Limits...
A Novel Percutaneous Electrode Implant for Improving Robustness in Advanced Myoelectric Control.

Hahne JM, Farina D, Jiang N, Liebetanz D - Front Neurosci (2016)

Bottom Line: Moreover, being percutaneous, it does not require power and data telemetry modules.The implants showed significantly lower impedance and greater robustness against mechanical interference than traditional surface EMG electrodes used for myoelectric control.Moreover, the EMG signals detected by the proposed systems allowed more stable control performance across sessions in different days than that achieved with classic EMG electrodes.

View Article: PubMed Central - PubMed

Affiliation: Institute of Neurorehabilitation Systems, University Medical Center Göttingen Göttingen, Germany.

ABSTRACT
Despite several decades of research, electrically powered hand and arm prostheses are still controlled with very simple algorithms that process the surface electromyogram (EMG) of remnant muscles to achieve control of one prosthetic function at a time. More advanced machine learning methods have shown promising results under laboratory conditions. However, limited robustness has largely prevented the transfer of these laboratory advances to clinical applications. In this paper, we introduce a novel percutaneous EMG electrode to be implanted chronically with the aim of improving the reliability of EMG detection in myoelectric control. The proposed electrode requires a minimally invasive procedure for its implantation, similar to a cosmetic micro-dermal implant. Moreover, being percutaneous, it does not require power and data telemetry modules. Four of these electrodes were chronically implanted in the forearm of an able-bodied human volunteer for testing their characteristics. The implants showed significantly lower impedance and greater robustness against mechanical interference than traditional surface EMG electrodes used for myoelectric control. Moreover, the EMG signals detected by the proposed systems allowed more stable control performance across sessions in different days than that achieved with classic EMG electrodes. In conclusion, the proposed implants may be a promising interface for clinically available prostheses.

No MeSH data available.


Related in: MedlinePlus

Real-time myoelctric control task. The subject controlled the red cursor with his EMG and had the task to hit the target (green circle) and remain in the target for 1 s. The subject had 10 s to perform this task.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4814550&req=5

Figure 2: Real-time myoelctric control task. The subject controlled the red cursor with his EMG and had the task to hit the target (green circle) and remain in the target for 1 s. The subject had 10 s to perform this task.

Mentions: The performance was evaluated in real-time control tasks, in which the two dimensional output of the linear regressor was used to control the position of a cursor in 2D (Figure 2). Specifically, the subject was asked to hit circular targets by moving the cursor into the target and remaining in the target for 1 s, before a 10 s time-out. The test was repeated over 2 days, twice per day, without re-training. The order in which the electrodes were tested was randomized and blind to the subject. To minimize the shifts over days of the classic electrodes, the position of the two classic electrodes was marked on the skin on the first day. The performance was quantified with two metrics: task completion rate and task completion time. The task completion rate is defined as the ratio of targets that were successfully hit before time-out and the total number of targets presented. The task completion time is the average time spent per target.


A Novel Percutaneous Electrode Implant for Improving Robustness in Advanced Myoelectric Control.

Hahne JM, Farina D, Jiang N, Liebetanz D - Front Neurosci (2016)

Real-time myoelctric control task. The subject controlled the red cursor with his EMG and had the task to hit the target (green circle) and remain in the target for 1 s. The subject had 10 s to perform this task.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4814550&req=5

Figure 2: Real-time myoelctric control task. The subject controlled the red cursor with his EMG and had the task to hit the target (green circle) and remain in the target for 1 s. The subject had 10 s to perform this task.
Mentions: The performance was evaluated in real-time control tasks, in which the two dimensional output of the linear regressor was used to control the position of a cursor in 2D (Figure 2). Specifically, the subject was asked to hit circular targets by moving the cursor into the target and remaining in the target for 1 s, before a 10 s time-out. The test was repeated over 2 days, twice per day, without re-training. The order in which the electrodes were tested was randomized and blind to the subject. To minimize the shifts over days of the classic electrodes, the position of the two classic electrodes was marked on the skin on the first day. The performance was quantified with two metrics: task completion rate and task completion time. The task completion rate is defined as the ratio of targets that were successfully hit before time-out and the total number of targets presented. The task completion time is the average time spent per target.

Bottom Line: Moreover, being percutaneous, it does not require power and data telemetry modules.The implants showed significantly lower impedance and greater robustness against mechanical interference than traditional surface EMG electrodes used for myoelectric control.Moreover, the EMG signals detected by the proposed systems allowed more stable control performance across sessions in different days than that achieved with classic EMG electrodes.

View Article: PubMed Central - PubMed

Affiliation: Institute of Neurorehabilitation Systems, University Medical Center Göttingen Göttingen, Germany.

ABSTRACT
Despite several decades of research, electrically powered hand and arm prostheses are still controlled with very simple algorithms that process the surface electromyogram (EMG) of remnant muscles to achieve control of one prosthetic function at a time. More advanced machine learning methods have shown promising results under laboratory conditions. However, limited robustness has largely prevented the transfer of these laboratory advances to clinical applications. In this paper, we introduce a novel percutaneous EMG electrode to be implanted chronically with the aim of improving the reliability of EMG detection in myoelectric control. The proposed electrode requires a minimally invasive procedure for its implantation, similar to a cosmetic micro-dermal implant. Moreover, being percutaneous, it does not require power and data telemetry modules. Four of these electrodes were chronically implanted in the forearm of an able-bodied human volunteer for testing their characteristics. The implants showed significantly lower impedance and greater robustness against mechanical interference than traditional surface EMG electrodes used for myoelectric control. Moreover, the EMG signals detected by the proposed systems allowed more stable control performance across sessions in different days than that achieved with classic EMG electrodes. In conclusion, the proposed implants may be a promising interface for clinically available prostheses.

No MeSH data available.


Related in: MedlinePlus