Analysis of Heart Rate Changes in Response to Sensory Stimulus Intensity Variations Using a Convolutional Neural Network-Based Remote Photoplethysmography Method

Abstract

The neurodevelopmental disease known as autism spectrum disorder (ASD) is frequently linked to deficiencies in sensory processing, which cause different physiological reactions to different environmental stimuli. Heart rate is one physiological sign that can be used to assess these reactions. This study uses a non-contact, camera-based approach to examine changes in heart rate in response to changes in sensory stimulus intensity in people with autism. The suggested system combines the Remote Photoplethysmography (rPPG) technique, which extracts physiological information from variations in facial skin color intensity, with a Convolutional Neural Network (CNN) to identify the facial region and establish the Region of Interest (ROI). Without physical contact with the individual, the system can use an RGB webcam to estimate heart rate in real time, in beats per minute (BPM). According to experimental findings, people with ASD showed discernible changes in heart rate in response to changes in sensory stimulus intensity. These results suggest that the proposed approach could support sensory evaluation for people with autism by offering a more pleasant and objective alternative to non-contact physiological-response monitoring.