SMART-TBI: Design and Evaluation of the Social Media Accessibility and Rehabilitation Toolkit for Users with Traumatic Brain Injury
Yaxin Hu, Hajin Lim, Lisa Kakonge, Jade T. Mitchell, Hailey L. Johnson, Lyn Turkstra, Melissa C. Duff, Catalina L. Toma, Bilge Mutlu
TL;DR
This work targets social participation barriers faced by individuals with traumatic brain injury (TBI) on social media. It introduces SMART-TBI, a suite of five aids (Writing Aid, Interpretation Aid, Filter Mode, Focus Mode, Facebook Customization) implemented as Facebook-oriented Chrome extensions to provide communication and cognitive support. The toolkit was evaluated with eight users with moderate-to-severe TBI and five rehabilitation experts, revealing both potential benefits and design challenges, including psychological safety, privacy, and mixed reactions to AI-based aids. The study offers design implications to guide the development of accessible social media platforms that accommodate cognitive and communicative needs in real-world use.
Abstract
Traumatic brain injury (TBI) can cause a range of cognitive and communication challenges that negatively affect social participation in both face-to-face interactions and computer-mediated communication. In particular, individuals with TBI report barriers that limit access to participation on social media platforms. To improve access to and use of social media for users with TBI, we introduce the Social Media Accessibility and Rehabilitation Toolkit (\textbf{SMART-TBI}). The toolkit includes five aids (Writing Aid, Interpretation Aid, Filter Mode, Focus Mode, and Facebook Customization) designed to address the cognitive and communicative needs of individuals with TBI. We asked eight users with moderate-severe TBI and five TBI rehabilitation experts to evaluate each aid. Our findings revealed potential benefits of aids and areas for improvement, including the need for psychological safety, privacy control, and balancing business and accessibility needs; and overall mixed reactions among the participants to AI-based aids.
