Principal components of wrist circumduction from electromagnetic surgical tracking

Abstract

Purpose

An electromagnetic (EM) surgical tracking system was used for a functionally calibrated kinematic analysis of wrist motion. Circumduction motions were tested for differences in subject gender and for differences in the sense of the circumduction as clockwise or counter-clockwise motion.

Methods

Twenty subjects were instrumented for EM tracking. Flexion–extension motion was used to identify the functional axis. Subjects performed unconstrained wrist circumduction in a clockwise and counter-clockwise sense. Data were decomposed into orthogonal flexion–extension motions and radial–ulnar deviation motions. PCA was used to concisely represent motions. Nonparametric Wilcoxon tests were used to distinguish the groups.

Results

Flexion–extension motions were projected onto a direction axis with a root-mean-square error of $5^\circ$. Using the first three principal components, there was no statistically significant difference in gender (all $p>0.20$). For motion sense, radial–ulnar deviation distinguished the sense of circumduction in the first principal component ($p<0.03$) and in the third principal component ($p<0.01$); flexion–extension distinguished the sense in the second principal component ($p<0.01$).

Conclusion

The clockwise sense of circumduction could be distinguished by a multifactorial combination of components; there were no gender differences in this small population. These data constitute a baseline for normal wrist circumduction. The multifactorial PCA findings suggest that a higher-dimensional method, such as manifold analysis, may be a more concise way of representing circumduction in human joints.