|
Abstract
|
This study examines how computational complexity and working memory load affect pupil responses during decision-making. Ninety participants completed a knapsack decision task, an NP-hard problem, across 72 trials. Before each trial, participants memorised either a 1-digit or 6-digit number, which they recalled after the decision task. Each trial began with a fixation period, followed by up to 25 seconds for participants to decide if a combination of six items met a minimum value and maximum weight constraint. The knapsack task varied in complexity, with half of the instances having a solution that met both constraints (“satisfiable”) and half lacking a feasible solution (“unsatisfiable”). Eye-tracking data, including gaze and changes in pupil diameter (CPD), were recorded at 300 Hz using Tobii Pro Spectrum eye trackers. Behavioural results showed that increased complexity reduced performance in satisfiable instances, while performance in unsatisfiable instances remained high. Working memory load further impaired performance, especially in low-complexity satisfiable instances, while time spent on the task remained unaffected by memory load. CPD revealed that unsatisfiable instances elicited slightly greater changes in pupil diameter under low complexity. Higher memory load significantly increased CPD, supporting the hypothesis that working memory load increases cognitive effort or arousal, as reflected in CPD. CPD during decision-making was influenced by an interaction between complexity, satisfiability, and memory load, highlighting the interplay of these factors. These findings provide insight into how working memory load and task complexity interact to influence biophysiological markers of decision-making effort.
|