Special collection in association with the 2023 International Conference on aging, innovation and rehabilitation

Babak Taati, Miloš R. PopovićBioMedical 2024 · DOI · OpenAlex

TL;DR

This editorial introduces a curated collection of 12 high-impact research articles stemming from the 2023 International Conference on Aging, Innovation, and Rehabilitation (ICAIR). The collection highlights how the convergence of artificial intelligence, computer vision, and functional electrical stimulation is creating more accessible, low-cost solutions for monitoring and treating age-related disabilities.

What Problem Does This Paper Address?

As global populations age, there is an urgent need for technologies that enhance the quality of life for individuals with disabilities. However, a significant gap exists between sophisticated laboratory "gold standard" equipment and what is actually feasible in a clinical or home setting.

For example, while visual-feedback balance training (VFBT) is effective for spinal cord injury rehabilitation, it traditionally relies on expensive force plates that are inaccessible to many clinics. This editorial frames the collection as a response to these challenges, seeking interdisciplinary solutions that are both technologically advanced and practically deployable.

Methods

As an editorial, this paper summarizes the methodologies found across the 12 accepted manuscripts. Key technical themes include:

  • Computer Vision & Pose Tracking: Utilizing depth cameras and machine learning to replace expensive hardware for tracking body center of mass and center of pressure.
  • Signal Processing & AI: Applying deep learning (e.g., convolutional neural networks) to complex biological data, such as peripheral nerve recordings and acoustic signals for disease detection (like ALS).
  • Hardware Innovation: Developing novel materials, such as dry polymer nanocomposite electrodes, to improve the delivery of functional electrical stimulation (FES).
  • Predictive Modeling: Using vision-based ML models to quantify the efficacy of medical interventions, such as deep brain stimulation for Parkinson’s disease.

Datasets

While the editorial does not introduce a single new benchmark dataset, it references several specialized data collection efforts described in the included papers:

  • Diabetic Foot Ulcer Metadata: A protocol for standardized image and metadata collection to enable deep learning in wound analytics.
  • Gait Videos: Natural gait video datasets of older adults with dementia used to estimate Parkinsonism severity.
  • Simulated Nerve Activity: Data used for training CNNs to record selective peripheral nerve activity.

Key Insights and Ideas

  • Democratizing Rehabilitation Tech: A recurring "aha moment" in this collection is the realization that consumer-grade sensors (like depth cameras) can achieve clinical-grade accuracy. Lim et al. [2] demonstrated that these sensors can accurately track balance metrics, potentially removing the "force plate barrier" for clinics.
  • AI as a Longitudinal Monitor: Rather than just a one-time diagnostic tool, AI is being positioned as a way to measure responsiveness. Sabo et al. [7] showed that ML models can detect subtle changes in gait that occur specifically in response to medication or deep brain stimulation.
  • The "KITE" Ecosystem: The editorial highlights the importance of dedicated research environments like the KITE Research Institute, which houses specialized facilities like DriverLab, ClimateLab, and StreetLab to test rehabilitation tech in realistic, simulated environments.

Key Results

  • Selection Rigor: Out of the numerous abstracts submitted to ICAIR 2023, only a small subset of the highest-scoring presentations were invited to submit full manuscripts. After a standard peer-review process, 12 papers were accepted.
  • Technical Validation: The referenced studies successfully validated that computer vision techniques can be applied to diverse domains: from shoulder exercise monitoring and stair recognition for human-robot locomotion to detecting bulbar ALS through acoustic analysis.
  • Institutional Success: The success of the 2023 conference has led to ICAIR becoming an annual event in Toronto, ensuring a steady pipeline of innovation in the aging and rehabilitation sector.

Limitations and Future Work

  • Clinical Implementation Gap: While the papers demonstrate experimental success, the authors acknowledge that the transition from "experimentally demonstrated" to "clinically implemented" remains a hurdle due to cost and accessibility of certain legacy systems.
  • Scalability: Future work involves refining these AI models to ensure they work across diverse populations and in "wild" (non-laboratory) environments.
  • Annual Expansion: The editors plan to continue this special collection annually, inviting the most "exciting research projects" from future ICAIR conferences to build a longitudinal body of work in this field.

Summary updated

BibTeX
@inproceedings{taati2024special,
  author = {Babak Taati and Miloš R. Popović},
  title = {{Special collection in association with the 2023 International Conference on aging, innovation and rehabilitation}},
  booktitle = {BioMedical},
  year = {2024},
  doi = {10.1186/s12938-024-01243-x},
}

Loading related content...

Figure