Signals of Success and Struggle: Early Prediction and Physiological Signatures of Human Performance across Task Complexity

Key findings from this study

This research indicates that:

• Ocular and cardiac signals collected during early task exposure prospectively predict later user performance with balanced accuracy of 0.86.
• Gaze targeting and visual sampling adjustment characterize high performers, with these patterns intensifying as task demands increase.
• High performers maintain more stable cardiac activation under increasing complexity and report more positive affective responses than low performers.

Overview

Early ocular and cardiac physiological signals predict user performance in complex interactive tasks. A within-subject experiment examined whether measures from eye tracking and heart rate monitoring collected during initial task exposure could prospectively identify high versus low performers as task complexity increased. The study tested ocular-cardiac fusion and ocular-only predictive models in a game environment with naturally escalating demands.

Methods and approach

Participants completed a within-subject experiment in a game with progressively increasing task complexity. The researchers collected ocular signals via eye tracking and cardiac signals via heart rate measurement during early task phases. They trained fusion models combining both signal types and ocular-only models to predict later performance outcomes. Physiological measures included gaze patterns, visual sampling characteristics, and cardiac activation stability. Participants also completed self-report measures of affective experience.

Results

The ocular-cardiac fusion model achieved balanced accuracy of 0.86 for cross-session performance prediction. The ocular-only model demonstrated comparable predictive power, indicating that gaze patterns alone captured most discriminative information. High performers exhibited more targeted gaze patterns and adjusted visual sampling strategies as task demands intensified. High performers maintained more stable cardiac activation throughout increasing complexity levels. High performers reported more positive affective experiences compared to low performers. These physiological and self-reported differences emerged systematically across task progression.

Implications

Early physiological assessment enables prospective identification of users at risk of performance difficulties before explicit task failure occurs. This approach supports the development of adaptive or proactive interventions timed to task demands. Eye tracking alone provides practical efficiency for performance prediction systems, reducing sensor requirements compared to multimodal fusion. The interpretable physiological signatures—targeted gaze, visual sampling adjustment, cardiac stability—establish measurable mechanisms linking physiology to performance variation. These findings establish feasibility for cross-session prediction frameworks in interactive system design.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.