HK1207326B - Apparatus and method for restoring voluntary control of locomotion in neuromotor impairments - Google Patents

Description

Apparatus and method for restoring voluntary control of locomotion in neuromotor impairment

Technical Field

The present invention relates to the field of medical engineering, in particular to devices and systems for rehabilitation of injured subjects, more particularly to devices and systems for rehabilitation of motor systems, in particular limbs.

Background

Neuromotor disorders such as Spinal Cord Injury (SCI) and stroke cause different impairments in motor pattern generation and balance (Courtine, G. et al, conversion of non-functional spinal circuits to a functional state after loss of brain input (Transformation of non-functional spinal circuits into the functional state of braine input), Nat Neurosci12, 1333 laid 1342 (2009); Harkema, S.J. et al, human lumbosacral spinal cord interpretation of loading during stepping (Humanmboscal spinal cord interpretation holding), J neurophysis 77, 797-1997).

Therefore, it is necessary to isolate these subfunctions for gait assessment and neurological rehabilitation. Conceptually, the neurorehabilitation system should function as a propulsive and postural neuroprostheses, assisting or perturbing the propulsion, the balance, or a combination of both, to varying degrees, depending on experimental objectives or patient-specific needs.

Existing systems for compensating for impaired propulsion and balance rely on passive spring-supported, weighted mechanisms or closed loop force control systems that generate vertical forces on the torso level during stepping confined to a stepper (Nessler, J.A. et al, Robotic devices for studying movements after rodent spinal cord injury (A cosmetic device for student coordination), IEEE transactions on neural systems and organizational methods: a publication of the IEEE Engineering in medical and biological 13,497 506 (2005); frequency, M. et al, new integrated electromechanical weight support systems (A cosmetic weight support system), Transmission on balance systems and Biology 14 (IEEE 311, 321, 311)). However, these methods show several disadvantages: (i) current systems provide support only in the vertical direction, while well-balanced movements require finely tuned torso movements in virtually every direction (Winter, d.a., MacKinnon, c.d., Ruder, G.K. & Wieman, c., integrated EMG/biomechanical model of upper body balance and posture in human gait (integrated EMG/biometrical model of upper body balance and position), Prog Brain Res 97, 359-; (ii) visual flow that significantly regulates movement (Orlovsky, G.N., Deliagina, T.G. & Grillener, S., Neuronal control of movement from mollusk to human (from molar man), Oxford University Press, Oxford,1999), is inhibited during stepping confined to a stepper; (iii) rehabilitation is limited to stepping on a treadmill (Musselman, K., Brunton, K., Lam, T. & Yang, J., functional walking profile of Spinal cord injury: a new measure of walking ability, neural rehabilitation and neural replay 25,285-293 (2011)); this is a significantly different situation than the rich instruction set of natural motor tasks.

Robotic systems have been designed to overcome these limitations. ZeroG (Hidler, j. et al, ZeroG: above-ground gait and balance training system), Journal of rehabilitation research and maintenance 48,287 and 298(2011)) uses lifting devices mounted on track guided trolleys to provide vertical support during walking on the ground. However, the track constrains the subject along a fixed direction, and the torso support is limited to a vertical direction. Navigator (Shetty, d., Fast, a. & Campana, c.a., ambulant suspension and rehabilitation device (US7462138)) allows translation in all directions with an overhead linear multi-axis system, but its bulky structure generates high inertia, preventing movement at normal steps.

Therefore, there is a problem of obtaining a robot system that overcomes the drawbacks of the prior art. In particular, there is a need for a multidirectional torso support system that addresses these various problems.

Another problem in the art is that assessment of motor function in subjects typically relies on visual scoring systems (Basso, D.M. et al, MASCIS assessment of outdoor sport scores: the effect of experience and team combination on reliability, multicenter Animal Spinal Cord Injury studies (MASCIS evaluation of experiences and team on reliability, multicenter Animal Spinal Cord Injury Study), Journal of neural Injury 13,343-359(1996)) or univariate analysis (Basso, D.M. et al, MASCIS assessment of outdoor sport scores: the effect of experience and team on reliability. multicenter Spinal Cord Injury Study)B. Etc., and the analysis movement is recovered: comprehensive quantification of post-CNS injury damage in rodents (Profiling logistic: comprehensive quantification of injuries after CNS injury, Nature methods 7, 701-.

It is well known that activity-based interventions utilize proprioceptive information to improve spinal cord movement output during training (H.Barbeau, S.Rossinol, Recovery of movement after long-term myelotomy in adult cats (Recovery of cognitive rehabilitation in the patient cat), Brain Res 412,84(May 26,1987); R.G.Lovely, R.J.Gregor, R.R.Roy, V.R.Edgeton, effect of training on Recovery of full weight in adult spinal cats (Effects of tracking on the Recovery of full weight-bearing in the patient spot), Experimental neurology 92,421(May, 1986); fb.Werng, S.Muller, severe Recovery of movement in persons with weight injury (road exercise) while promoting severe Recovery of movement of spinal cord in individuals with weight injury (late walking), and improvement of spinal cord movement after walking, early rehabilitation, muller, improved walking supported using body weight Laufband locomotion in persons with severe spinal cord injury (Laufband coordination with body weight supported in people with segment spinal cord injuries), Paraplegia30,229(Apr, 1992); laufband therapy based on the "rule of spinal cord movement" is effective in humans with spinal cord injury (Laufband therapy based on 'rules of spinal cord coordination' ideal in spinal cord affected individuals), Eur J Neurosci 7,823(Apr 1, 1995).

Recent case studies have shown that activity-based rehabilitation, in combination with epidural electrical stimulation of the lumbosacral segment, can also restore spine-mediated locomotion after motor complete paraplegia (Harkema, S. et al, effects of epidural stimulation of the lumbosacral spinal cord on voluntary locomotion, stance and assisted stepping after motor complete paraplegia. case study (a case study), Lancet,377,1938(Jun 4,2011)).

There is combined evidence that gait Rehabilitation should be performed on the ground (Wessels, M., Lucas, C., Eriks, I. & de group, S., weight support gait training for Rehabilitation of walking in persons with incomplete Spinal cord injury: systematic review (Body weight-supported gait training), Journal of Rehabilitation medium: of firm joint office MS European bone of Physical and Rehabilitation Medicine 42, 513-with) walking across multiple paradigms (Musselmans, K., Lam, T.A. Yang, J., Spinal cord injury function support capacity (Brussels, 293, K.: walking measurement of walking support capacity), new weight support training for Rehabilitation of Spinal cord injury (Brussels, M.A.) (Body weight-supported gait training for Rehabilitation, S.)), walking training for Rehabilitation of walking with incomplete Spinal cord injury, S., walking training for Rehabilitation of walking, walking training, and walking training, etc. (branched walking training, 293, walking training, using proper walking training, 293, walking training for Rehabilitation, K.: white bone injury, K., walking training, K., walking, 293, and walking training, using proper measures, 293, walking training, walking training, for Rehabilitation, walking, for Rehabilitation, etc. (branched Spinal cord injury, K., walking training, similar training, lucas, c., Eriks, I. & de Groot, s., weight-bearing gait training for rehabilitation of walking in persons with incomplete spinal cord injury: systematic overview (body-supported gate tracking for restoration of walking in peptide with synthesized peptide in: a systematic review), Journal of reconstruction media: of the UEMS European Board of physical and reconstruction media 42,513-519 (2010); reinkensmeyer, D.J., et al, Tools for understanding and optimizing robotic gait training (Tools for understating and optimizing), Journal of rehabilitating research and evaluation 43,657-670 (2006); ada, l., Dean, c.m., Vargas, J. & Ennis, s., mechanically assisted walking with weight support in patients unable to walk early after stroke results in more independent walking than assisted ground walking: a systematic overview (mechanical lost with body weight support resources in more than one independent walking and walking in no-album matrix approaches: a systematic review), Journal of physiological therapy 56, 153-; harkema, s. et al, effect of epidural stimulation of the lumbosacral spinal cord on voluntary locomotion, stance and assisted stepping following locomotion with complete paraplegia: case study (Effect of anatomical simulation of the luminal aspect on volume movement, standing, and contained constructing after motor complete parplegia: a case study), Lancet377, 1938-; kwakkel, g., Kollen, B.J. & Krebs, h.i., effect of robotic adjuvant therapy on upper limb recovery after stroke: systematic overview (Effects of robot on upper limber recovery after spoke: a systematic review), neural toxicity and neural repair 22, 111-; edgerton, V.R. & Roy, r., Robotic training and spinal cord plasticity (robotics training and spinal cord plasticity), Brain research bulletin 78,4-12 (2009); reinkensmeyer, d.j., et al, Tools for understanding and optimizing the gait training of robots (Tools for understating and timing the gait training), Journal of rehabilitating research and evaluation 43, 657-; harkema, s. et al, effect of epidural stimulation of the lumbosacral spinal cord on voluntary locomotion, stance and assisted stepping following locomotion with complete paraplegia: case study (Effect of anatomical simulation of the anatomical specific on volume movement, standing, and assisted steering after motor completed Patient: a case study), Lance 377,1938-1947(2011)) and Patient active cooperation (Duschau-Widke, A., Capez, A. & Riener, R., control of Patient cooperation during robot-assisted gait training enhances active participation of SCI individuals (Patient-assisted cognitive training of individual identification with SCI reduced group-aided gait training), Journal of neurological engineering and regeneration 7,43 (2010); edgerton, V.R. & Roy, r., robotic training and spinal cord plasticity (robotically training and spinal cord plasticity), Brain research bulletin 78,4-12(2009)), but these concepts remain discrete and do not indicate how to obtain a unified therapeutic tool for assessing and restoring motor function following CNS disorders in both animals and humans.

Furthermore, according to the state of the art, the subject still cannot achieve autonomous control of the movement.

There remains a problem to provide a method for rehabilitation of subjects suffering from neuromuscular disorders, in particular partial or complete paralysis of the limbs, and to achieve autonomous control of movements.

There is also a need for a device that provides autonomous control for restoring motion in neuromotor impairments that can be used as both propulsive and postural neuroprostheses, assisting or perturbing propulsion, balance, or a combination of both, to varying degrees depending on experimental objectives or patient-specific needs. In particular, such a device should be able to perform an objective assessment of the motion function, capturing the multi-dimensional associative structure of the motion function. Furthermore, such a device should be able to direct the subject in need to regain autonomous control of movement, and should also be "transparent" to the subject as the case may be.

Summary of The Invention

It has now been found that the problems of the prior art are solved by the combination of a multidirectional torso support system with a device for epidural electrical stimulation.

Therefore, as defined in the claims, the object of the present invention is an apparatus for restoring voluntary control of locomotion in a subject suffering from neuromotor impairment, said apparatus comprising a multidirectional trunk support system and a device for epidural electrical stimulation.

