The effect of wrist posture on extrinsic finger muscle activity during single joint movements

Carl R. Beringer, Misagh Mansouri, Lee E. Fisher, Jennifer L. Collinger, Michael C. Munin, Michael L. Boninger, Robert A. Gaunt

Abstract

The extrinsic finger muscles, which originate in the forearm and cross the wrist joint, are subject to significant changes in length and moment arms depending on wrist orientation. This study quantifies how wrist posture modulates the electromyographic (EMG) activity of these muscles during single-joint finger movements. Using fine-wire intramuscular electrodes in 11 able-bodied subjects, we observed that EMG activity varies by up to 70% across different wrist postures. Notably, finger extensors showed peak activity when the wrist was extended, while flexors were most active during wrist flexion. Furthermore, these muscles demonstrated significant activity during wrist movements even when digits remained stationary, suggesting a role in wrist torque generation. Finally, we demonstrate that finger muscle EMG can be used to classify wrist posture, offering potential improvements for the control of myoelectric prostheses.

Introduction

Developing reliable myoelectric prosthetics—artificial limbs controlled by muscle signals—remains a significant challenge in rehabilitative engineering. A primary hurdle is the biomechanical complexity of the "extrinsic" finger muscles. These muscles, which reside in the forearm, span multiple joints including the wrist. Consequently, their mechanical advantage and length change dynamically as the wrist moves.

Previous research into these signals has largely relied on surface EMG (sEMG). However, sEMG is limited by "crosstalk" between adjacent muscles and signal instability caused by the skin shifting over the forearm. To precisely quantify how wrist posture alters the neural signature of finger movements, this study utilizes intramuscular fine-wire electrodes, the gold standard for high-fidelity muscle recording.

Methods

The study involved 11 healthy subjects (7 male, 4 female, average age 28). To capture high-resolution data, researchers employed a sophisticated mapping and recording protocol:

  • Intramuscular EMG: 16 fine-wire bipolar electrodes were inserted into specific muscle compartments, including individual sections of the flexor digitorum profundus, under ultrasound guidance. This allowed for clean signal isolation from deep muscles inaccessible to surface sensors.
  • Kinematic Tracking: Subjects wore a glove with electromagnetic sensors to monitor the precise angles of finger joints during tasks.
  • Experimental Tasks: Subjects performed single-digit flexion and extension movements while maintaining the wrist in five distinct postures: neutral, flexed, extended, pronated (palm down), and supinated (palm up).
  • Classification Analysis: Linear Discriminant Analysis (LDA) classifiers were developed to determine if wrist posture could be predicted solely from the EMG patterns of the finger muscles.

Results

Posture-Dependent Modulation

The study found that EMG activity for identical finger movements differed significantly across the five wrist postures ($p < .05$). The magnitude of this modulation was substantial, with activity levels varying by up to 70%.

A distinct pattern emerged regarding muscle activation: finger extensors were most active when the wrist was extended, and finger flexors were most active when the wrist was flexed. For the flexor muscles, these variations were most pronounced during index finger movements compared to other digits. In contrast, the extensor muscles showed similar modulation across all fingers.

Wrist Torque Contribution

Interestingly, the extrinsic finger muscles showed significant activity during wrist movements even when the digits were held stationary. This suggests that these muscles are not exclusively dedicated to finger manipulation but also play a functional role in generating torque and providing stability for the wrist joint.

Posture Classification

The researchers successfully demonstrated that the EMG signal is sufficiently modulated by posture to allow for automated detection. Using only finger extensor EMG, the system predicted wrist posture with an average accuracy of 63.5% (where chance is 33%). The highest accuracy achieved was 86.9% for specifically detecting an extended wrist posture.

Discussion

The Mechanics of Passive Tension

The finding that finger extensors are most active when the wrist is extended—a seemingly counterintuitive result—can be explained by the "fighting the tension" effect. When the wrist is extended, the opposing flexor muscles are stretched, creating passive tension. The extensors must exert significantly more force to overcome this "rubber band" effect of their antagonists to move the finger.

Compartmental Independence

The use of fine-wire electrodes confirmed that different compartments of the same muscle (e.g., the portion of the muscle controlling the index finger versus the pinky) respond independently to wrist posture. This highlights the high degree of complexity in the neural control of the hand and the limitations of broader surface recordings.

Conclusion

This study demonstrates that wrist posture is a critical variable in the interpretation of finger muscle signals. The 70% variance in EMG activity suggests that current prosthetic control algorithms, which often ignore wrist position, may be prone to errors. By integrating posture-aware signals or using classifiers to predict wrist orientation, future "biomimetic" control systems could become significantly more intuitive for users.

While these results are promising, the study was limited to able-bodied subjects. In transradial amputees, the mechanical coupling of tendons is altered. Future research must determine if these neural "feed-forward" patterns persist after amputation to ensure these findings can be successfully translated into clinical prosthetic design.

Summary updated

Abstract

Abstract Wrist posture impacts the muscle lengths and moment arms of the extrinsic finger muscles that cross the wrist. As a result, the electromyographic (EMG) activity associated with digit movement at different wrist postures may also change. Therefore, we sought to quantify the posture-dependence of extrinsic finger muscle activity. Fine-wire bipolar electrodes were inserted in the extrinsic hand muscles of able-bodied subjects to record EMG activity during wrist and finger movements in various postures. EMG activity of all the recorded finger muscles were significantly different (p&lt;.05, ANOVA) when performing the same movement in five different wrist postures. EMG activity varied by up to 70%, with the highest levels of activity observed in finger extensors when the wrist was extended. Similarly, finger flexors were most active when the wrist was flexed. For the finger flexors, EMG variations with wrist posture were most prominent for index finger muscles, while the EMG activity of all finger extensor muscles were similarly modulated. The extrinsic finger muscles also showed significant activity during wrist movements with the digits held still regardless of finger posture, suggesting that they may play a role in generating torque during wrist movements. Finally, we developed a pair of generalized classifiers that show that finger muscle EMG can be used to predict wrist posture. These results may impact the design of biomimetic control algorithms for myoelectric prosthetic hands, but further work in transradial amputees is necessary to determine whether this phenomenon persists after amputation.
BibTeX
@inproceedings{beringer2020effect,
  author = {Carl R. Beringer and Misagh Mansouri and Lee E. Fisher and Jennifer L. Collinger and Michael C. Munin and Michael L. Boninger and Robert A. Gaunt},
  title = {{The effect of wrist posture on extrinsic finger muscle activity during single joint movements}},
  booktitle = {Scientific},
  year = {2020},
  doi = {10.1038/s41598-020-65167-x},
}
Figure