Demonstrating InteractiSense: An Interactive Tool for Supporting Social Engagement for Children with ASD
Won Kim, Chaeyeon Lim, Minwoo Seong, Kangbeen Ko, SeungJun Kim · 2026 · Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems (CHI EA '26) · doi:10.1145/3772363.3799141
Summary
InteractiSense is a sensing-integrated tangible prototype paired with a cooperative serious game called Treasure Hunters, designed to support social engagement for children with autism spectrum disorder (ASD). The authors argue that ASD-related differences in cognitive, motor, and social development frequently lead to disengagement during interactive activities, and that existing engagement-supporting toys and wearables either offer rich interaction without sensing or rich sensing without ASD-appropriate form factors. InteractiSense addresses both: a soft, yellow, felt-covered sphere (18 cm diameter, sized to a 10-year-old's palm) embeds seven sensors - audio, heart rate, button input, skin temperature, accelerometer/gyroscope, galvanic skin response, and palm pressure - and streams multimodal data over Bluetooth Low Energy at up to 2.1 Mbps for roughly 10 hours on battery. The spherical form was selected over torus and cube alternatives to avoid sensor-placement constraints and edge-related safety concerns, and to align with transitional-object literature on soft, graspable, non-vulnerable materials preferred by many children with ASD. The paired Treasure Hunters game, built in Unity3D, structures play around three social skills commonly addressed in ASD therapy: imitation (players must embrace and jointly hold the sphere before a fall zone), joint attention (synchronised input on shared monsters and treasures in a timing slider), and turn-taking (alternating inputs gated by an in-game waiting state). Sessions consist of 21 modules in a balanced Latin square with a monotonous 'Game Refresher' running-on-rails scene between modules to provide a sensory pause.
Key findings
A preliminary study with 15 children with ASD (mean age 11, SD 3; 11 male, 4 female; 6 with comorbid intellectual disability, 2 with comorbid ADHD) examined data quality and user experience rather than clinical outcomes. The physiological signals were stable: signal-to-noise ratio (calculated via autocorrelation with a second-order polynomial fit) averaged above 15 dB across GSR (mean 15.6 dB), HR (15.7 dB), and skin temperature (26.8 dB), supporting the feasibility of capturing engagement-related signals during natural play. Sixty-seven percent of participants preferred InteractiSense to a general-purpose controller from prior work; with the small sample, a frequentist test was not significant, but Bayesian analysis returned a posterior probability of 0.89 favouring InteractiSense. System Usability Scale scores were comparable across the two conditions (65.8 vs 67.1) with no significant difference. The authors are explicit that this is a feasibility-and-preference signal, not a measure of social or therapeutic effect, and frame the CHI demo as a way to gather further preference data through paired gameplay and a real-time sensor visualisation interface.
Relevance
For practitioners working at the intersection of accessibility, developmental disability, and game design, this paper is a useful exemplar of how multimodal physiological sensing can be embedded in a child-friendly tangible form rather than imposed as wearables or fixed sensors. The deliberate reasoning about form factor - sphere over cube to avoid edges, soft felt for tactile preference, palm-sized for the target age - shows accessibility-first hardware design and is transferable beyond ASD to other contexts where sensory preferences shape acceptance. The translation of imitation, joint attention, and turn-taking into discrete, structurally simple gameplay modules with sensory-pause buffers between them is a concrete pattern for designing inclusive serious games. Limitations worth noting: the sample is small, validation is preference-and-signal-quality rather than clinical, and the paper does not report on accessibility for children with co-occurring blindness, motor impairment, or limited grip strength. The use of physiological data also raises privacy and consent questions for child participants that future work should address explicitly.
Tags: autism spectrum disorder · children with ASD · social engagement · serious games · tangible interaction · multimodal sensing · collaborative play · joint attention · turn-taking · imitation · physiological sensing