Another object of the invention is a robotic interface capable of assessing, enabling and training the generation and balance of movement patterns in a variety of natural walking behaviors in a subject with neuromotor impairment as defined in the claims. Surprisingly, providing such a robotic interface with a device for epidural electrical stimulation and optionally with a pharmacologically active mixture, together with certain improvements of said robotic interface, results in a device for restoring voluntary control of movement in subjects suffering from neuromotor impairment, enabling rehabilitation results that are much higher than those of the devices of the prior art.

As defined in the claims, as a further object of the present invention, a method has also been found for evaluating, enabling and training subjects suffering from neuromotor impairments by combining robotically-assisted evaluation tools with sophisticated neurobiomechanical and statistical analyses. The method provides a means to fine-objectively assess the control of gait and balance and the interactions between them.

As an object of the present invention, a method has also been found for rehabilitation of a subject suffering from neuromotor impairment, in particular partial or complete paralysis of the limbs (which term also includes restoring voluntary control of movement), which method enables voluntary control of movement, comprising applying electrical and optionally pharmacological stimuli in an above-ground training program and using the above-mentioned robotic interface.

In an embodiment of the invention, in the apparatus, the multi-directional torso support system provides support for the subject against gravity.

In another embodiment of the invention, the multidirectional torso support system comprises:

a. a robotic interface having an end effector with n actuated degrees of freedom;

b. a mechanism integrated in or attached to the robotic interface for providing compliant/elastic or viscoelastic behavior in the degrees of freedom at an end effector of the robot;

c. a sensor for measuring the movement of the end effector resulting solely from such compliance; or sensors for measuring the force (torsion) generated by such compliant motion (compliant deformation);

d. an interface connected to the subject using the apparatus to transfer any twist in the degrees of freedom to the subject.

In another embodiment of the invention, the sensor is a position sensor or a force sensor.

In another embodiment of the invention, the multidirectional torso support system comprises:

i. a multidirectional elastic decoupling system; three motor-driven linear actuation modules having horizontal, orthogonal axes X and Y along an X, Y, Z cartesian coordinate system and a vertical axis Z, each defining 4 degrees of freedom, and one motor-driven rotary actuation module about the vertical axis Z; wherein the linear actuation modules are simultaneously decoupled by a suspension system having a compliant element oriented in each of the four degrees of freedom;

a parallel Delta motion system to prevent tilt.

Optionally, the device of the invention may be equipped with a robot leg.

Any type of position sensor (rotational or longitudinal) or force sensor may be used. In one embodiment of the invention, the sensor is selected from the group consisting of a contactless magnetic encoder, a potentiometer, and a laser. For the purposes of the present invention, it is contemplated that any type of suitable sensor may be used in accordance with the knowledge of one skilled in the art. For example, in the device, 4 contactless magnetic encoders are located at the junction of the Delta system.

According to another object of the invention, the device also comprises a computer which communicates with the module and acquires information from the encoder, and optionally exchanges information with a second computer running a user interface.

In an embodiment of the invention, in the device, the motor-driven actuation modules provide constant force modes independent of each other.

In an embodiment of the invention, in the apparatus, the motor-driven linear actuation modules along the horizontal, orthogonal axes X and Y and the motor-driven rotational actuation module around the vertical axis Z provide a transparent mode, and the motor-driven linear actuation modules along the vertical axis Z provide a constant force mode.

In another embodiment of the invention, in the device, a constant force mode can be used in all directions (mainly X, Y, Z), especially in a training mode.

In another embodiment of the invention, all modules can also be actuated in a variable force mode (e.g. gate phase-dependent support).

The device of the invention is used for rehabilitation (including rehabilitation of voluntary control of locomotion) of subjects suffering from neuromotor impairments selected, for example, from partial and complete paralysis of the limbs.

As is apparent from the above description, in the integrated concept of the present invention, on the basis of the combination of a multidirectional torso support system with a device for epidural electrical stimulation, a mixture comprising a combination of agonists of monoaminergic receptors may be used to improve the recovery of locomotor autonomic control in a subject in need of said apparatus. In this sense, another object of the present invention is a pharmaceutical composition for restoring voluntary control of locomotion in a subject suffering from a neuromotor disorder, comprising a combination of agonists of 5HT1A, 5HT2A/C, 5HT 7and DA 1-like receptors.

Another object of the present invention is a pharmaceutical composition for restoring voluntary locomotion in a subject suffering from a neuromotor disorder, comprising a combination of agonists of monoaminergic receptors, in particular serotonergic receptors, dopaminergic receptors and adrenergic receptors.

According to some embodiments of the invention, the neuromotor disorder is selected from the group consisting of spinal cord injury and stroke consequences.

Another object of the invention is a method of restoring voluntary control of locomotion in a subject suffering from a neuromotor disorder, the method comprising:

a. using the apparatus disclosed above;

b. providing electrical stimulation, in particular to a site of neuromotor pathology, more in particular to a site of myelopathology, and optionally administering a pharmaceutical composition comprising a combination of agonists of monoaminergic receptors as disclosed above.

In the context of the present invention, the above-described method does not intend that steps a) and b) must be performed one after the other, but rather that they are used according to the teachings of the present invention, in particular that the electrical stimulation using the device for epidural stimulation can be set at different times of the method, and that the device can be used alone even after epidural stimulation has stimulated spinal neurons and established traffic with the brain.

In an embodiment of the invention, the method for restoring voluntary control of movement further comprises providing a treadmill exercise to the subject prior to using the apparatus disclosed above and applying the epidural electrical stimulation.

Another object of the present invention is a method for operating the above disclosed apparatus, said method comprising the steps of:

a. an evaluation mode, in which the device provides support against gravity in a spring-like or gravity-reduced state, using a motor-driven actuation module along the vertical axis Z;

b. an energized mode, wherein the device provides propulsive and/or postural assistance using constant speed forward motion with the motor driven actuation modules along the horizontal axis X, while the motor driven actuation modules along the vertical axis Z provide constant force vertical support at a percentage of body weight, and the motor driven actuation modules along the horizontal axis Y and the motor driven rotational actuation modules around the vertical axis Z provide rigid support in the lateral direction;

c. a training mode wherein the device provides postural support against gravity with a motor-driven actuation module along the vertical axis Z, the motor-driven actuation module along the horizontal axis X being set to be transparent, the motor-driven rotation actuation module around the vertical axis Z being set to be rigid or transparent, the motor-driven actuation module along the horizontal axis Y being set to be rigid or transparent.

In an embodiment of the present invention, in the above method, Principal Component (PC) analysis is performed on a gait cycle.

Advantageously, the present invention provides a device that solves the problem of inertia that circumvents the cumbersome robotic structures of the prior art, and effectively solves the major problems associated with prior support systems, such as unidirectional torso support, high inertia, or stepping limited to treadmills.

In addition, the devices disclosed herein may provide an objective assessment of the complexity of gait and stepping to create motor function. The device may also provide a fine-tuned energizing and training program during rehabilitation.

Drawings

The invention will now be disclosed in detail in exemplary embodiments of the invention on experimental animals, also using the figures and examples. The system may be scaled up to humans.

In the attached drawings

Fig. 1 shows a perspective view of an exemplary embodiment of a robotic interface of the present invention. The degree of freedom (X, Y, Z,) Indicated by arrows. The subject to which the device is to be used is connected to the device by suitable means, for example a skin-like jacket attached to the back plate at torso height. The subject is also provided with a device for epidural electrical stimulation, said device being placed according to known methods.

FIG. 2 shows a detailed view of the multidirectional elastic decoupling system of an embodiment of the present invention.

Fig. 3A and 3B show the following bar graphs: bar graphs reporting the mean (n ═ 7 rats) 3D distance between conditions (distance of each rat from the mean of all gait cycles without using a robot) (a, top panel) and 3D dispersion (gait variability) (a, bottom panel); bar graph of PC analysis (B, bottom) of gait and 3D distance between conditions during movement along the ladder (B, top). a.u. arbitrary units. Error bar, s.e.m.; bar graphs of mean distance (C, top graph) and gait variability (C, bottom graph) from intact rats calculated by PC analysis (a.u. arbitrary units) in the robot interface in the assessment of pattern generation and balance are reported; bar graphs (D) of the average (n-5 rats) 3D distance from the pre-lesion trial in experiments using a robotic postural neuroprosthesis to enable skilled motor control after cortical stroke were reported (a marked difference from all pre-lesion conditions at p < 0.01); respectively related to the distance a from the lesion front), the percentage of steps accurately positioned on the step b) (white bar: step on, gray bar: toddlers), distance c) from the lesion, percentage of stepping d) accurately positioned on the step (white bar: step on, black bar: drop, grey bar: toddlers) (a.u. arbitrary units). Error bars, s.e.m.. x: significantly different from pre-lesion conditions at p < 0.01). This bar graph correlates the condition of significant difference at p <0.01 in experiments using a robotic postural neuroprosthetic to enable coordinated motion on steps after moderate and severe SCI (for all graphs: white: pre-lesion, grey: using robot, black: not using robot; dashed line: no stimulation); steering to restore equilibrium in rats with severe SCI in training experiments powered by robotic postural neuroprostheses, bar graphs (F) of the mean distance between each trajectory of motion and the optimal trajectory (left panel) and the maximum deviation of the pelvic segment from the vector of the heading (right side) (error bars, s.e.m.: significant differences from all other unlabeled conditions at p <0.01) were reported.

FIG. 4: is a technical description of a robot interface and control scheme.

Detailed Description

According to the general concept of the present invention, the goal of autonomous control of motion is made possible by the basic combination of a multidirectional torso support system and a device for epidural electrical stimulation. In principle, any type of known multidirectional torso support system and any type of device for epidural electrical stimulation are suitable for carrying out the present invention. The above description also provides details of certain embodiments which are intended to improve certain aspects of the present invention.

The multi-directional torso support system conveniently provides support for the subject against gravity.

In a preferred embodiment of the present invention, the multi-directional torso support system comprises a robotic interface comprising an end effector having n-actuated degrees of freedom; a mechanism integrated in or attached to the robotic interface for providing compliant/elastic or viscoelastic behavior in the degrees of freedom at an end effector of the robot; a sensor for measuring the movement of the end effector resulting solely from such compliance; and an interface connected to the object using the apparatus to transfer any twist in the degrees of freedom to the object.

According to the invention, the robot interface has at least 1, preferably at least 2, more preferably at least 3, even more preferably at least 4 degrees of freedom. Mechanisms that are integrated in or attached to the robotic interface and provide compliant/elastic or viscoelastic behavior in the degrees of freedom at the robot's end effector are well known in the art and need not be described specifically herein, as are the sensors and interfaces described above.

