In a number of patient categories, problems in performing activities of daily living are expressed in upper limb (non)-usage. So far, there were no objective outcome measures available to determine functional problems during daily living of patients with diseases involving the upper limb(s) [1]. Therefore, we performed a feasibility study to extend the possibilities of the Activity Monitor (AM), which has been developed at our department. The AM consists of a number of piezoresistive acceleration sensors connected to a portable recorder. Reliability and validity of measuring basic activities (e.g. lying, sitting, walking) using the standard configuration of 3 or 4 sensors attached on thighs and trunk were found to be good [2]. To measure upper limb (non)-usage, however, this standard configuration was not sufficient.

The aim of this study was to develop a reliable and valid instrument, which can discriminate between upper limb usage and non-usage during performance of activities of daily living. Upper limb usage was defined as active displacement of (parts of) the upper limb(s) in relation to proximal parts, holding objects and/or leaning. We formulated several requirements that had to be met: measurement periods of minimally 24 hours, usage of small low power consuming sensors, quantifiable bilateral upper limb measurement for comparison, applicable for both healthy subjects and patients and comfortable wearing. Sixteen subjects performed an activity protocol while wearing the AM (for a picture of the standard configuration of the AM, see [4]; for this study we used two additional sensors on both forearms of the subjects). Video recordings were made as a reference. For analysis we synchronised the video recordings and the raw acceleration signals. We used Vitagraph Basic Software and the time-series oriented Signal Processing and Inferencing Language (S.P.I.L.). In the current method of signal processing, three feature signals are derived from each raw acceleration signal: the angular, motility and frequency feature [3]. It was found that the motility feature was most suitable to discriminate between usage and non-usage. This feature is created after high pass filtering, rectifying and smoothing and depends on the variability of the raw signal around the mean. The standard configuration to measure basic activities was maintained to support and improve the discrimination between usage and non-usage.

Preliminary results with two extra sensors on each forearm, just proximal to the dorsal side of the wrist, with their sensitive axis perpendicular to the body segment offer good perspectives. Use of the motility feature to detect upper limb displacement was almost without any problems, while detection of holding and leaning was clearly more difficult. Future research will be aimed at improving the discrimination between upper limb usage and non-usage during measurements in the patients' home environment.

References

1. Schasfoort, F.C.; Bussmann, J.B.J.; Stam, H.J. (2000). Outcome measures for complex regional pain syndrome type I: an overview in the context of the international classification of impairments, disabilities and handicaps. Disability and Rehabilitation, 22, 387-398.
2. Bussmann, J.B.; Tulen, J.H.M.; van Herel, E.C.G.; Stam, H.J. (1998). Quantification of physical activities by means of ambulatory accelerometry: a validation study. Psychophysiology, 35, 488-496.
3. Bussmann, J.B.; Veltink, P.H.; Koelma, F.; van Lummel, R.C.; Stam, H.J.. (1995). Ambulatory monitoring of mobility-related activities; the initial phase of the development of an Activity Monitor. European Journal of Physical Medicine and Rehabilitation, 5, 2-7.
4. Bussmann, J.B.J.; Schasfoort, F.C.; Tulen, J.H.M.; van den Berg, H.J.G.; Stam, H.J. (2000). Accelerometry-based activity monitoring: objective measurement of postures and motions. This volume.

© 2000 Noldus Information Technology b.v.