valence, N = Negative Valence. Each point on the graph corresponds to a stimulus, i.e., a video clip. The valence and arousal values are rescaled in the range [-2, 2], where 0 indicates a neutral, average value.