Effects of Myoelectric Channel Count and Targeting for Upper Limb Prosthetic Control

The objective of this study is to systematically evaluate whether or not two commonly held fundamental assumptions for pattern recognition control, that have been established nearly exclusively on offline data analysis from able-bodied subjects, translates to functional performance when tested on persons with upper limb difference wearing a physical prosthesis.

The two prior assumptions that will be evaluated are: 1) there is no advantage to targeting myoelectric channels to specific muscle sites, and 2) there is no advantage to using more than 4-6 myoelectric channels for trans-radial prosthesis users.

The investigators will perform a randomized crossover study to compare system configurations with different number of sensor channels and with various strategies for sensor placement on the limb. Unlike much of the existing research, this study will test these configurations with subjects with limb absence controlling physical prosthetic arms to directly conduct activities of daily living.

The specific aims of this study are:

1. To evaluate the effect of changing untargeted myoelectric channel count on the functional performance of transradial prosthesis users wearing a fully functional prosthesis via functional outcomes measures such as the AMULA, ACMC, and Clothespin relocation tests (described below). 2. To evaluate the effect of changing myoelectric channel targeting of a fixed number of myoelectric channels on the functional performance of transradial prosthesis users wearing a fully functional prosthesis via functional outcomes measures such as the AMULA, ACMC, and Clothespin relocation tests.

Experimental check sockets will be made for each participant. Two sockets will be made (one with targeted channels and another with untargeted channels). Sockets are for in lab use only. The Coapt Pattern Recognition control system, including Coapt Gen1 and Coapt Gen2 hardware, will be used for this study. Coapt control system is a clinically focused software application interface with channel selection, control calibration tools, and a gaming interface. A Coapt EMG Cuff will be provided for home use.

Commercial prosthetic components are from multiple manufacturers. Commercial wrist units include devices such as the Motion Control wrist rotator. Commercial hand units include devices such as the TASKA CX Hand. These listed devices are commercially available.

Once the subject has provided consent, they will be assigned to a random test sequence that includes the following four different EMG channel conditions.

Condition A: 4 untargeted channels Condition B: 8 untargeted channels Condition C: Up to 16 untargeted channels Condition D: 8 targeted channels

Block randomization based on random number generation will be used to equally distribute the 32 subjects into the 4 possible intervention sequences of the Williams design (e.g., ABDC, BCAD, CDBA, and DACB). The order of completion of the condition sequences will be randomized, each visit will involve completing training and testing for the specific assigned condition.

Participants will attend up to 5 in-lab sessions that are expected to last 4 hours. Participants will be monitored during all sessions and will be allowed to take breaks at any time during training and testing as needed. Additional in-lab sessions may be scheduled if any issues arise during the visit (time constraints, issues with completing or recording the intended data from the visit, subjects gets too fatigued to complete the protocol, etc).

During the first visit, the socket will either be duplicated or a cast will be taken to fabricate two experimental sockets, one with targeted electrode locations and another with non-targeted electrode locations. The first socket will follow Coapt's clinical instructions for targeted site location to identify and mark locations for 8 targeted EMG channels (Condition D). Untargeted sites, for the second socket, will ideally be located by creating a band of electrode pairs at the largest muscular circumference distal to the elbow. Three untargeted channel conditions will be evaluated. One will use the full set of 16 channels (Condition C). The 4-channel untargeted condition (Condition A) will utilize every fourth channel, and their location will be targeted at positions that are 'on the 45s' when examining a cross section of the forearm with the elbow bent to 90 degrees and the thumb pointed up (blue-colored channels. The 8-channel condition (Condition B) will use the same channels as the 4-channel condition but add the four channels that bisect the original four channels (blue and orange colored channels. Due to different limb sizes a full singular ring of 16 channels may not be possible on all human subjects. In these situations, the clinicians fabricating will use their best clinical judgement in order to fit as many electrode channels as possible onto the arm.

During the first visit, participants will also be fit with a Coapt EMG Cuff for at-home use between the visits. During visits 2-5, participants will don the appropriate socket for the condition being tested and connect to the Coapt software. Participants will run through the coapt calibration procedure and conduct training with the study team OT which will be similar to training received as standard of care when receiving a new prosthetic device. Participants will then conduct a series of functional outcome measures designed for upper limb prosthesis users such as the Assessment for the Capacity for Myoelectric Control (ACMC), Activities Measure for Upper Limb Amputees (AMULA), and the Clothespin Relocation Test. Subjects may also be asked to complete virtual games within the Coapt software and complete custom surveys about their experiences with the control configuration during that testing session. Additionally, subjects will be asked to answer custom surveys about their experiences with the control configuration during that testing session, such as "rate how responsive the device was to your intended movements on a scale of 1-10." The participant will then go home for approximately 1-2 weeks. When they return for their next visit, they will follow the randomized order of test conditions and repeat training and testing described above for the specific assigned condition.

Between visits, participants will be sent home with the EMG Cuff and a tablet that will allow them to use the Coapt software to practice using pattern recognition control in a virtual environment. Participants will be asked to play virtual games for 15 minutes a day, 3 days a week between visits. The time between visits is expected to be about 1-2 weeks. The goal of the games is to strengthen the participant's ability to isolate individual motions, understand the intensity of their muscle contractions, to improve control in their muscle contraction intensity to better utilize proportional speed when controlling a prosthesis.