Handling of the patient.

Do we need machines to move the limbs of our patients on the treadmill? Our current believe is no, no, possibly in the future.

No, because all patients trained thisway stop using their own voluntary activity during the whole or part of the gait cycle. This violates one of the main principles (see above): Use as much and as intensively as possible your remaining activity and thereby enhance it. Activity related motor learning will not happen with passive movements. Elegant animal experiments are currently being performed by the Edgerton group in Los Angeles, which show precisely that this prediction is correct.

No, because nothing can be better in setting, tilting, dorsiflexing etc a limb and can react immediately if the gait does not develop properly than a skilled therapist. An impractical number of sensors on each joint would be needed to report - very fast - the emerging movement, the computing device calculate the deviation and several motors correct the movement via a feedback loop. The active component contributed or anticipated by the patient (which might then be pre-settable) needs to be worked in.

Possibly in the future: Interesting features might evolve from such presetting (see above), if they allow the patient to realize his achieved or expected active contribution. Thisway a feedback type of setting is created, which when properly set, demands active movement by the patient which can be graded according to the patients momentary conditions or the state of the training schedule.

Most recent models of the Locomat (Colombo ETH Zürich, 2005) include some feedback features. However, when we recently (Jan. 2005) tested the device together with a mid thoracic completely paralyzed girl which had been using it for a few weeks, we realized that she had learned to trick the robot: By using her latissimus dorsi muscles she moved her pelvis enough for the machine to assume active hip flexion. And it had taken a considerable amount of time to mount the robot, too much for everyday use. One argument pro robot we often heard (and might in fact be a major reason why it was purchased by some rehab institutions) is that "patients like it". An unusual level of decision making indeed, if it is not accompanied by superior progress in walking capability (which there is no published evidence for up till today, July 2005). The Step Trainer introduced by Hesse has a similar problem: While during stance phase active work by the patient is demanded, swing phase happens passively lest the patient is continually reminded to actively contribute.