Investigators:
Jon Dekar, DESIN LLC
Zackory Erickson, CMU Robotics Institute

MassAITC Cohort: Year 4 (Aging)

Initial Proposal Abstract: Feeding difficulties are common in individuals with Alzheimer’s Disease and Related Dementias (AD/ADRD), often leading to malnutrition, increased caregiver burden, and higher healthcare costs. While assistive technologies can support motor impairments, few address the cognitive challenges—particularly inattention—that disrupt self-feeding in this population. DESĪN LLC’s commercially available Obi robotic feeding device has proven effective for individuals with physical disabilities, but it lacks the capability to support users with attention-related deficits common in mid to late-stage AD/ADRD.

This project aims to enhance Obi with 3D perception cameras and AI algorithms capable of detecting inattention and delivering gentle, real-time prompts to guide users back to eating. The system will be trained and validated using data from healthy volunteers (Aim 1), then tested in a feasibility study with 30 older adults with mild to moderate AD/ADRD in a long-term care setting (Aim 2). Outcome measures will include inattention detection accuracy, redirection effectiveness, adherence to the device, and feedback from both users and caregivers.

By integrating cognitive support into assistive feeding technology, this project seeks to restore independence in eating for individuals with AD/ADRD, reduce caregiver workload, and improve mealtime outcomes. The results will establish the feasibility of an AI-enhanced Obi, paving the way for future commercialization and broader adoption in both home and institutional care settings.