To address the problem of avoiding the inertia of bulky robot structures, the robot interface of the present invention is provided with a multidirectional elastic decoupling system (also known as a multidirectional torso support system) that provides transparency to the robot. Such a robotic interface effectively addresses major problems associated with existing support systems, such as one-way torso support, high inertia, or stepping limited to a stepper. The present invention provides means in the form of a robotic interface that assists or perturbs propulsion and balance continuously and independently along n, preferably 4, degrees of freedom (DoF) as an object using or assisted by the interface progresses over the ground in a large workspace. In particular, the invention provides said device as a means for rehabilitation of subjects suffering from an impairment of the motor system, in particular caused by a neuromotor impairment, in particular suffering from partial or complete paralysis.

In a first embodiment, the robotic interface is used for rehabilitation of a subject suffering from Spinal Cord Injury (SCI).

In a second embodiment, the robotic interface is used for rehabilitation of a subject suffering from stroke consequences.

Advantageously, the robotic interface is capable of assessing, enabling and training pattern generation and balance between walks under natural conditions encompassing a wide range of motor behaviors and advanced capabilities.

In one embodiment of the invention, the multidirectional torso support system comprises:

i. a multidirectional elastic decoupling system; three motor-driven linear actuation modules having horizontal, orthogonal axes X and Y along an X, Y, Z cartesian coordinate system and a vertical axis Z, each defining 4 degrees of freedom, and one motor-driven rotary actuation module about the vertical axis Z; wherein the linear actuation modules are simultaneously decoupled by a suspension system having compliant elements oriented in each of the 4 degrees of freedom;

a parallel Delta motion system to prevent tilt.

Referring now to FIG. 1, an exemplary embodiment of a robotic interface of the present invention comprises:

(1) a continuous robot module consisting of three translation axes defining a Cartesian coordinate system (x, y, z) and oneRotation axis () Is constituted and denoted by the general reference numeral (1);

(ii) a parallel Delta motion system that prevents tilting and allows the position of the object to be measured, and is denoted by the general reference numeral (2);

(iii) a suspension system having springs (fig. 2) oriented in each of the 4 dofs of the continuous structure to decouple the inertia of the cumbersome robotic structure from the end effector. This suspension system utilizes the high performance of a series of elastic actuators to achieve a behavioral transparent haptic device (Pratt, G.A. et al, Stiffness not all (Stiffness Isn't Everering), International symposium on Experimental Robotics (ISER) (Springer, Stanford, USA, 1995); Valley, H. et al, Compliant actuation of rehabilitation robots-the Benefits and limitations of series of elastic actuators (compliance actuation of rehabilitation robots-fibers actuation of series of elastic actuators), ie bot Autom Mag 15,60-69 (2008)).

The robotic interface of the present invention advantageously allows real-time control of body displacement (propulsion) and Body Weight Support (BWS) (balance) along 4 independent dofs, which can be continuously adjusted, i.e. from rigid position control to transparent zero force control.

In more detail and with reference to fig. 1, component (i) of the robotic system of the present invention is adapted to provide adjustable torso support along 4 independent degrees of freedom (DoF).

3 motor-driven linear actuation modules (3, 4, 5) are provided. These types of modules are commercially available, see, for example, CKK 20-145, CKK 15-110, and CKK12-90 (BoschRexroth AG), and define a large Cartesian workspace that enables objects to be translated in the direction X, Y, Z. The first two axes for movement in the horizontal plane (fig. 1, (X) and (Y)) cover a large area (6) that is estimated to be sufficient for the object using the interface. The third axis (fig. 1, (5, Z)) provides the object with resistance to gravitySupport and allow a sufficient range of vertical motion for rehabilitation purposes. At the end of the cartesian structure, a fourth motor (7) drives a shaft around a vertical axis (figure 1,) E.g. 300 degrees), of a commercially available model such as RE25, Maxon motor AG, Sachseln, Switzerland. This continuous configuration provides a large working space in which a force can be applied to the object while preventing tilting to the horizontal direction.

The assembly of 4 motor-driven modules may be securely supported by a suitably constructed frame (fig. 1, (8) only showing one support for module 4. for simplicity, the rest of the frame is not shown as they may be configured differently according to common general knowledge), wherein the motor-driven modules may translate along X, Y and the Z-axis. The frame may be provided with frame elements adapted to support the motor-driven modules and allow movement in their direction. For example, the modules (3), (4) and (5) may be provided with a frame in the form of a rail on which they are mounted in a conventional manner. A vertical structure is used to support the motor-driven module (5) in such a way that it can move along a vertical axis Z. The mounting of the three modules and the frame supporting them is conventional and within the ability of one of ordinary skill in the art.

The area (5) may be provided with different means for training a subject in need of rehabilitation, such as straight or different curvature roads, obstacles, ladders, steppers.

If desired, in order to provide a highly flexible robotic system capable of guiding objects along any desired trajectory, but which may also appear transparent, i.e. allowing the patient to walk freely throughout the workspace without "feeling" the robot, the interaction forces between the objects and the robot must be minimized. The inertia of the robot is significantly greater than the mass of the object in which it is used.

Generally, using conventional rigid force sensors and force controls, the inertia of the robot cannot be circumvented from the subject due to theoretical stability limitations of force control (Colgate, E. & Hogan, N., Analysis of Contact instability based on Passive Physical Equivalents, Proceedings-1989 Ieeeeerntial Conference on Robotics and Automation, Vol 1-3, 404-. Thus, a direct coupling between the robot and the object will generate significant interaction forces, which will disturb the natural motion of the object. In order to make significantly lighter interacting objects evade the inertia of the robot structure, Pratt, g.a. et al (rigid not all (stiff Isn't evolution), International Symposium on Experimental Robotics (ISER) (Springer, Stanford, USA,1995) propose to couple actuators to objects by means of compliant elements; this configuration is known as a Series Elastic Actuator (SEA). Furthermore, the interaction force and torque can be directly measured by monitoring the deformation of the compliant element. However, the concept of SEA has so far only been used for a single actuator, i.e. a single DoF.

In embodiments of the invention, in order to best use the SEA concept for the robotic interface of the invention, all 4 actuation modules need to be decoupled simultaneously, which requires that all deformable elements are as close to the object as possible.

It has been found (see fig. 2) that this problem is solved by providing a lightweight, low friction compliance module consisting of a base platform with three protruding legs to form a cage (10), a spring suspended platform (9) inside the cage and a Delta structure to constrain the un-actuated DoF (i.e. the tilt of the subject).

With reference to fig. 2, the suspended platform (9) is connected to the cage (10) by 6 linear springs (11, one coupling located behind the cage not shown) calibrated according to the weight of the subject to be treated (for example for small animals such as rats or mice the following settings can be adopted: angle in the horizontal plane, 120 degrees; stiffness, 112N/m for the upper springs and 57N/m for the lower springs). Another pair of springs (not shown) is attached to the central rotational axis of the suspension platform (9), providing elastic decoupling around the vertical axis. Taken together, this configuration decouples the inertia of successive modules from the suspended platform over 4 actuated dofs.

The Delta configuration (12) allows for measurement of the displacement of the suspended platform and thus the deflection of the springs along each DoF, providing an inexpensive way to measure the interaction force or torque.

For the measurement of the interaction force, any known means may be used. In one embodiment of the invention, 4 contactless magnetic encoders (sensors) (commercially available from, for example, 12-bit, Austria microsystems, Austria) are placed at the junction of the Delta structure. The position of the end effector relative to the continuous robot is calculated by combining the information from these angular sensors with a forward dynamical model of the Delta structure. The relative position of the platform encodes the spring length and thus the interaction forces and torques resulting from the linear spring characteristics.

These forces and torques are used in the force control loop of the robot. The control strategy was implemented in MATLAB/Simulink and executed in real time on a desktop computer running an xPC target (sampling frequency, 1 kHz). The computer is in communication with the motor drive of the actuator and acquires information from the sensor. It also exchanges information with a second computer running a user interface for changing the control parameters of the robot online.

The SEA-based elastic decoupling allows setting a very high control gain without affecting stability. Due to the use of multi-dimensional SEA, this inertia controls only the perceived dynamics for low frequency excitation with low inertial forces (valley, h. et al, Compliant drive of rehabilitation robots-Benefits and limitations of series elastic actuators, Ieee Robot Autom Mag 15,60-69 (2008)). For high frequency excitation, which is generally associated with low amplitude motion, the physical properties of the spring control the response, and low forces are also generated. Thus, the object primarily feels the inertia of the suspended platform.

Thus, the robotic interface combines the advantages of continuous motion (large workspace), parallel motion (low inertia), and a series of elastic actuations extending in multiple dimensions (compliant interactions). Taken together, this new robotic device provides real-time control of body translation (propulsion) and Body Weight Support (BWS) conditions (balance) along 4 independent dofs within a configurable environment.

With reference to fig. 4, the control of the robot interface is further disclosed.

User interface

A user-friendly GUI (graphical user interface) is implemented in MATLAB/Simulink (The MathWorks, CA) or other similar programs. The interface allows the user to create a virtual environment (shown as a "virtual world" in fig. 4) in which the applied force or end effector can be adjusted for each individually driven DoF of the robot. For example, the user may independently set any one of the 4 drive shafts to operate transparently. Meanwhile, the vertical axis provides a constant force proportional to the body weight of the subject for supporting the subject against gravity. The axes may also be constructed to be rigid in order to prevent lateral falls or to guide the object along a user-defined trajectory. Alternatively, the user may control the displacement of the end effector (position control) for pushing the object in a given direction or along a user-defined trajectory. Finally, the user can introduce abrupt changes (arbitrary twists) in the virtual environment. For example, user-defined perturbations may be superimposed on any control scheme based on external triggers or the location of objects in the real world. For example, the user may create a virtual environment for the subject for a straight road or a road containing at least one turn, or a road containing a length of irregularly spaced horizontal piles (struts), or a straight gait of constant velocity, or a straight road in which lateral movement is induced, or a road containing up and down stairs. The 4 motor-driven actuation modules can be set by the user in different modes: rigidity (100% constant force), transparency (not felt by the subject), constant force (%), and constant velocity.

Universal impedance control implementation

Referring to fig. 4, an impedance control scheme is implemented that can adjust the force independently applied by each driven DoF of the robotic interface in real time (1 kHz). The controller is cascaded: the external loop processes the location of the object relative to a virtual environment, such as a world having a guide wall or gravity-reduced condition. The algorithm transforms the user-defined virtual environment into vectors of required forces and torques

The force controller adjusts the desired motor speed q delivered to the module drive in 4 degrees of freedom (DOF) according to the error between the desired force and the force measured by the spring bias of the decoupling system_mot,des. Internal speed controller by commanding the appropriate actuator torque τ_motEnsuring true motor speed q_motFollowing the desired motor speed. The external loop runs on the Matlab xpc real-time operating system. The speed control runs on the actuator driver.

