AUTHOR=Vásquez-Pinto Sebastián , Morales-Bader Diego , Cox Ralf F. A. , Munoz-Rubke Felipe , Castillo Ramón D. TITLE=The nonlinearity of pupil diameter fluctuations in an insight task as criteria for detecting children who solve the problem from those who do not JOURNAL=Frontiers in Psychology VOLUME=Volume 14 - 2023 YEAR=2023 URL=https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2023.1129355 DOI=10.3389/fpsyg.2023.1129355 ISSN=1664-1078 ABSTRACT=This study investigated whether certain features associated with self-organization in dynamical systems can detect differences between individuals who succeed and those who do not succeed in solving an insight task. Specifically, the study examined fluctuations in entropy and fractal scaling exponents. The 8-coin task was used as it is considered a better insight task than other tasks. Pupillary diameter fluctuations of children aged between 6 and 12 years (M=9.39; SD=1.53) were analyzed, and the children were divided into two groups based on who successfully (n=24) and unsuccessfully (n=43) solved the insight problem task. Recurrence Quantification and Power Spectrum Density analyses were used to estimate Entropy, Determinism, Recurrence Ratio, and the β scaling exponent. The results showed that the solver group had greater uncertainty (between 102 - 86 and 58 – 30 seconds) and less predictability (52 and 44 seconds) in their pupillary diameter fluctuations before finding the solution. The RQA measures could find changes that the mean and standard deviation did not detect. However, the β scaling exponent could not differentiate between the two groups at any moment. The findings suggest that entropy and determinism in pupillary diameter fluctuations can detect differences in problem-solving success early on in problem-solving processes. However, it is speculative to conclude that problem-solving processes depend only on perceptual and motor activity and do not require high-level representational mechanisms. Further research is necessary to support such claims reliably. Moreover, it is important to investigate whether these findings generalize to other tasks and populations.