File
M. Guemann1, C. Halgand1, A. Bastier2, L. Borrini3, D. Cattaert1, A. de Rugy1
1INCIA, Bordeaux, France, 2IRMA, Valenton, France, 3SSA Percy, Clamart, France
Background: Loss of feedback such as proprioception after an amputation is devastating and has many consequences for patients. Multiple sensory feedback systems have been designed to overcome the lack of feedback in prosthesis control. As non invasive devices, Both vibro and electro-tactile stimulations enable good discrimination of skin stimulation which may be a good candidate to use it durion pérosthetic command. However, most studies exploring sensory-substitution have been conducted on healthy subjects, and only few included patients. More studies évaluation functional tasks are needed.
Purpose: With this study, we explore the effect of vibrotactile feedback on myoelectric control along with the workload in a reaching target task under 4 feedback conditions.
Methods: Two groups of people (healthy (16) and limb loss (8)) took part of the experiment. All participants were equiped with a EMG wristlet at the humeral level and a ribbon integrating the 6 actuators that feedback the sensory information. They were seated in front of a computer showing an avatar who's arm was moving thansk to the EMG activity of the participant (biceps for flexion and triceps for extension).
The experiment was composed of two phases. During the first one, participants had to reach targets at different angular positions between 20° to 120° in one movement. Time to reach was proportional to the angular distance. After a short familiarization phase, the test consisted in reaching 30 targets either in flexion or in extension under 4 different feedback conditions : vision, vision + vibration ; vibration ; no feedback.
Distances and feedback were randomly assigned to avoid any learning effect.
The second part of the expriment was more difficult. Participants had to realize back and forth movements between two targets and stabilized each movement before changing of direction. Familiarization and test procdeure were the same than in the first part.
Distance to the target (in degree) was the depdendant variable allowing the comparison between the 4 feedback. Workload and preference were also evaluated using the NASA-TLX.
The experiment was composed of two phases. During the first one, participants had to reach targets at different angular positions between 20° to 120° in one movement. Time to reach was proportional to the angular distance. After a short familiarization phase, the test consisted in reaching 30 targets either in flexion or in extension under 4 different feedback conditions : vision, vision + vibration ; vibration ; no feedback.
Distances and feedback were randomly assigned to avoid any learning effect.
The second part of the expriment was more difficult. Participants had to realize back and forth movements between two targets and stabilized each movement before changing of direction. Familiarization and test procdeure were the same than in the first part.
Distance to the target (in degree) was the depdendant variable allowing the comparison between the 4 feedback. Workload and preference were also evaluated using the NASA-TLX.
Results: For the first part of the experiment (movement in one direction), statistically significant differences were found between the condition without feedback (NONE) and the three other conditions, and between vibration only (VIB) and the three other conditions. No statistical difference was found between vision only (VIS) and vision and vibration (VIS+VIB).
For the back and fort movements, a statistical significate difference between all conditions except between conditions VIS and VIS+VIB were found.
NASA-TLX revealed the best mean (SD) for the VIS+VIB condition 37.20 (14.17). Visual condition obtain 39.03 (19.58) and vibration 70.83 (14.25).
Ten out of 16 participants (62.5%) preferred the VIS+VIB condition.
For the back and fort movements, a statistical significate difference between all conditions except between conditions VIS and VIS+VIB were found.
NASA-TLX revealed the best mean (SD) for the VIS+VIB condition 37.20 (14.17). Visual condition obtain 39.03 (19.58) and vibration 70.83 (14.25).
Ten out of 16 participants (62.5%) preferred the VIS+VIB condition.
Conclusion(s): After only one session of practice, participants showed a preference for the multisensory modality. They quickly understand and integrate the meaning of the vibro-tactile feedback. Those preliminary results are encouraging for the developement of new sensory-substitution devices.
Implications: This work mya helps to reduce the gap between reasearch results and available technology for patients. It may helps to improve daily usage of upper limb prosthesis and decrease the number of prosthesis abandon.
Funding, acknowledgements: Thanks to the Agence Innovation Defense from the french army that have founded the PhD of M. Guemann
Keywords: myoelectric control, upper limb amputation, sensory substitution
Topic: Innovative technology: robotics
Did this work require ethics approval? Yes
Institution: CHU Besançon
Committee: CPP Est II
Ethics number: IDRCB 2017-A03609-44
All authors, affiliations and abstracts have been published as submitted.