Robot

Cartesian positioning system: the robot consists of an actuated cartesian positioning system allowing the translation of the object in the horizontal plane (x, y) while providing vertical support (z). At the end effector of the continuous structure, another motor rotates (). This continuous configuration provides a large workspace in which forces can be applied to the object over 4 dofs.

Force module: in order to avoid the inertia of bulky positioning robots and to measure the minimal interaction forces between the robot and the object in which the robot is used, the invention provides a new "series elastic actuator" (SEA) based force module. The SEA consists of a series of actuators implemented by passive compliance elements. This compliance improves force control performance and effectively decouples actuator inertia to achieve a transparent interface. In the force module of the present invention, the SEA concept is extended to 4 dofs by providing multi-dimensional compliance at the end effector of the positioning system.

Motion constraint for undriven DoF: the mechanical "Delta" connection prevents the object from tilting on the 2 un-actuated DoFs, creating a restraining force F_c. Delta configuration also provides for measuring the end effector position (object position q)_sub) And subsequently measuring the interaction force F between the robot and the object_elSee the above equation, wherein in this case, F is used_elIn place of F_desAnd each variable is el instead of des.

Elastic decoupling of driven dofs: the compliance of the remaining dofs is,by a plurality of linear springs attached to the suspended platform and by another pair of springs attached to a rotating shaft within the platform.

Real world

The object is placed in a customized means for receiving the object, such as a harness or skin-like casing preferably made of lightweight fabric. Closures, such as hook and loop fasteners, allow for attachment of an object to a back plate with a rigid strip from an end effector of a robot. The position of the object and the interaction force with the robot are fed back to the impedance controller.

The motor ability of intact and motor-damaged objects can be evaluated, for example, in many tasks. a. Movement along a straight horizontal runway. b. Movement along a horizontal runway that is curved at 90 degrees. c. Movement of a linear horizontal ladder along rungs with irregular spacing. d. Motion along a straight horizontal runway, wherein the robot pushes the object forward at a constant speed. e. Lateral disturbances introduced during the continuous movement along the straight horizontal runway (task a). f. Continuous motion on a motor-driven treadmill belt. g. Moving along regularly spaced steps on the stairs. For each task, the degree of compliance is adjusted independently for each translational and rotational axis. The control strategy comprises the following steps: stiffness control, zero force control, adjustable constant force (constant force set to a certain percentage of body weight) and constant speed (position control).

The results of the exercises performed using the robotic interface of the present invention are elaborated using suitable statistical methods. In a representative embodiment performed on laboratory animals (rats), the experimental data sets are processed in a multi-step statistical analysis applicable to all experiments described herein. Step 1: for all experimental conditions, a recording system was used to collect movement, power and EMG data during continuous movement. Step 2: a large number of parameters are calculated to provide comprehensive quantification of gait characteristics. The analytical procedures and calculations are detailed in Courtine, G.et al, the Transformation of non-functional spinal loops into a functional state after loss of brain input (Transformation of non-functional spinal circuits after the loss of the brain input), Nat Neurosci12, 1333 + 1342 (2009); musienko, p. et al, in J Neurosci 31, 9264-. And step 3: we applied Principal Component (PC) analysis to all variables (n-144) calculated from all gait cycles from all rats and experimental conditions. The gait cycle is represented in the new 3D space generated by the 3 first PCs (accounting for variance, 39%). A least squares sphere is traced to highlight the overlap between the gait with and without use of the robot. This analysis constructs a new variable, PC, that linearly combines the original variables and maximizes the amount of interpretation variance for each successive PC. Due to the high correlation between gait parameters during movement, a few PCs are sufficient to account for most of the variance. And 4, step 4: the gait cycle can be represented in the new "de-noising" space generated by the PC 1-3. In the proposed embodiment, the data points associated with each experimental condition are clustered in well-defined positions, indicating that the rat exhibits an intervention-specific gait pattern. In general, PC1 strongly distinguishes the gait cycle from intact rats (or pre-lesion) from altered gait from rats with SCI or stroke and motion improvement using a robotic interface. In some cases, the PC2 captures other features. In the proposed embodiment, PC2 is involved in the specific characterization of the intervention as compared to intact and no intervention. To provide a straightforward representation of the differences between conditions, we applied a least squares ellipse fit to the 3D data points. And 5: to quantify the quality of gait expression, we measured the 3D geometric distance between the mean position of the gait cycle from each rat and the mean position of all gait cycles from all intact (or pre-lesioned) rats under the given conditions. For each rat and condition, we also measured (in au, arbitrary units) the 3D dispersion of gait cycles to provide a measure of gait variability. Step 6: the score (position of gait cycle in PC space) reveals which conditions are differentiated along each PC. And 7: we then extract the factor load, i.e. the correlation between each variable and each PC. We select the target PC according to step 6 and use the highest factorial load (| value | >0.5, p <0.05) to regroup the variables into functional clusters, which we name for clarity. The variables loaded on the same PC are related to each other. For example, in one embodiment, improvements in hindlimb motion are directly related to improved postural control. And 8: to provide a more classical representation of the differences between the conditions, we generate a histogram for one variable of each extracted functional cluster.

In a preferred embodiment of the invention, a motor-driven actuation module is used in a constant force mode, which results in improved motor performance compared to spring-like supports in rats with complete SCI.

Mode of operation

The robotic interface of the present invention can operate in three different modes: 1) an evaluation mode for evaluating motion pattern generation and balance; 2) an enabling mode for motor control of robot energization after a neuromotor injury; 3) a training mode for robot-enabled training, which latter mode may be used for example for rehabilitation of subjects with paralytic SCI.

1) Evaluation mode

With the benefit of constant force support, the robotic interface of the present invention is able to assess motion pattern generation and balance.

Most BWS systems rely on passive spring mechanisms that provide support against gravity proportional to the vertical position of the subject. Although a tailored motion configuration can achieve position-independent constant force support (Nessler, J.A., et al, A robotic device for studying motion after spinal cord injury in rodents (IEEE transactions on neural systems and mobility Engineering: a publication of the IEEE Engineering in Medicine and Biology Society 13,497-506(2005)), there is a problem that these passive systems do not compensate for rapid movement.

Advantageously, in this evaluation mode embodiment, the robotic system of the present invention can apply well-controlled arbitrary vertical force profiles that can simulate spring-like conditions or gravity-reduced environments. In fact, the constant force BWS of the present invention significantly improves the quality and consistency of gait characteristics and promotes a movement pattern that is more favorable to that of healthy subjects when compared to spring-like BWS.

The evaluation mode of the present invention provides heuristic conditions to evaluate motor pattern generation and balance after neuromotor impairment.

2) Enabling mode

According to the present invention, the robotic interface may be used as a propulsive and/or postural neural prosthesis, which provides adjustable assistance to propel the body forward and restore postural orientation and stability.

It is well known that electrical and pharmacological stimulation can produce motion in subjects with severe SCI, possibly in humans (Harkema et al, Lancet), but subjects cannot produce the necessary force to propel their body above ground. Instead, they show tonic activity in the extensor muscle, behaviorally as a stand. To compensate for the lack of propulsive force, the robotic interface of the present invention functions as a propulsive neural prosthesis that moves the subject forward at a constant speed while providing constant force vertical support as a percentage of body weight (e.g., 60 +/-10% of BWS) that is adjusted according to the needs of the subject and the rehabilitation program. When robot guidance is initiated, the subject transitions smoothly from quiet standing to continuous motion. When the pusher neural prosthesis stops translating the subject forward, the rhythmic movement stops immediately.

The energized mode will now be described in an exemplary embodiment on a laboratory animal.

Rats with unilateral left cortical stroke, when traversing a horizontal ladder with irregularly spaced rungs, exhibited a lesion contralateralSignificant impediment to paw placement (B. Etc., and the analysis movement is recovered: comprehensive quantification of lesions after CNS injury in rodents (Profiling lococontroovery: comprehensive quantification of injuries after animals CNS damagein variants), Nature methods 7,701-708 (2010)). In all trials from all rats not supported and supported with a constant force robot, the relative positioning of the lesion contralateral hind paw with respect to the two consecutive rung positions was evaluated. The assessment was made by a graphical decomposition of hindlimb movements during trials with and without the use of a robot. Hindlimb oscillations and EMG activity of TA and Sol muscles were recorded. PC analysis (accounting for variance, 28%) was performed to separate the exact step from the missing step to emphasize that the robot improved the percentage of exact steps, but had no effect on the motion strategy itself. The results (FIGS. 3A, A and B) show the mean 3D distance (.: at p) from the pre-lesion trial<0.01 significantly different from all pre-lesion conditions).

These impairments have been attributed to the loss of visual motor control of the motor cortex that is heavily dependent on the impairment (Drew, t., Andujar, j.e., Lajoie, K. & Yakovenko, s., Cortical mechanisms involved in visual motor coordination during precise walking (Cortical mechanisms involved in visual motor coordination during walking), Brain Res Rev 57,199-211 (2008)). Impaired balance maintenance may also contribute to changes in proficiency following cortical stroke. The robotic interface of the present invention functions as a postural neuroprosthesis.

In this energized mode embodiment, the robot provides constant force support in the vertical direction (z-axis, 27 ± 4% of BWS) and rigid support in the lateral direction (y and rotation axis). The robotic postural neuroprostheses immediately improved the ability of subjects to accurately position their diseased contralateral limb on the irregularly spaced rungs of the ladder. Statistical analysis shows that the robot significantly reduces the number of mistakes/slips, which correlates to improved postural stability.

Thus, the robot postural neuroprosthesis of the invention enables motion control in subjects with motion impairment, in particular caused by SCI or stroke.

Surprisingly, the energized pattern of the robotic interface immediately restores motor capabilities in a wide range of natural walking behaviors following moderate to severe neuromotor impairment.

3) Training mode

In an embodiment of the training mode, the robotic interface uses repeated exercises to improve functional capabilities. In this mode, the robotic postural neuroprosthesis provides support against gravity (z-axis), but in other directions (x, y andshaft) is made transparent. Locomotion is enabled by, for example, electrical stimulation and optionally pharmacological stimulation. The training mode of the robot interface significantly improves the motion capabilities. In one embodiment of the invention, such a robotic interface is suitable for training programs in subjects with paralytic movement disorders, such as SCI.

When used as a postural or propulsive neural prosthesis, the robotic interface of the present invention is immediately capable of providing unexpected motor capabilities in the affected subject.

