Myoelectric prosthesis control using recurrent convolutional neural network regression mitigates the limb position effect

Many myoelectric upper limb prosthesis controllers use pattern recognition, a method that learns and recognizes patterns of electromyographic (EMG) signals produced by the user’s residual limb muscles to predict and execute device movements. Such control becomes unreliable in high limb positions—a problem known as the limb position effect. Pattern recognition often uses a classification algorithm; simple to implement, but limits user-initiated control to only one device movement at a time, at a single speed. To combat position-related control deficiencies and classification controller constraints, we developed and tested two recurrent convolutional neural network (RCNN) pattern recognition-based solutions: (1) an RCNN classification controller that uses EMG plus positional inertial measurement unit (IMU) signals to offer one-speed, sequential movement control; and (2) an RCNN regression controller that uses the same data capture technique to offer simultaneous control of multiple moveme..., , # Myoelectric Prosthesis Control using Recurrent Convolutional Neural Network Regression Mitigates the Limb Position Effect

Dataset DOI: [10.5061/dryad.rv15dv4ks](10.5061/dryad.rv15dv4ks)

## Description of the data and file structure

This dataset contains calculated metrics from 16 non-disabled participants, each testing two myoelectric prosthesis control strategies: 1) either a recurrent convolutional neural network-based classification model (RCNN-Class) or a recurrent convolutional neural network-based regression model (RCNN-Reg), and 2) a linear discriminant analysis classification baseline (LDA-Baseline). 

### Files and variables

#### File: RCNN\_vs\_LDA-Baseline.csv

**Description:** 

##### Variables

* ParticipantID: Randomly assigned 3-digit participant identification number.
* ControlStrategy: Which control strategy was used for the given trial: linear discriminant analysis baseline classification (LDA-Baseline) or recurrent convolutional neural network classification wit..., All participants provided explicit written consent for their de-identified data to be shared in the public domain. Prior to data sharing, all personally identifiable information (PII) was removed to ensure participant anonymity. Electromyographic (EMG) and inertial measurement unit (IMU) data were anonymized by assigning randomized participant codes. The resulting dataset contains only non-identifiable signal data and task labels, ensuring compliance with ethical standards for human subject research and data sharing.