EMG signal analysis to assess prosthetic hand failure during activities

Abstract

Introduction: myoelectric upper-limb prosthetic retention is low [1] due to prosthetic hand failure caused by EMG electrode artifacts [2]. Observational studies using a range of specific activities have previously assessed prosthetic hand failure [3]. However, there lacks real-time signal processing studies to help underpin the mechanism and assessment of failure in hand activities. This preliminary study synchronises clinical (Ottobock) and research tools (Delsys) to analyse prosthetic hand failure risk.

Methods: a female, able-bodied participant was recruited. The VariPlus Speed prosthetic hand (Ottobock) is mounted to the participant, using a custom-designed 3D-printed cuff, shown in Figure 1. The TRIGNO system (Delsys), was attached proximally to the forearm. Two activities were studied: lifting a cup and placing a cup on a plinth. EMG signals were processed and separated into intended activation and idle windows. These were analysed based on signal proximity to the activation threshold.

Results & Discussion: figure 2 shows typical EMG signals for placing a cup on a plinth. Figure 3 shows percentage of readings crossing the activation threshold during different phases of the movement. Whilst it was observed that 60% of the recording sessions resulted in prosthetic failure, all sessions showed high percentages (~20%) of EMG data near the threshold of movement throughout the whole session. Most activities showed comparable percentages of EMG data crossing the threshold during periods of no intended activation (~9%) to periods of intended activation (~12%), signifying risk of prosthetic failure. Notably, placing the cup showed higher chance of prosthetic failure (~14%) than simply lifting (~10%), suggesting the lateral movement influences electrode artefacts.

Conclusion: by using observational data alone, it may be challenging to fully understand to risk of prosthetic failure for a given action. Signal analysis helps to determine when prosthetic failure is more likely to occur and during which activities. This is made possible by the unique combination of engineering and clinical tools.

More information

Identifiers

ORCID for Kirstie Devin: orcid.org/0000-0001-6794-2375

ORCID for J. Tang: orcid.org/0000-0003-3359-5891

ORCID for L. Jiang: orcid.org/0000-0002-3400-825X

Catalogue record

Export record