Dataset Information

Determining grasp selection from arm trajectories via deep learning to enable functional hand movement in tetraplegia.

ABSTRACT: Background:Cervical spinal cord injury severely affects grasping ability of its survivors. Fortunately, many individuals with tetraplegia retain residual arm movements that allow them to reach for objects. We propose a wearable technology that utilizes arm movement trajectory information and deep learning methods to determine grasp selection. Furthermore, we combined this approach with neuromuscular stimulation to determine if self-driven functional hand movement could be enabled in spinal cord injury participants. Methods:Two cervical SCI participants performed arbitrary and natural reaching movements toward target objects in three-dimensional space, which were recorded using an inertial sensor worn on their wrist. Time series classifiers were trained to recognize the trajectories using either a Dynamic Time Warping (DTW) algorithm or a Long Short-Term Memory (LSTM) recurrent neural network. As an initial proof-of-concept, we demonstrate real-time classification of the arbitrary movements using DTW only (due to its implementation simplicity), which when used in combination with a high density neuromuscular stimulation sleeve with textile electrodes, enabled participants to perform functional grasping. Results:Participants were able to consistently perform arbitrary two-dimensional and three-dimensional arm movements which could be classified with high accuracy. Furthermore, it was found that natural reaching trajectories for two different target objects (requiring two different grasp types) were distinct and also discriminable with high accuracy. In offline comparisons, LSTM (mean accuracies 99%) performed significantly better than DTW (mean accuracies 86 and 83%) for both arbitrary and natural reaching movements, respectively. Type I and II errors occurred more frequently for DTW (up to 60 and 15%, respectively), whereas it stayed under 5% for LSTM. Also, DTW achieved online accuracy of 79%. Conclusions:We demonstrate the feasibility of utilizing arm trajectory information to determine grasp selection using a wearable inertial sensor along with DTW and deep learning methods. Importantly, this technology can be successfully used to control neuromuscular stimulation and restore functional independence to individuals living with paralysis. Trial registration:NCT, NCT03385005. Registered September 26, 2017.

SUBMITTER: Bhagat N

PROVIDER: S-EPMC7449026 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Determining grasp selection from arm trajectories via deep learning to enable functional hand movement in tetraplegia.

Bhagat Nikunj N King Kevin K Ramdeo Richard R Stein Adam A Bouton Chad C

Bioelectronic medicine 20200825

<h4>Background</h4>Cervical spinal cord injury severely affects grasping ability of its survivors. Fortunately, many individuals with tetraplegia retain residual arm movements that allow them to reach for objects. We propose a wearable technology that utilizes arm movement trajectory information and deep learning methods to determine grasp selection. Furthermore, we combined this approach with neuromuscular stimulation to determine if self-driven functional hand movement could be enabled in spin ...[more]

PMID: 32864392

Similar Datasets

Project description:BackgroundParalysis after spinal cord injury involves damage to pathways that connect neurons in the brain to peripheral nerves in the limbs. Re-establishing this communication using neural interfaces has the potential to bridge the gap and restore upper extremity function to people with high tetraplegia.ObjectiveWe report a novel approach for restoring upper extremity function using selective peripheral nerve stimulation controlled by intracortical microelectrode recordings from sensorimotor networks, along with restoration of tactile sensation of the hand using intracortical microstimulation.MethodsA right-handed man with motor-complete C3-C4 tetraplegia was enrolled into the clinical trial. Six 64-channel intracortical microelectrode arrays were implanted into left hemisphere regions involved in upper extremity function, including primary motor and sensory cortices, inferior frontal gyrus, and anterior intraparietal area. Nine 16-channel extraneural peripheral nerve electrodes were implanted to allow targeted stimulation of right median, ulnar (2), radial, axillary, musculocutaneous, suprascapular, lateral pectoral, and long thoracic nerves, to produce selective muscle contractions on demand. Proof-of-concept studies were performed to demonstrate feasibility of a bidirectional brain-machine interface to restore function of the participant's own arm and hand.ResultsMulti-unit neural activity that correlated with intended motor action was successfully recorded from intracortical arrays. Microstimulation of electrodes in somatosensory cortex produced repeatable sensory percepts of individual fingers for restoration of touch sensation. Selective electrical activation of peripheral nerves produced antigravity muscle contractions. The system was well tolerated with no operative complications.ConclusionThe combination of implanted cortical electrodes and nerve cuff electrodes has the potential to allow restoration of motor and sensory functions of the arm and hand after neurological injury.

Project description:BackgroundAssessment of arm-hand use is very important in children with cerebral palsy (CP) who encounter arm-hand problems. To determine validity and reliability of new instruments to assess actual performance, a set of standardized test situations including activities of daily living (ADL) is required. This study gives information with which such a set for upper extremity skill research may be fine-tuned, relative to a specific research question. Aim of this study is to a) identify upper extremity related ADL children with CP want to improve on, b) determine the 10 most preferred goals of children with CP, and c) identify movement components of all goals identified.MethodThe Canadian Occupational Performance Measure was used to identify upper extremity-related ADL preferences (goals) of 53 children with CP encountering arm-hand problems (mean age 9 ± 4.5 year). Goals were ranked based on importance attributed to each goal and the number of times a goal was mentioned, resulting in a gross list with goals. Additionally, two studies were performed, i.e. study A to determine the 10 most preferred goals for 3 age groups (2.5-5 years; 6-11 years, 12-19 years), based on the total preference score, and study B to identify movement components, like reaching and grasping, of all goals identified for both the leading and the assisting arm-hand.ResultsSeventy-two goals were identified. The 10 most preferred goals differed with age, changing from dressing and leisure-related goals in the youngest children to goals regarding personal care and eating for children aged 6-11 years. The oldest children preferred goals regarding eating, personal care and computer use. The movement components 'positioning', 'reach', 'grasp', and 'hold' were present in most tasks. 'Manipulating' was more important for the leading arm-hand, whereas 'fixating' was more important for the assisting arm-hand.ConclusionThis study gave insight into the preferences regarding ADL children with CP would like to improve on, and the movement components characterizing these activities. This information can be used to create a set of standardized test situations, which can be used to assess the validity and reliability of new measurement instruments to gauge actual arm-hand skilled performance.

Dataset Information

Determining grasp selection from arm trajectories via deep learning to enable functional hand movement in tetraplegia.

Publications

Determining grasp selection from arm trajectories via deep learning to enable functional hand movement in tetraplegia.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets