TL;DR
While wearable exosuits successfully reduce the physical strain of lifting, these biomechanical benefits do not translate directly into how users perceive, learn, or make decisions. This study reveals that the cognitive and perceptual consequences of force augmentation are complex, often decoupled from muscle activity, and heavily influenced by a user's very first experience with the device.
What Problem Does This Paper Address?
The development of wearable robotics, like soft exosuits, has traditionally focused on "reducing the load"—lowering metabolic costs and muscle activation (EMG). However, human performance is not just about raw physical output; it requires accurate perception of the environment, the ability to adapt to changing conditions, and the cognitive weighing of effort against reward.
The researchers identified a critical gap: we don't fully understand how "making an action easier" affects the brain's internal models. If an object feels lighter because of a suit, does the brain update its expectations of the world, or does it become less flexible in learning new rules? This paper investigates whether the well-documented biomechanical benefits of augmentation actually lead to predictable improvements in perception and cognition.
Methods
The study employed a within-subject, two-session design with 32 healthy participants. In one session, they used a bimanual soft exosuit that assisted elbow flexion; in the other, the suit was worn but powered off. Crucially, the order of these sessions was counterbalanced to test for the effects of initial experience.
Participants performed three distinct tasks:
- Weight Estimation: Lifting boxes of different sizes and masses to test weight perception and the "size-weight illusion" (where smaller objects feel heavier than larger ones of the same mass).
- Probabilistic Learning: A volatile task where participants had to predict if a box would be "heavy" or "light" based on visual cues, with the rules reversing periodically.
- Effort-Reward Decision-Making: Choosing between a low-effort/low-reward lift and a high-effort/high-reward lift to see how augmentation alters the subjective cost of work.
The team recorded surface EMG from the biceps, triceps, deltoids, and lower back to quantify physical effort, alongside verbal ratings of perceived weight and fatigue.
Datasets
The study relied on primary data collected from 32 healthy adults (19 female, mean age ~24.5). The researchers did not use external public datasets but generated a rich internal dataset of EMG amplitudes, perceptual ratings, and choice behavior across 70+ trials per task.
Key Insights and Ideas
- The "First Impression" Anchor: The effect of the exosuit on perceived weight was dominated by session order. If participants started without the suit, the suit made things feel significantly lighter later. If they started with the suit, the perceptual benefit was muted. This suggests that our first interaction with a new technology sets a persistent "reference point" that shapes all subsequent experiences.
- Cognitive Flexibility Trade-off: Surprisingly, assistance actually reduced behavioral flexibility. In the learning task, participants were slower to adapt when old rules were reintroduced while wearing the suit.
- Perception vs. Reality: Perceived fatigue did not directly track muscle activity. While the suit reduced overall fatigue, the rate at which fatigue accumulated was actually faster in the assisted condition. This was partly because the reduced physical cost encouraged participants to choose heavier, higher-reward loads more often.
- Biomechanical Decoupling: The most significant "aha moment" is that the 31% reduction in biceps activity did not lead to a 31% change in how heavy things felt or how hard people were willing to work. The brain's interpretation of the "assisted" signal is not a linear reflection of the muscle's relief.
Key Results
- Physical Benefit: The exosuit reliably reduced biceps EMG activity by approximately 31% across all lifting tasks.
- Weight Perception: Assistance reduced perceived weight, but this was strongly shaped by initial experience (Interaction: $F(1, 1591) = 121.98, p < .001$). The size-weight illusion itself remained robust regardless of assistance.
- Learning Dynamics: Accuracy in the volatile environment increased more steeply in the unassisted condition compared to the assisted condition for specific contingencies ($p < .001$), indicating a selective reduction in learning flexibility under augmentation.
- Decision-Making: Participants who started with the assisted session maintained high-effort choices longer, whereas those who started unassisted showed a decline in high-effort choices over time, even when the suit was eventually provided.
Limitations and Future Work
The authors acknowledge that this study looked at a single experimental session. It remains unknown whether these perceptual and cognitive "lags" persist after weeks or months of training. As users become "embodied" with the device, their internal models might eventually align with the biomechanical benefits.
Future research should focus on longitudinal studies and explore these effects in clinical populations (e.g., stroke rehabilitation) or high-stakes occupational settings where accurate perception and rapid learning are mission-critical. The takeaway for engineers is clear: evaluating an exosuit based on EMG alone is only telling half the story.
Summary updated