There is a correlation between the multi-directional torso balance restored by the robot and improved lower limb motion control. This immediate functional improvement underscores the importance of extending the current torso support system, which in the prior art was only one-way, to multiple dimensions. Likewise, robotic exoskeletons that provide multidirectional support against gravity can improve upper limb recovery in stroke survivors (KWakkel, G., Kollen, B.J. & Krebs, H.I., impact of robotic-assisted therapy on post-stroke upper limb recovery: systematic reviews (Effects of robot-assisted thermal on upper limb recovery: systematic review), neurological and neural repair 22,111 repair 121(2008)) and improved movement in humans with partial SCI (Dual-wicker, A., Capez, Riener, R., cooperative Patient control during robot-assisted gait improves active participation of SCI individuals (tissue-assisted surgery, human-aided diagnosis of repair, joint repair of human tissue, joint repair of joint repair, joint repair of repair, joint repair and repair of repair, joint repair, and repair, joint repair, and repair, joint repair, and repair).

The robotic postural neuroprostheses of the present invention not only provide multidirectional trunk support, but also restore limb and trunk orientation. As a result, the extension and load-related afferent inflow from the hip and ankle joints (Pearson, K.G., producing a walking gait: the role of sensory feedback (Generating the walking gap: roll of sensory feedback), Prog Brain Res 143, 123-. This emphasizes the restoration of key sensory feedback and task-specific regulation thereof, significantly contributing to reestablishing gait control. For example, a robotic postural neuroprostheses may be able to improve hip extension during stair climbing compared to horizontal motion. This information appears to be sufficient to modulate the increased step height and accurate placement of the feet on the stairs. Likewise, side-dependent modulation of the receptors from the load and stretch sensitivity of the ankle and torso muscles during curvilinear walking results in an asymmetric force pattern that maintains balanced steering. To this end, the interface of the present invention is conveniently equipped with sensors to measure force. These sensorimotor processes are improved by training. Taken together, these findings confirm and extend the current view of the ability of sensory information to influence the lost function as a source of motor control on spines (Courtine, G. et al, Transformation of non-functional spinal circuits into a functional state after loss of brain input, Nat Neurosci12, 1333-. In this regard, the interface of the present invention may be equipped with Robotic legs (exoskeletons) attached to the lower extremities (Nessler, J.A., et al, Robotic devices for studying movement after spinal cord injury in rodents (A Robotic device for student coordination after spinal cord injury), IEEE transactions on neural systems and rehabilitation Engineering: application of the IEEE Engineering in Medicine and Biology Society 13,497-506(2005)), to ensure proper task-specific sensory feedback during rehabilitation (Edgerton, V.R. & Roy, R.R., Robotic training and spinal cord plasticity), Brain research bolus 78,4-12 (2009)).

In another aspect, the invention relates to a method for restoring voluntary control of movement in neuromotor impairments, for example after spinal cord injury causing paralysis, and for rehabilitation of subjects suffering from neuromuscular disorders, in particular partial or complete paralysis of the limbs, such method enabling voluntary control of movement comprising applying electrical and optionally pharmacological stimuli and using the above-mentioned robotic interface in an above-ground training program.

In a preferred embodiment, the method of the invention comprises a first step of stepping machine exercise, and a second step of above-ground training using the robotic interface of the invention in combination with electrical stimulation, optionally in combination with pharmacological stimulation.

It is important to note that in the training mode, the subject may gain sufficient motion control so that electrical stimulation may be abandoned and assistance provided only with the robotic interface.

PC analysis (accounting for variance, 48%) was applied to all gait cycles and rats. A least squares fit and exponentiation was performed independently for each rat. The average of the scores on PC1 for gait cycles was recorded in intact and spinal rats using the same level of spring-like versus constant force vertical direction support stepping. The parameters with the highest factorial load (| value | >0.5, p <0.05) on PC1 were regrouped into functional clusters. For intact rats and spinal rats using spring-like versus constant force vertical support stepping, the average of one variable per functional cluster was calculated.

In rats with complete SCI, the effect of loading conditions on motor pattern generation was evaluated. Rats received complete SCI. After 5 weeks of recovery, rats received energizing factors to encourage biped movement on the treadmill (13 cm. s)^-1). 10 gait cycles were recorded for each constant force BWS level (40-90%). Exercise was recorded in healthy rats at 60% BWS, the level being the weight of the hind limb normally loaded during a four-footed gait. For each BWS level and for intact rats, a representative graphical map decomposition of hindlimb movement during stance, trawling and swing was obtained. The trajectory of the hind limb endpoint is traced together with the orientation and intensity of the foot velocity vector at the beginning of swing. The average vertical ground reaction force (combined left and right hind limbs) and the relative duration of the stance, swing and tow phases of gait are determined. The relationship between the level of BWS and the degree of similarity of gait pattern compared to healthy rats was measured as the 3D distance from the gait cycle in the PC analysis. A second order polynomial fit is applied to the data points to highlight the U-shaped relationship between step quality and BWS level. Will have the highest factorial load (| value! non-woven on PC1>0.5，p<0.05) into functional clusters. An average of one variable per functional cluster at different BWS levels is obtained.

The robot-propelled neural prosthesis of the invention enables coordinated above-ground movement of the spinal rat. Spinal rats were placed on both feet in a robotic interface. The robot is configured to operate at a constant speed (13 cm-s)^–1) The body is moved forward while providing constant force vertical support. In the graphical decomposition of the hindlimb movement and limb endpoint trajectories, the traces show the angular oscillations of both hindlimbs. To enable hindlimb locomotion, rats received stressors epidural electrical stimulation at spinal segments S1 and L2, in combination with agonists of 5HT1A, 5HT2A/C, 5HT 7and DA 1-like receptors. Using these stimuli, spinal rats showed tonic activity in the left and right extensors and were able to stand for longer periods of time. In the front of the robotWhen translated to replace the lost propulsive ability, the animal immediately exhibited a coordinated plantar step that rotated between the two hind limbs.

Using the robotic interface of the invention, improved balance control using a postural neuroprostheses during movement along a ladder with irregularly spaced rungs in rats with cortical stroke is associated with improved hindlimb movement and performance. The combination of the device of the invention, i.e. the robotic interface, with a device for epidural electrical stimulation and a pharmaceutical composition comprising a mixture of a combination of 5HT1A, 5HT2A/C, 5HT 7and DA 1-like receptor agonists, provides improved balance control, which is associated with improved hindlimb movement and performance during movement, e.g. along a ladder with irregularly spaced ladder stirrups, in rats with cortical stroke.

The PC analysis was applied to all gait cycles recorded along the ladder in all rats, before and 2 days after the lesion, with and without vertical constant force support. Both accurate and faulty steps are included in this analysis, but are not distinguished in the figures to emphasize the contrast between conditions with and without the use of the robot. The average of the scores on PC1 was obtained. The variables with the highest factorial load (| value | >0.5, p <0.05) on PC1 were regrouped into functional clusters.

Improved balance control using postural neuroprostheses using the robotic interface of the present invention was associated with improved hindlimb locomotion during straight horizontal runway locomotion in rats with moderate and severe SCI.

Graphical depictions of hindlimb movement, hindlimb oscillations and EMG activity of the SoL and TA muscles were recorded before pathology and 10 days after lateral cervical spine (C7) hemisection with and without constant force robotic support. PC analysis was applied to all gait cycles recorded in all rats before and 10 days after the lesion with and without robotic support. Hindlimb movement and EMG activity of MG and TA muscles before pathology and 12 days after staggered lateral hemisection were recorded without and with constant force robotic support, without energizing factor (no stimulation) and with stimulation. PC analysis was applied to all gait cycles recorded in all rats before and 10 days after the lesion without stimulation and with and without robotic support. The mean of the 3D distances between the different experimental conditions and the mean position of the pre-lesion gait in PC space were calculated. PC1 distinguishes between actual stepping and paralysis, while PC2 highlights the improvement in movement when using postural neuroprostheses.

Using the robotic interface of the invention, improved balance control using postural neuroprostheses was associated with improved hindlimb locomotion during locomotion on stairs in rats with moderate SCI (lateral cervical spine (C7) hemisection).

Experiments, evaluations and results analysis were performed as disclosed above.

In a similar manner, the robot interface of the present invention shows improved balance control using a postural neuroprosthesis, associated with improved hindlimb locomotion during locomotion on stairs in rats with severe SCI (staggered lateral hemisection).

Experiments, evaluations and results analysis were performed as disclosed above.

A method of restoring voluntary control of locomotion in a subject suffering from a neuromotor impairment, e.g. a disorder selected from the group consisting of spinal cord injury and stroke consequences, will now be disclosed in detail.

Using the controls of the robotic interface of the present invention, the X-axis (forward direction) is typically arranged to be transparent and the Z-axis is arranged to provide a constant force proportional to the subject's body weight. Transverse (Y) and rotation () The shaft maintains rigidity to prevent lateral falls. For certain experiments and training, the robot may move the torso of the subject forward at a constant speed. As a result, the limbs move backwards and the hip joint angular extension increases, thereby creating a similar shape to stepping on a treadmillThe method is described. While performed above ground, these stepping movements are still involuntary.

Training consists of a combination of 4 different paradigms, broadly divided into 3 stages specifically tailored to the performance and training objectives of the subject. To be able to enter a highly functional locomotor state, subjects may optionally receive a monoamine agonist 10min prior to training and a bi-site EES during the entire course of treatment. And (1). The primary purpose of the early training phase is to optimize the functionality of the lumbosacral loop. The subject undergoes treadmill-based training using vertical directional support. Sensory input induced by the moving treadmill belt acts as a source of control over the stepping of the limb. Manual assistance is provided in an on-demand assistance manner to present appropriate sensory cues to the lumbosacral circuit. At the end of each session, the subject is placed in the robotic gesture interface and encouraged to walk towards the target in front of it. The robot is configured to establish the most moderate side and vertical weight support. To provide background information about the task required, the robot translates the object forward at a constant speed. The objective is to force the brain to regain on-spine control of the electrochemically energized lumbosacral loop. And (2). The duration of the above-ground movement is gradually increased as the subject gradually regains the ability to produce an autonomous step. The objective is to encourage repeated and quantitative activation of the lumbosacral circuit by newly formed intra-and supraspinal connections. However, training confined to treadmills is still performed daily in order to mill spinal motion circuits over a consistent period of time to maintain their functionality. And (3). When the subject regains a steady above-ground hind limb movement, a complex task is introduced that requires fine adjustment of the hind limb movement, i.e. climbing stairs and avoiding obstacles. The objective is to facilitate improved supraspinal contributions in order to restore qualitative control over the electrochemically energized lumbosacral loop.

