Comprehensive Summary
This study analyzes the use of AI in instrumental gait analysis to find specific challenges and limitations, and proposes steps to take toward alleviating these struggles. Scientific literature surrounding the use of AI in instrumental gait analysis was studied and combined with the authors’ own clinical experiences. After research, six main challenges of the use of AI in this context were identified. The challenges present were: scarcity of data, large number of influencing factors, unclear target variables, high measurement variability, generalization difficulties, and limited clinical acceptance. The study also listed solutions for these struggles, in the same order: transfer learning and data augmentation, including clinical expert knowledge, identifying meaningful clinical endpoints, standardization of data collection protocols, larger and more robust training strategies, and including physicians in development.
Outcomes and Implications
The use of AI in instrumental gait analysis could simplify a difficult process for physicians. These analyses can tell a physician a lot about a patient and give important insight into possible difficulties they may be having. However, per the results of this study, the use of AI in instrumental gait analysis is not yet as developed as it needs to be to be an effective tool in a clinical setting. Once the struggles found can be overcome, AI could play a massive role in clinical instrumental gait analysis.