Résumé :
|
There is currently no standardized method to assess upper limb activity in non-ambulatory patients. Actimetry, with use of motion sensors, are probably the most promising method. Indeed, it provides continuous monitoring of specific part of body whose movement is being studied. Moreover, it offers the opportunity to be used in free-living environment for longitudinal follow-up. The main objective of this study is to develop an actimetric method for the movement characterization of upper limbs. However, we still have to demonstrate that pertinent indicators of upper limb functional activity can be obtained from sensors output. We used a prototype integrating three-axis accelerometer, gyroscope and magnetometer. This device measures accelerations, angular velocities, magnetic field and gives an estimation of spatial orientation. Several healthy subjects were asked to perform the validated task Box & Block Test while wearing the device on the wrist of dominant hand. They were also asked to perform a task representative of daily life: typing a predefined text on a computer, score being the number of characters typed in one minute. Each test was repeated fifty times with various velocities. Then, signals captured by sensors were processed to obtain parameters representing angular velocity and acceleration magnitudes. The angles were used to dissociate gravity effect from acceleration due to real movement. Lastly, computed parameters were plotted with the scores obtained in the different tasks. For each task, exponential regressions indicate a strong correlation between average acceleration magnitude and corresponding scores (R=0.95). In the same way, linear regression indicates a strong correlation between average angular velocities magnitude and corresponding scores (R=0.97). Consequently all these results suggest that subject's efficiency could be well predicted by accelerations and angular velocities in some validated and daily life tasks. In conclusion, this preliminary study demonstrates that clinically meaningful indicators can be extracted from sensors output to predict the functional status of the upper limbs. This could be further applied to non ambulatory patients.
|