Together with the trajectory of the hind limb endpoint, a graphical decomposition of the hind limb motion is generated. The hind limbs are defined as virtual sections that connect the pelvis to the foot. The vector representing the direction and intensity of the hindlimb endpoint velocity at the onset of swing is used to assess the progress of rehabilitation. A multi-step statistical analysis of athletic performance and control strategies is performed. Step 1: advanced recording of hind limb movements during bipedal ground movements. Step 2: a number of variables are calculated that provide a comprehensive quantification of gait. And step 3: principal Component (PC) analysis was applied to all variables and recorded gait cycles. And 4, step 4: the single gait cycle is then represented in the new "de-noising" space generated by the PC 1-3. A least squares ellipse fit is used to facilitate visualization of differences between subsequent rehabilitation steps. And 5: the athletic performance was quantified as the 3D euclidean distance between the gait cycle position and the average position over all gait cycles. Step 6: the scores indicate the training periods distinguished by each PC. And 7: extraction of the factor load (i.e. the correlation between each gait variable and each PC) is performed. And 8: the highest factorial load (| value | >0.5, p <0.05) was used to classify variables into functional Clusters (CL) PC1 and revealed recovery of voluntary movements in subjects trained on the ground, resulting from strong synergy between ankle extension, torso extension and hip flexion, as well as improved interphalangeal cooperation, increased weight bearing capacity, improved lateral foot movement and near normal hind limb endpoint trajectory control. PC2 indicates that the subject of treadmill training exhibits a highly stable posture, but cannot initiate forward motion. Furthermore, the above-ground trained subject exhibits improved lateral body movement during exercise that alternately loads the left and right hind limbs and thus helps maintain dynamic balance. PC3 highlights the flexion posture and slow hindlimb movement of the subject in sub-acute states.

Detailed description of the inventiona combination of multidirectional torso support and a device for epidural electrical stimulation is essential to the apparatus for restoring autonomic control of motion of the present invention.

The following examples further illustrate the invention. It will be apparent to those skilled in the art that well-known technical changes may be made to the robots of the exemplary embodiments without departing from the teachings of the present invention, and in particular the functional concepts and methods illustrated herein.

Example 1

General procedure

Animal and animal care

All procedures and procedures were approved by the zurich veterinary administration of Switzerland (witinerian OfficeZurich, Switzerland). The experiments were performed on adult female Lewis rats (200 g body weight, Centre d' elevoge r. janvier, France). Animals were housed individually in 12h light/dark cycles with food and water ad libitum.

Surgical procedure and postoperative care

All procedures have been described in detail previously (Courtine, G. et al, Transformation of non-functional spinal loops into a functional state after loss of brain input (Transformation of non-functional spinal circuits into the functional state of spinal inputs), Nat Neurosci12, 1333-. Surgical intervention was performed under general anesthesia and sterile conditions. Rats underwent two surgical interventions. Bipolar intramuscular EMG electrodes (AS 632; cooper Wire, Chatsworth, CA) are first implanted into their selected hind limb muscles (Courtine, g. et al, non-functional spinal circuits transform into a functional state after loss of brain input (Transformation of non-functional circuits internal functional states of brain input), Nat Neurosci12, 1333-1342 (2009)). For some experiments, the electrodes were also secured at spinal level L2 and S1 at the midline of the spinal cord by suturing a lead (identical to an EMG lead) over and under the dura (Courtine, 2009). Rats were allowed to recover for 2 weeks after implantation. In the course of the lesion, the disease is preceded byAfter completion of the recording, the rats underwent a second surgical intervention during which they received SCI or stroke. SCI includes a complete transection of the thoracic (T7) spinal cord (Courtine, 2009), a right cervical (C7) lateral hemisection (Courtine,2008), or two lateral hemisections (Courtine2008) on opposite sides and at different spinal heights (T7 and T10). Ischemic lesions of the cortex (stroke) were induced by injection of vasoconstrictive endothelin-1 (ET-1, 0.3 μ g · μ l-1; Sigma-Aldrich) at 14 sites in the left motor cortex (forelimb and hindlimb areas). We used 6nl · s^-1At a depth of 1.2mm, a volume of 500nl was injected. After each injection, the needle was left in place for 3min and then carefully removed (B. Etc., and the analysis movement is recovered: overall quantification of damage following CNS injury in rodents (Profiling loomotor recovery: comprehensive quantification of injuries after CNS damage in rodents), Nature methods 7,701-708 (2010)). The extent and location of the lesions was confirmed at necropsy. The complete transected SCI was checked visually. The extent of chest and neck hemisection was measured on 40- μm thick transverse sections incubated in serum with anti-GFAP (1:1000, Dako, USA) antibodies. We measured the extent of transverse lesions at 5 equally spaced locations on the dorsal-ventral surface of the spinal cord. These values are expressed as a percentage of the total medial length and averaged to obtain a uniform measure of lesion size. The half-cut SCI ranged from 49.8% to 54% (50.8 +/-0.48%). In addition, a quantitative check was performed to ensure that the lesions met the following specific criteria: (i) minimal lesion ipsilateral spinal cord residual, defined as the absence of white matter residual, (ii) minimal damage to the lesion contralateral spinal cord, defined as near complete integrity of dorsal and ventral white matter tracts.

Exercise task

A total of 7 exercise tasks were used in this experiment: on a moving treadmill belt (13 cm. s)^–1) The feet of the robot, walking along the feet of the linear track, walking along the four feet of the linear track, and traversing during walking along the four feet of the linear trackTo a disturbance, walking along the four feet of an irregularly spaced smooth rung, climbing the four feet on a staircase, and turning along the four feet of a 90 degree curved runway. The attachment of rats to the back plate was different in different tasks and between various different types of injury. For bipedal exercise, the rat wears an upper body coat that extends from behind the neck to the iliac crest. The back panel is attached to the entire extent of the garment by hook and loop fasteners. For quadruped sports, the rat wears a one-piece garment, which exhibits two attachment points, namely at the pelvis or mid-chest height. The location of the backplate attachment is selected according to the particular gait impairment exhibited by the rat. Typically, the robot is attached to the pelvis when the rat exhibits changes in hindlimb motor control, while a mid-thoracic attachment is selected when the rat exhibits balance impairment.

Behavioral training in rats

When rats first wear the full body garment, they exhibit a change in gait pattern. Thus, rats were adapted to navigate freely along the runway for 1-2 weeks while wearing a custom-made coat. When no significant difference (p) is observed between the movements of wearing and not wearing the garment>0.1), we trained animals 1 or 2 courses per day until they crossed the runway at a constant speed. Positive reinforcement (food reward) was used to encourage the rats to perform the required tasks. Rats were trained on ladders with a regular array of rungs. For trials, the rung sequences were irregular and varied to avoid adaptation to a particular rung pattern (B. Etc., and the analysis movement is recovered: overall quantification of damage following CNS injury in rodents (Profiling loomotor recovery: comprehensive quantification of injuries after injury of animals CNS in rodentions), Nature methods 7,701-708 (2010).

Motion control enabling factor

To facilitate locomotion in paralyzed rats, we used a mixture of epidural electrical stimulation and a monoamine agonist (Musienko, p. et al, by spinal cord)Multidimensional monoamines in the medullary loop can regulate and control specific motor behaviors (Controlling specific logic or biochemical strategies of spinal circuits, J Neurosci 31,9264-9278 (2011)). Rectangular pulses (0.2ms duration) were sent at 40Hz using two constant current stimulators (AM-Systems, WA, USA) connected to L2 and SI electrodes. The stimulation intensity (50-200. mu.A) was adjusted to visually obtain the best promotion of stepping. The rats also received 5HT1A/7(8-OH-DPAT, 0.05-0.1 mg. multidot.Kg^–1) 5HT2A/C (quinazidine, 0.2-0.3 mg. Kg)^–1) And SKF-81297(0.15-0.2 mg. Kg)^–1) Systemic administration of the agonist of (a).

Test protocol

Typically, 10 stepping cycles (treadmill) or 10 trials (runway) are recorded per rat under the given experimental conditions. The conditions with and without the use of a robot were randomly assigned among rats. During walking with and without the use of a robot, rats wear body garments to maintain the same experimental conditions for both types of recordings. When using electrical and pharmacological stimulation to promote exercise, stepping was recorded about 10min after drug injection.

Neural rehabilitation training

Rats underwent a training session of 30min 6 days per week, starting 12d after injury. They were trained for 7 weeks. It can be moved by electric stimulation and pharmacological stimulation. During each training session, rats practice quadruped movements along a horizontal linear runway, on stairs, and along a 90 degree runway. We adjusted the respective duration of each task according to the current ability of the animal. For example, rats performed only a few walks along a curved runway during each training session before the animals began to show a return in balance control at weeks 4-5.

Kinematic, kinetic and EMG recording

And (4) kinematics. 3-D video recording (200Hz) was performed using a motion capture system (Vicon, Oxford, UK). The motion of reflective markers bilaterally attached at the distal end of the scapula, iliac crest, greater trochanter of femur (hip), lateral condyle (knee), lateral malleolus (ankle), fifth Metatarsal (MTP) and toe tip was tracked using 12 infrared T10 cameras. Nexus (Vicon, Oxford, UK) was used to obtain the 3D coordinates of the markers. The body is modeled as an interconnected chain of rigid segments and joint angles are generated accordingly. The principal axis of the limb is defined as the virtual line connecting the greater trochanter to the lateral malleolus.

An EMG. EMG signals (2kHz) are amplified, filtered (10-1000 Hz bandpass), stored and analyzed offline to calculate the amplitude, duration and timing of individual bursts (Courtine, G. et al. non-functional spinal circuits are transformed into a functional state after loss of brain input (Transformation of non-functional spinal circuits into functional states, Nat Neurosci12, 1333-1342 (2009)). To assess temporal coordination between muscles, we generated a probability density distribution of normalized EMG amplitudes of the agonistic and antagonistic muscles as previously described (Courtine, g, et al, non-functional spinal circuits transformed into a functional state after loss of brain input (Transformation of non-functional spinal phases after the loss of brain input), Nat Neurosci12, 1333- > 1342 (2009)).

Kinetics. The ground reaction torque and ground reaction force in the vertical, fore-aft, and medial directions were monitored using a force plate (2kHz, HE6X6, AMTI, USA) located under the treadmill belt or in the middle of the runway.

Data analysis

For each experimental condition and each rat, a minimum of 10 stepping cycles were extracted for both the left and right hind limbs. According to the methods described in detail previously (Courtine, G. et al, the non-functional spinal loops are transformed into a functional state after loss of brain input (Transformation of non-functional spinal circuits inter functional state of gait input), Natneurosci 12,1333-, a total of 148 parameters for kinetic and EMG characteristics. These parameters provide a comprehensive quantification of movement patterns from the overall characteristics of gait and performance to the fine details of limb movements.

