Skin Conductance as an In Situ Marker for Emotional Arousal in Children with Neurodevelopmental Communication Impairments: Methodological Considerations and Clinical Implications
Mariana Aparicio Betancourt, Laura S. Dethorne, Karrie Karahalios, Jennifer G. Kim · 2017 · ACM Transactions on Accessible Computing · doi:10.1145/3035536
Summary
This paper explores the use of wearable biosensors to measure electrodermal activity (EDA) — specifically skin conductance — as a proxy for emotional arousal in children with neurodevelopmental communication impairments. While EDA has been used in psychological research for over 130 years, applying this technology in situ (within everyday activities rather than controlled laboratories) remains challenging, particularly with children who have sensory sensitivities, difficulty expressing emotions, and higher rates of challenging behaviors. The researchers draw on experience from two experimental interventions with eight children ages 2-11: the SL-EDA study monitored skin conductance during speech-language therapy for children at the single-word developmental stage, while the OT-EDA study examined the effects of pressure vests on academic engagement and behavior in children with intellectual disabilities during occupational therapy. Both studies used Affectiva Q sensors placed primarily on ankles to collect continuous EDA data synchronized with video recordings. The paper provides an accessible overview of skin conductance science, explaining both phasic responses (transient peaks indicating reactions to specific stimuli) and tonic levels (baseline conductivity associated with overall arousal). It examines the neural mechanisms underlying EDA and discusses its potential for understanding the experiences of children who cannot easily communicate their internal states through language.
Key findings
The studies revealed significant challenges in using wearable EDA sensors with this population, with 29% of data being discarded due to behavioral and technical issues. Three behavioral themes emerged: (1) developmental differences, with younger children (ages 2-4) requiring multiple desensitization sessions before tolerating sensors, while older children (ages 9-11) wore them immediately without difficulty; (2) sensor tolerance varied substantially — most children eventually wore sensors for 61% of sessions, but one child (Heidi) never tolerated them across 25 sessions; (3) ankle placement was better tolerated than wrist placement, with children more likely to forget about hidden sensors and less likely to manipulate them. Technical challenges included sensor functioning issues (sensors required electrode gel despite manufacturer claims, Bluetooth connectivity was unreliable), synchronization difficulties (matching EDA timestamps with video was described as "more art than science"), and data preprocessing complexity (existing software like ELAN proved unreliable for this purpose, leading to development of BEDA). Despite these challenges, the SL-EDA study found significant associations between skin conductance response frequency and behavioral coding of emotional valence, with both measures increasing over treatment sessions — consistent with children becoming more engaged over time.
Relevance
This paper provides essential practical guidance for researchers and practitioners considering physiological monitoring with children who have communication disabilities. The detailed methodological considerations — including specific desensitization protocols, sensor placement recommendations, data quality thresholds, and synchronization strategies — fill a significant gap in the literature. The behavioral recommendations are immediately actionable: use social stories to prepare children, place sensors on ankles under socks, allow four or more desensitization sessions for children under five, pair sensor introduction with positive experiences, and have backup sensors available. The technical guidance specifies parameters like recommended SCR thresholds (>0.05 μS amplitude), temperature-controlled recording environments (22-24°C), and the need for 10-15 minute acclimation periods. For accessibility practitioners, this research highlights both the promise and limitations of using physiological sensing to understand the experiences of individuals who cannot easily self-report. While wearable biosensors are increasingly marketed as tools to "predict meltdowns" in autistic children, this paper grounds such claims in the practical realities of signal acquisition with this population.
Tags: electrodermal activity · wearable sensors · emotional arousal · autism · neurodevelopmental disorders · communication impairments · physiological computing · biosensors
Standards referenced: DSM-5