Limb Manipulation

Limb Manipulation under patient control

Charles G. Burgar MD; H. F. Machiel Van der Loos, PhD; Peter S. Lum, PhD; Felix E. Zajac, PhD; Michelle Johnson, MS

Stroke is the most common inpatient rehabilitation diagnosis. The prevalence of this diagnosis is increasing within the DVA, as aging World War II veterans join the population at highest risk. Upper limb paralysis is often resistant to therapeutic efforts. The resulting impairments in mobility and activities of daily living have a significant negative impact on quality of life, independent living, and hospital length of stay. Until significant gains are made in prevention of stroke and its sequelae, more effective methods are needed to maintain or restore the dexterity necessary for daily activities.

In this project, we are evaluating a new approach for assessing and facilitating upper limb movement. To test the benefit and feasibility of semi-automated therapy systems, we have developed a prototype device that allows patient control of active-assisted movement. Normally, this commonly used therapeutic technique is implemented by a therapist who moves the paretic limb as the patient attempts to contribute to the movement. Using our experience with servomechanisms, we have assembled a device which places control of these movements in the hands of the patient: the Mirror-Image Movement Enabler (MIME).

Figure 1. MIME: A robot servomechanism provides the assistance necessary for left arm movement in trajectories determined by movement of the right arm.

MIME incorporates two mobile arm supports that allow free movement of the forearms in a horizontal plane, while supporting the arms against gravity (Fig. 1) Movements of the "normal" limb (ipsilateral to the cerebrovascular accident) are sensed by position encoders, reflected symmetrically about the sagittal plane, and reproduced in the paretic limb with the assistance, as necessary, of external forces applied by a small industrial robotic arm. In preliminary testing, MIME has safely enabled movement of the left arms of three healthy volunteers and three persons with left hemiplegia, in trajectories commanded by right arm movements under voluntary control. The external forces supplied by MIME reflect the differences in level of motor recovery of the subjects (Fig. 2).

The top view (above) of the subject shows the initial arm position (solid) and the final arm position (dashed) of the movements. The top panels (left) show the y movement of the left forearm center of mass for a passive (light lines) and an active trial (heavy lines). The bottom panels show the y forces from the left arm during these movements. The active Fy profile minus the passive Fy profile is the force generated.

Figure 2. Typical trials for one normal and two stroke subjects when assisted by MIME

In the current pilot project, we will test MIME with ten hemiplegic stroke patients having a wide range of motor impairment and with three neurologically normal, non-disabled individuals. We expect all hemiplegic subjects, when assisted by MIME, to perform bilateral mirror-image movements as accurately as unassisted normal subjects. We also expect the assistance forces to vary systematically and predictably with the clinical condition of that limb (as measured by an upper extremity Fugl-Meyer assessment). The forces that MIME applies to the paretic limb will increase in magnitude and variability as a subject progresses from the flaccid stage to the hypertonic and synergistic stages. The forces will then decrease in magnitude, variability, and phase error as the subject recovers volitional control.

A second project, spawned by the MIME, is Driver's SEAT (Simulated Environment for Arm Therapy). This project allows us to explore the main issues raised by MIME concerning self-initiated bimanual upper limb therapy in a simplified, one-degree-of-freedom control environment. Currently, the simulator is being used to assess and assist the upper limb motor function of persons with hemiplegia due to stroke. The design has been implemented to allow hemiplegic subjects to experience three steering wheel torque control strategies corresponding to normal driving, assisted, or resisted movement. The hypothesized effect on a subject's rehabilitation is an increase in the level of upper-limb bimanual function and a decrease in the total time needed to achieve recovery. Preliminary testing with normal subjects has confirmed that the torque control and tracking strategies result in meaningful torque profiles. These findings support the MIME results, and represent the first step in validating the Driver's SEAT concept.

The pilot data collected in this proposal will allow us to design future experiments to study the loss and recovery of upper limb function following stroke. In these future studies, we will use the kinematics and kinetics of active-assisted movements, along with muscle electromyographic measurements, to develop a musculoskeletal model of the upper extremity and to investigate the mechanisms of motor impairment due to stroke.

We view MIME, in its current state, as a research tool which will aid in validating the merits of patient-controlled limb manipulation for stroke rehabilitation. In future studies, we will also investigate the potential utility of this technique in the very acute stages of stroke recovery when it might have a beneficial effect on functional outcomes.

Republished from the 1996 Rehabilitation R&D Center Progress Report. For current information about this project, contact: Charles G. Burgar.