Pupillary activity in areas of interest from visual stimuli for neonatal pain assessment

This paper compares the pupillary activity index to traditional eye-tracking metrics like the fixation count and duration in assessing neonatal pain. It explores the benefits of incorporating pupillary activity measures to improve methods that lead to an understanding of cognitive processing and performance evaluation. The estimation of cognitive load using pupil diameter typically involves measures relative to a baseline. Instead, we conducted an eye-tracking study using the Low/High Index of Pupillary Activity to evaluate healthcare experts and non-experts analyzing the faces with and without pain from a dataset of newborn faces. This data was recorded by the Tobii TX300 eye-tracking system in a closed room with controlled lighting. Our contribution is to introduce the LHIPA calculation considering the areas of interest segments of the pupil diameter signal. The results suggest that the visual attention reflected by the traditional metrics may not correspond directly to the respective cognitive load for both sample groups of participants.

Revisiting N-CNN for Clinical Practice

This paper revisits the Neonatal Convolutional Neural Network (N-CNN) by optimizing its hyperparameters and evaluating how they affect its classification metrics, explainability and reliability, discussing their potential impact in clinical practice. We have chosen hyperparameters that do not modify the original N-CNN architecture, but mainly modify its learning rate and training regularization. The optimization was done by evaluating the improvement in F1 Score for each hyperparameter individually, and the best hyperparameters were chosen to create a Tuned N-CNN. We also applied soft labels derived from the Neonatal Facial Coding System, proposing a novel approach for training facial expression classification models for neonatal pain assessment. Interestingly, while the Tuned N-CNN results point towards improvements in classification metrics and explainability, these improvements did not directly translate to calibration performance. We believe that such insights might have the potential to contribute to the development of more reliable pain evaluation tools for newborns, aiding healthcare professionals in delivering appropriate interventions and improving patient outcomes.