Statistical analysis

Various experimental conditions were associated with substantial modulation of gait patterns, as is evident in the modification of a large proportion of the calculated parameters. To evaluate the more important and reproducible modulation patterns mediated by different conditions and the correlation between the modulated parameters, we performed a multi-step statistical procedure based on Principal Component (PC) analysis (Courtine, g. et al, Transformation of non-functional spinal circuits into a functional state after loss of brain input (Transformation of non-functional spinal functional states after the loss of brain input), Nat Neurosci12, 1333-1342 (2009)). PC analysis was applied to data from all individual gait cycles of all rats taken together. The data were analyzed using a correlation method that adjusted the mean of the data to zero and the standard deviation to 1. This is a conservative approach for variables with different variances (e.g. kinematics versus EMG data).

Processing the data to mean ± s.e.m. The difference between data from normal distributions for each experimental condition was tested using repeated measures ANOVA and Student's paired t-test (Kolmogorov-Smirnov test). When the distribution is not normal, a nonparametric test (Wilcoxon and Kruskall Wallis) is used instead.

Robot interface

A robotic system was constructed that provided rats with adjustable torso support along 4 independent degrees of freedom (DoF). The three linear actuation modules CKK 20-145, CKK 15-110 and CKK12-90(Bosch Rexroth AG, distributor: Amsler AG, Feuerhalen, Switzerland) were aligned to define a large Cartesian workspace that enables the rat to translate in the x, y, z directions. The first two axes for movement in the horizontal plane (see fig. 1, x and y for reference) cover 1.2m²The area of (a). The third axis (see, for reference, fig. 1, z) provides support against gravity for the rat and allows vertical movement within a range of 35 cm. At the end of the cartesian structure, a fourth motor (RE25, Maxon motor AG, Sachseln, Switzerland) drives a rotation about a vertical axis (for reference, see figure 1,) Rotation (300 degrees). This continuous configuration provides a large working space in which force can be applied to the rat while preventing tilting about the horizontal direction.

In order for the robot system to behave transparently (i.e. to allow the rat to walk freely throughout the workspace without "feeling" the robot), the interaction forces between the object and the robot must be minimized. The inertia of the robot (106 kg in x-direction, 32kg in y-direction, 29kg in z-direction) is significantly higher than the mass of the rat (<0.25 kg).

A lightweight, low friction (<10g), compliant module is provided, consisting of a base platform with three protruding legs to form a cage, a spring suspended platform within the cage, and a Delta structure to constrain un-actuated DoF (i.e., tilt of the subject) (see fig. 2 for reference). The suspended platform was connected to the cage by 6 linear springs (angle 120 degrees in the horizontal plane; stiffness, 112N/m for the upper springs and 57N/m for the lower springs) (see FIG. 2 for reference). Another pair of springs is attached to the rotating shaft in the center of the suspended platform, providing elastic decoupling about the vertical axis. Taken together, this configuration decouples the inertia of successive modules from the suspended platform over 4 actuated dofs.

The Delta configuration allows for measurement of the displacement of the suspended platform and thus the deflection of the springs along each DoF, providing an inexpensive way to measure the interaction force or torque. 4 contactless magnetic encoders (12-bit, Austria microsystems, Austria) were placed at the junction of the Delta structure. The position of the end effector relative to the continuous robot is calculated by combining the information from these angular sensors with a model of the forward motion of the Delta structure. The relative position of the platform encodes the spring length and thus the interaction forces and torques resulting from the linear spring characteristics.

These forces and torques are used in the force control loop of the robot (see fig. 4 for reference). The control strategy was implemented in MATLAB/Simulink and executed in real time on a desktop computer running an xPC target (sampling frequency, 1 kHz). The computer is in communication with the motor drive and acquires information from the sensors. It also exchanges information with a second computer running a user interface for changing the control parameters of the robot online.

The SEA-based elastic decoupling allows setting a very high control gain without affecting stability. The resulting reflected mass of the rigid robot is: 787g in the x-direction, 104g in the y-direction, 22g in the z-direction, 998g cm in the rotational direction^-2. Due to the use of multi-dimensional SEA, this inertia only controls the perceived dynamics for low frequency excitations where the inertial forces are low. For high frequency excitation, which is generally associated with low amplitude motion, the response is controlled by the physical properties of the spring, and low forces are also generated. Thus, the rat predominantly felt the inertia of the suspended platform, which was 109.1 g. The bandwidth of the SEA system is 2.5Hz in the x-direction, 2.8Hz in the y-direction, 13Hz in the z-direction, and 2.2Hz in the rotation.

To confirm the transparency of the robot, we compared the kinematics and muscular activity implied under the motion of healthy rats (n-7) walking along a straight course with and without the robot. The results were evaluated by a graphical decomposition of hindlimb movements during stance and swing, together with EMG activity of limb end trajectories, hindlimb joint angles and Medial Gastrocnemius (MG) and Tibialis Anterior (TA) during locomotion along a linear runway without and with robotic support. Despite detailed analysis, we did not detect significant differences between these conditions (p >0.3, fig. 3A, fig. a), indicating that the cumbersome robot did not interfere with gait. We confirmed these results during walking on a horizontal ladder (n-5). Even under such challenging conditions, the precise paw placement (p >0.4) and gait characteristics are virtually unaffected by the robot interface (p >0.3, fig. 3A, fig. B).

Evaluation mode

The objective of this experiment was to compare the effect of prior art spring-like on the generation of movement patterns compared to constant force BWS conditions in rats with complete SCI (n-5). Rats received complete SCI causing permanent hind limb paralysis.

To enable stepping, we applied a combination of epidural electrical stimulation and monoamine agonists (Courtine, g., et al, Transformation of non-functional spinal circuits into a functional state after loss of brain input (Transformation of non-functional spinal circuits into functional states of brain input), Nat Neurosci12, 1333-channels 1342 (2009)). We evaluated performance by elaborating a graphical breakdown of hindlimb motion compared to constant force BWS using a spring-like, combined with subsequent limb endpoint trajectories (n ═ 10 steps), motion of TA and MG muscles and vertical ground reaction forces. We adjusted the spring-constant force to the optimum to promote stepping (Courtine, G., et al, conversion of non-functional spinal cord circuits to a functional state after loss of brain input (Transformation of non-functional spinal circuits of the loss of brain input), Nat Neurosci12, 1333-1342(2009)) and maintain the same amount of support throughout the constant force conditions. Constant force BWS significantly improves the quality and consistency of gait characteristics (p <0.01) and promotes motor patterns towards healthy rats (p <0.01, see fig. 3A, panel C) compared to spring-like BWS.

Human (Harkema, S. et al, epidural stimulation of the lumbosacral spinal cord after locomotor complete paraplegia versus voluntary locomotor, standing and assistedInfluence of stepping assistance: case studies (Effect of anatomical translation of the luminal spinal cord on luminal movement, standing, and associated stabilizing after motor complete parplegia: a case study, Lancet,377,1938(Jun 4,2011)) and rats (Courtine, G. et al) with conversion of non-functional spinal loops to functional states after loss of brain input (Transformation of non-functional spinal loops into functional states of brain input), Nat Neurosci12, 1343 (2009); timoszyk, w.k. et al, Hindlimb loading determines the number and quality of steps following spinal transection (hindimbimb loading determinations and quality influencing spinal cord transformation), Brain Res 1050,180-189(2005)) lumbosacral spinal cord can account for weight bearing information during steps. We evaluated whether weight input also determines gait quality in rats with complete SCI (n-4). Lowering BWS levels results in a graduated adjustment of hind limb kinematics, power and muscle activity (p)<0.01) that demonstrates the ability of the lumbosacral circuit to translate weight bearing information into specific motion patterns. However, we have found an inverted U-shaped relationship (R) between gait mass and BWS level²＝0.87)。

These findings confirm that optimal constant force support conditions can be used for enabling and training exercises in subjects with gait disorders.

In rats, unilateral cortical stroke has limited effect on basic locomotion, but behavioral observations indicate a defect in balance control (ii) a deficiency in balance controlB. Etc., and the analysis movement is recovered: overall quantification of damage following CNS injury in rodents (Profiling loomotor recovery: comprehensive quantification of injuries after injury of animals CNS in rodentions), Nature methods 7,701-708 (2010). To verify the balance impairment after stroke, we utilized the robot's ability to superimpose any force on the transparent control mode at any time and any driving DoF. Specifically, as the rat proceeded freely along the straight runway, we applied a sudden triangular force of 1s in the medial-lateral direction (y-axis, pushing right)(2.5N, 1s, right during runway movement in rats with left cortical stroke). Shortly after left stroke (6d), the rat was unable to compensate for the disturbance. They exhibit a large rightward deviation (p)<0.002, mean transverse trunk displacement before, during and after the perturbation (n ═ 5 rats)), and frequent runouts (56 ± 39%, mean ± s.d.). After one month of recovery, rats responded to the perturbation by using controlled co-activation of the extensors and flexors followed by long-term activity of the contralateral extensor with lesions (360 +/-80%, p)<0.001). This muscle synergy stabilizes the trunk and hind limbs and produces a significant medial-lateral force (p) that restores the motion trajectory<0.001, compared to 0.60 ± 0.07N at 6d post-lesion, 1.54 ± 0.18N at 30d post-lesion).

Taken together, these results demonstrate that the assessment mode of the robotic interface provides heuristic conditions to assess motor pattern generation and balance after neuromotor impairment.

Enabling mode

Next, we sought to utilize the robotic interface as a propulsive and postural neuroprostheses that provided adjustable assistance to propel the body forward and restore postural orientation and stability. We show that this so-called energized mode will reveal unexpected motor abilities masked by propulsion and/or balance impairments.

Electrical and pharmacological stimulation are able to produce movement in rats with complete SCI, but animals are unable to produce the necessary force to propel their body forward above ground. Instead, they exhibit tonic activity in the extensor muscle, behaving as a stand. To compensate for the lack of propulsive force, we constructed a robot that acted as a propulsive neuroprosthesis that moved the rat forward (x-axis, 13cm · s)^–1) While providing constant force vertical support (e.g., 60 +/-10% of BWS). When the robot sequence was initiated, the rat smoothly transitioned from resting to continuous motion. When the pro-neural prosthesis stopped to translate the rat forward, the rhythmic movement stopped immediately.

