Measuring upper limb activity following stroke using inertial sensors

The overall study aims to measure meaningful upper limb activity using inertial sensors to discourage the upper limb activity of the unaffected arm following stroke and therefore increased activity of the affected arm.

We have measured the affected and non-affected arm movements of 30 chronic stroke survivors (over 2 years post stroke) and 30 age matched controls over 4 days in the context of their home and everyday environment. This has provided comparison data for affected v non-affected arm, dominant v non-dominant and stroke v age matched controls. We also explored the association with a standardised measure (ARAT) to compare accelerometry measures with capability measures. 


The graph above shows affected versus unaffected limbs.


We are still in the process of analysing results however; early indications suggest that controls tend to use their dominant and non-dominant arms at a ratio of 45 – 55% whereas stroke survivors have ratios of as much as 10 – 90% differences. We also found that controls tend to use both arms together more than stroke survivors. The association between the ARAT demonstrated that in some cases, despite scoring maximum scores, the stroke survivors still did not use their affected arm even if it was their dominant arm.



Control group, left versus right arm.



Stroke survivors group, left versus right arm.

 

These results suggest that after a number of years following stroke, over a four-day period in the context of everyday life, stroke survivors do not use their affected arm as much as controls even if they have full capability.

Next steps will involve exploring the accuracy of the sensors by adding gyroscope data to detect meaningful movement (as opposed to involuntary arm use i.e. walking) to provide feedback to discourage the non-affected arm and therefore encourage the affected arm to carry out meaningful activities.


Meet the team:

Dr Jack Parker (Project lead, University of Sheffield): Jack.Parker@sheffield.ac.uk

Miss Lauren Powell (Research Associate, University of Sheffield): L.a.powell@sheffield.ac.uk

Dr Ben Heller (Biomechanist, Sheffield Hallam University): B.heller@shu.ac.uk

Ms Eileen Schweiss (Data analyst, Runscribe): Eileen@runscribe.com