Shoulder and arm robot

The arm rests in a spring-loaded support that offsets its own weight. A weak arm that cannot lift against gravity can often move well once gravity is removed, and this is what makes practice possible from an early stage. The amount of support is reduced step by step as strength returns.

Movements drive tasks on a screen — reaching, grasping, sorting — so the patient is doing an activity rather than performing an exercise, and the repetition count climbs accordingly. The system records range, speed and smoothness for every repetition, so change is measured rather than estimated.

Arm and hand weakness after stroke, traumatic brain injury, spinal cord injury at cervical level, in multiple sclerosis, and in children with cerebral palsy affecting the upper limb. It suits patients across a wide range of severity, because the level of support can be set very high at the start.

The arm needs enough passive range at the shoulder to be positioned safely, and a painful subluxed shoulder is treated before loading it into the device.

Upper limb recovery after stroke is slower and less complete than lower limb recovery, and no device changes that. What arm robotics does reliably is increase the number of quality repetitions per session, and repetition volume is one of the few things consistently associated with better outcomes.

Gains in the device transfer to real tasks only if real tasks are trained too. We pair robot time with task practice using actual objects, because a patient who improves on screen and cannot pick up a cup has not improved.

The device supports the weight of the arm, so movement becomes possible for someone who cannot lift against gravity. Hundreds of guided repetitions per session become achievable where manual therapy manages a fraction.

Evidence supports robot-assisted arm training as an addition to conventional therapy rather than a replacement, and that is how it is scheduled here.

A modest one, on average. A Cochrane review of 45 trials with 1,619 people after stroke found small improvements in arm function, arm strength and daily activities with robot-assisted arm training, and adverse events were rare 2. The evidence was rated high quality, but the trials differed widely in training dose, device and patients 2. A later meta-analysis of 46 trials found that the gains were present at the end of treatment but not at follow-up, and appeared only when training added up to more than 15 hours and the patient took an active part 3.

The largest single trial, with 770 people with moderate or severe arm limitation, found that training with a different arm robot did not improve arm function compared with usual care 4.