Contrasting action and posture coding with hierarchical deep neural network models of proprioception

Biological motor control is versatile, efficient, and depends on proprioceptive feedback. Muscles are flexible and undergo continuous changes, requiring distributed adaptive control mechanisms that continuously account for the body's state. The canonical role of proprioception is representing the body state. We hypothesize that the proprioceptive system could also be critical for high-level tasks such as action recognition. To test this theory, we pursued a task-driven modeling approach, which allowed us to isolate the study of proprioception. We generated a large synthetic dataset of human arm trajectories tracing characters of the Latin alphabet in 3D space, together with muscle activities obtained from a musculoskeletal model and model-based muscle spindle activity. Next, we compared two classes of tasks: trajectory decoding and action recognition, which allowed us to train hierarchical models to decode either the position and velocity of the end-effector of one's posture or the character (action) identity from the spindle firing patterns. We found that artificial neural networks could robustly solve both tasks, and the networks' units show tuning properties similar to neurons in the primate somatosensory cortex and the brainstem. Remarkably, we found uniformly distributed directional selective units only with the action-recognition-trained models and not the trajectory-decoding-trained models. This suggests that proprioceptive encoding is additionally associated with higher-level functions such as action recognition and therefore provides new, experimentally testable hypotheses of how proprioception aids in adaptive motor control.

A simulation-based framework with a proprioceptive musculoskeletal model for evaluating the rehabilitation exoskeleton system

BACKGROUND AND OBJECTIVE

Various rehabilitation exoskeletons have been designed to help people regain normal gait from stroke effects. However, the evaluation and further optimization of these exoskeletons are not convenient and usually need complicated experimental works. The present study aims to establish a simulation-based method with a proprioceptive musculoskeletal model to conveniently evaluate the efficiency of a self-developed exoskeleton for further optimization.

METHODS

Three volunteers who suffer from dyskinesia due to stroke were recruited for gait experiments with and without the self-develop exoskeleton. The corresponding simulations were implemented based on the proprioceptive model, the exoskeleton model, and the input kinematic data obtained from the experiments. The joint angles, muscle activations, and metabolic costs as well as the proprioceptor feedback stimulation were extracted for comparative analysis.

RESULT

Several positive effects of the exoskeleton were noted based on the simulation results when using it to aid the patients' rehabilitation during the gait training. The CORA scores of the patients' joint angle to the normal data increased by 11.6~37.8% with the assistance of the exoskeleton. The wave frequency of proprioceptive feedback stimulation that can be directly correlated to the neural rehabilitation obviously inclined during a gait cycle. The muscle activations were also rearranged to better support the patient's walk when using the exoskeleton, while the metabolic costs were reduced for all the patients.

CONCLUSION

In summary, the present simulation-based method can be practical for pre-evaluation and optimization of various exoskeleton design in the future.

A leg to stand on: computational models of proprioception

Dexterous motor control requires feedback from proprioceptors, internal mechanosensory neurons that sense the body's position and movement. An outstanding question in neuroscience is how diverse proprioceptive feedback signals contribute to flexible motor control. Genetic tools now enable targeted recording and perturbation of proprioceptive neurons in behaving animals; however, these experiments can be challenging to interpret, due to the tight coupling of proprioception and motor control. Here, we argue that understanding the role of proprioceptive feedback in controlling behavior will be aided by the development of multiscale models of sensorimotor loops. We review current phenomenological and structural models for proprioceptor encoding and discuss how they may be integrated with existing models of posture, movement, and body state estimation.

Recruitment of upper-limb motoneurons with epidural electrical stimulation of the cervical spinal cord

Epidural electrical stimulation (EES) of lumbosacral sensorimotor circuits improves leg motor control in animals and humans with spinal cord injury (SCI). Upper-limb motor control involves similar circuits, located in the cervical spinal cord, suggesting that EES could also improve arm and hand movements after quadriplegia. However, the ability of cervical EES to selectively modulate specific upper-limb motor nuclei remains unclear. Here, we combined a computational model of the cervical spinal cord with experiments in macaque monkeys to explore the mechanisms of upper-limb motoneuron recruitment with EES and characterize the selectivity of cervical interfaces. We show that lateral electrodes produce a segmental recruitment of arm motoneurons mediated by the direct activation of sensory afferents, and that muscle responses to EES are modulated during movement. Intraoperative recordings suggested similar properties in humans at rest. These modelling and experimental results can be applied for the development of neurotechnologies designed for the improvement of arm and hand control in humans with quadriplegia.

The efficacy of epidural electrical stimulation (EES) to engage arm muscles and improve movement after spinal cord injury is still unclear. Here, the authors investigated how EES can recruit upper-limb motor neurons by combining computational modelling with experiments in non-human primates.