Exploration of A Brain Activity-Metabolic Cost Relationship for Human-in-the-loop Optimization during Incline Walking

Exoskeletons have the potential to enhance human performance, with the design of effective human-machine interaction (HMI) playing a crucial role in improving operability. However, the quest for optimal human-machine control remains an open field for further investigation. The key to success lies in establishing a flexible mode of communication between humans and machines, employing diverse methods and adjustments of parameters. Tailoring optimization to each individual user is pivotal for achieving significant advancements. This study presents a novel approach to human-in-the-loop optimization that leverages the correlation between brain activity and metabolic cost. This approach not only aims to reduce the energy expenditure of users during walking but also seeks to optimize gait patterns for healthy individuals across various walking environments. Our findings reveal that the $\mu$ and $\gamma$ frequency bands display notable event-related synchronization (ERS) and event-related desynchronization (ERD) phenomena during walking. Notably, the $\gamma$ band power is positively correlated with changes in inclined terrain. Pearson correlation indicates a stronger correlation between $\gamma$ band power and metabolic cost than in other bands. This study thus built a regression model that links brainwave patterns to metabolic rates, enabling the prediction of current metabolic costs based on brain activity. This relationship could facilitate the realization of human-in-the-loop optimization, enhancing the walking economy.