Objective: Wheelchair safety is of great importance since falls from wheelchairs are prevalent and often have devastating consequences. We developed an automatic system to detect destabilizing events during wheelchair propulsion under real-world conditions and trigger neural stimulation to stiffen the trunk to maintain seated postures of users with paralysis.

Design: Cross-over intervention

Setting: Laboratory and community settings

Participants: Three able-bodied subjects and three individuals with SCI with previously implanted neurostimulation systems

Interventions: An algorithm to detect wheelchair sudden stops was developed. This was used to randomly trigger trunk extensor stimulation during sudden stops events

Outcome Measures: Algorithm success and false positive rates were determined. SCI users rated each condition on a seven-point Usability Rating Scale to indicate safety.

Results: The system detected sudden stops with a success rate of over 93% in community settings. When used to trigger trunk neurostimulation to ensure stability, the implant recipients consistently reported feeling safer (P<.05 for 2/3 subjects) with the system while encountering sudden stops as indicated by a 1–3 point change in safety rating.

Conclusion: These preliminary results suggest that this system could monitor wheelchair activity and only apply stabilizing neurostimulation when appropriate to maintain posture. Larger scale, unsupervised and longer-term trials at home and in the community are indicated. This system could be generalized and applied to individuals without an implanted stimulation by utilizing surface stimulation, or by actuating a mechanical restraint when necessary, thus allowing unrestricted trunk movements and only restraining the user when necessary to ensure safety.

Trial Registration: NCT01474148

目的：轮椅安全非常重要，因为从轮椅上摔下来很普遍，而且往往会带来毁灭性的后果。我们开发了一种自动系统，用于检测现实条件下轮椅推进过程中的不稳定事件，并触发神经刺激以使躯干变硬，以保持瘫痪用户的坐姿。

设计：交叉干预

环境：实验室和社区环境

参与者：三名身体健全的受试者和三名脊髓损伤患者，之前植入了神经刺激系统

干预：开发了一种检测轮椅突然停止的算法。这用于在突然停止事件期间随机触发躯干伸肌刺激

结果测量：确定算法成功率和误报率。SCI 用户使用 7 点可用性评级量表对每个条件进行评级，以表明安全性。

结果：系统在社区环境中检测到突然停止的成功率超过 93%。当用于触发躯干神经刺激以确保稳定性时，植入物接受者一致报告说在系统遇到突然停止时感觉更安全（2/3 受试者P <.05），如安全等级变化 1-3 点所示。

结论：这些初步结果表明，该系统可以监测轮椅活动，并且仅在适当时应用稳定神经刺激以保持姿势。需要在家里和社区进行更大规模、无监督和长期的试验。该系统可以通过利用表面刺激或在必要时通过启动机械约束来推广并应用于没有植入刺激的个体，从而允许不受限制的躯干运动并且仅在必要时约束用户以确保安全。

试用注册： NCT01474148