This study reports an investigation into the use of embroidered electrodes for recording surface EMG data from amputees, and the feasibility of using that data as a discrete control interface for a prosthesis through classification of muscle patterns. During data recording, participants are asked to visualise performing one of four gestures with their ghost limb (corresponding to the desired movement of a prosthetic hand) while surface EMG data is collected using the textile-based system. The four classes of gesture are those most commonly used in human hand motion: fist contraction, index finger extension, wrist extension, wrist flexion.Total data size 54.3MB To make downloading quicker and easier the files are accessible as a single zipped folder (11.9MB)