Rats with unilateral cortical stroke, when traversing horizontal ladders, show significant obstruction of lesion-contralateral paw placement (ii)B. Etc., and the analysis movement is recovered: overall quantification of damage following CNS injury in rodents (Profiling loomotor recovery: comprehensive quantification of injuries after injury of animals CNS in rodentions), Nature methods 7,701-708 (2010). These defects have been attributed to the loss of visual motor control that is heavily dependent on the injured motor cortex (Drew, t., Andujar, j.e., Lajoie, K).&Yakovenko, S., cortical mechanisms involved in coordination of visual movement during precise walking (cortical mechanisms involved in visual motor coordination stimulating walking), Brain Res Rev 57,199-211 (2008)). We tested the hypothesis that impaired balance maintenance may also contribute to changes in proficiency in locomotion following cortical stroke. We constructed a robotic interface as a postural neuroprosthesis. In this energized mode, the robot provides constant force support in the vertical direction (z-axis, 27 ± 4% of BWS) and rigid support in the lateral direction (y and rotation axis). Robotic postural neuroprostheses immediately improved the ability of rats to accurately position their lesion contralateral hind paw on the irregularly spaced rungs of the ladder (p)<0.002). Statistical analysis shows that the robot significantly reduces the number of errors/slips (p)<0.01, fig. 3A, fig. D), which correlates with improved postural stability (p)<0.01)。

Next, we evaluated the ability of the robotic postural neuroprostheses to enable motion control in rats with a transversal C7 hemisection (n-5). A graphical breakdown of hind limb movements during stair climbing in front of the lesion is recorded, showing hind limb oscillations and EMG activity of MG and TA muscles. PC analysis was applied to all gaits and rats. 10 days after the lesion, the rat drags the same hind limb of the lesion during movement, particularly during climbing stairs, without and with constant force robot support. Without robot support, they stumble on stairs and rarely step up stairs. The robot postural neural prostheses immediately enable them to coordinate footsteps during level walking (32 ± 4% of BWS) and stair climbing (28 ± 3% of BWS). The robot support restores torso orientation and stability (p <0.001), which is related to near normal hind limb motion and placement of the diseased co-lateral paw on the stairs (p <0.001, fig. 3B, fig. E, left).

We then investigated whether the robotic postural neuroprostheses are capable of motion control shortly after the more severe SCI consisting of two transverse half cuts at opposite sides and different vertebral levels (T7 and T10). This SCI completely interrupts direct supraspinal input, causing permanent hind limb paralysis (e.s. rosenzweig et al, Extensive spontaneous plasticity of corticospinal spinal projections after spinal cord injury in primates (Extensive spinal surgery efficacy of corticospinal surgery), Nat Neurosci 13,1505(Dec, 2010)). To enable exercise as early as 12 days after SCI, we applied electrical and pharmacological stimulation. At 12d after the staggered half cut, the movement of the animal on stairs was tested without and with the constant force robot support. Exercise was tested without (spontaneous) and with electrical and pharmacological stimulation. Without robotic support, rats exhibited rhythmic hind limb movement, but they were unable to footstep (91 ± 7% drag) during walking and generally fell laterally. Using a robotic postural neuroprosthesis, all rats tested (n-5) showed bilateral weight bearing footsteps. Despite the interruption of the direct supraspinal approach, rats immediately regained the ability to place both hind paws exactly on the stairs (p < 0.001). Otherwise paralyzed rats exhibited gait patterns that were almost indistinguishable from healthy rats both during horizontal movement and during stair climbing (fig. 3B, panel E, right). For both tasks, improvement of hindlimb motion is associated with restoration of robot-energized torso position and stability.

Taken together, these findings confirm that the energized pattern of the robot interface immediately restores unexpected motor abilities in a wide range of natural walking behaviors after moderate to severe neuromotor impairment.

Training mode

Finally, we use the enabling mode of the robot interface to improve functional capabilities using repeated exercises; this control scheme is referred to as a training mode. We performed a 30-min exercise training session every other day for rats with staggered half-cut SCI (n-6) for 8 weeks (see methods). The rat quadruped was placed in a robotic interface that provided constant force vertical support (z-axis) against gravity, but in other directions (x, y and y)Shaft) is provided to be transparent. Rats walked along a 90 degree curved runway. The torso orientation is measured as the angle between the pelvis and the orientation of the upper body velocity vector, which is called the heading and also defines the trajectory of the movement. Locomotion is enabled by electrical and pharmacological stimulation. At 9 weeks post-lesion, untrained rats exhibited weight-bearing stepping, but they were unable to control body inertia and balance (p) during robot-assisted movements along a curved runway<0.001, fig. 3B, fig. F, left). In contrast, the trained rats were able to curve around (fig. 3B, panel F, right) while maintaining balanced torso movement (p)<0.001)。

These results reveal that the training mode of the robot interface significantly improves locomotor ability in rats with paralytic SCI.

Example 2

Materials and methods

Animal and behavioral training

The experiments were performed on adult female Lewis rats (200-220g body weight) individually housed in a 12 hour light-dark cycle with food and water ad libitum. All experimental procedures were approved by the Veterinary administration of Canton of Zurich, Zurich. Prior to surgery, all rats (untrained and trained) were first adapted to navigate freely along the runway for 1-2 weeks while wearing a custom made coat. The rats were then trained to walk on both feet for an additional 1-2 weeks. All rats quickly learned this task. Typically, they produce a consistent stepping pattern over 1-2 treatment sessions. Positive reinforcement (food reward) was used to encourage the rats to perform the required tasks.

Surgical procedure

All basic surgical procedures and post-operative care for SCI rats have been described in detail previously (r.g. lovely, r.j. gregor, r.r.roy, v.r.edgeton, training the Effect on recovery of full weight stepping in adult spinal cats (Effects of tracking on the recovery of full weight-bearing steering in the adaptive spinal cat), experimental physiology 92,421(May, 1986); a.werign, s.murr, improved walking using the weighted lafband and motor support in persons with severe spinal cord injury (lau-and-bearing steering in patients with severe spinal cord injury), mapping 30,229 (pacific 30,229, pacifying of cases of spinal cord exercise, acute spinal cord exercise, lumbar spinal cord exercise, and spinal cord exercise assistance in the study of spinal cord exercise and lumbar spinal cord injury, lancet377,1938(Jun 4,2011)). Briefly, bipolar EMG electrodes were inserted into hind limb muscles under general anesthesia and sterile conditions. Two stimulation electrodes were secured to the dura at the midline of the vertebral levels L2 and S1. After pre-lesion recording, rats received a left lateral cross-dissection at T7and a right lateral dissection at T10 (court, g. et al, Recovery of spinal control of spinal index repair after spinal cord injury, Nature Medicine2008), on spine control of stepping through indirect spinal intrinsic relay junctions Recovery after spinal cord injury. For the T7 overcut, our purpose is to interrupt the bilateral dorsal column on the opposite side while leaving the ventral access. The completeness of the half-cuts was assessed on 30- μm thick longitudinal sections incubated in serum with anti-GFAP (1:1000, Dako, USA) antibodies. In addition, we demonstrated on cross-sectional slices the absence of BDA-labeled corticospinal axons in the dorsal column of the T8 spinal segment.

Multi-system neural prosthesis training

Rats received systemic (I.P.) administration of quinazidine (5-HT2A/C, 0.2-0.3mg/kg), SKF-82197(D1, 0.1-0.2mg/kg) and 8-OH-DPAT (5-HT1A/7, 0.05-0.2mg/kg) 10min prior to training. During training, we send unipolar electrical stimulation (0.2ms, 100-. The exercise training was performed on a treadmill (9cm/s) while using a vertical robot support, and with feet on the ground while using a robot pose interface. The content of each training session evolves with the actual capacity and training objectives of the rat. Positive reinforcement was used to encourage rats to perform the required tasks. Another group of rats was trained using the same frequency and length of time, but rehabilitation was limited to stepping training on a treadmill. Two weeks prior to the lesion, these rats were trained to walk on both feet above the ground using a robotic postural interface. They were also tested in this paradigm at weeks 1 and 9 post-lesion. At the end of the training session, treadmill-trained rats practice ground movement for about 10min using a robotic postural interface over a 4-8 session per day to ensure that the specificity of the task is not the reason they are unable to initiate and sustain movement.

Kinematics, kinetics, EMG recording and analysis

Biped movements were recorded on a treadmill (9cm/s) and on the ground. Kinematic (12 cameras, 200Hz), kinetic (force plate, 2kHz) and EMG (2kHz, 10-1000 Hz bandpass) recordings were made using an integrated motion capture system. Procedures for data collection, data analysis and calculation have been described in detail previously (Courtine, G. et al, Transformation of non-functional spinal circuits into a functional state after loss of brain input (Transformation of non-functional spinal circuits after the loss of the brain input), Nature Neuroscience 2009). To quantify athletic performance, we applied Principal Component (PC) analysis (Courtine, g., etc., to all calculated variables, and transformed the non-functional spinal circuits into a functional state after loss of brain input (Transformation of non-functional spinal circuits over the loss of brain input), Nature Neuroscience 2009), providing a step-by-step explanation of procedures and comments. We quantified the recovery of motor function as the distance between the gait cycles of intact and injured rats in the 3D space generated by PC1-3 (m. hagglund, l. borgius, k.j. dougherty, o. kiehn, Activation of groups of excitatory neurons in the mammalian spinal cord or hindbrain evokes movement), Nat Neurosci 13,246(Feb, 2010).

Brain stimulation and recording

The monopolar electrode was implanted epidurally on the left hind limb motor cortex. In the fully awake state, stimulation bursts (0.2ms, 10ms pulse length, 300Hz, 0.5-1.5mA) were delivered during bipedal stance. The test was performed without and with electrochemical stimulation. Peak-to-peak amplitude and latency of evoked responses were calculated from EMG recordings of the left TA muscle.

Neuronal modulation

60-70 days after injury, a micro-wire array (16 or 32 channels) was stereotactically implanted into the V-layer of the hindlimb region of the left motor cortex. Recordings were made 5-7 days post-surgery. Neuronal signals (24.4kHz) were acquired using a neurophysiological workstation synchronized with the motion recordings. All spike sorting was performed off-line by superparamagnetic clustering (j. liu, l.m. jordan, Stimulation of the parapyramidal regions of the neonatal rat brainstem produces motor-like activity involving the spinal cord 5-HT7and 5-HT2A receptors (Stimulation of the paragammal area of the neural rat brain receptor 5-HT7and 5-HT2 arereactors), Journal of neural science 94,1392(Aug, 2005)). Groups were manually adjusted according to established principles (g. court et al, do experiments in non-human primates accelerate the transformation of treatments for spinal cord injury in humans. Modulation was analyzed during a single experimental session to avoid potential confounding factors of instability. Two recurrent behaviors were used to assess the importance of neuronal modulation. (i) Triggering is defined as rocking from rest. (ii) Correction is defined as the onset of the swing phase after an irregular gait. The two-sample Kolmogorov-Smirnov test compares the excitation rates (evaluated in a 250ms window) over successive 1 second time periods covering both initiation and correction to determine if modulation is important.

Microinjection of NMDA and muscol

To ablate the T8-T9 neurons, we perfused NMDA (1% in dH2O) in 14 sites (1 mm deep, 3. mu.l total volume) covering the height of the vertebrae T8-T9. Rats were tested 5 days after the lesion and sacrificed the following day. Ablation of neurons was verified on histological sections stained with mouse anti-NeuN (1:500, Chemicon, USA) antibodies at necropsy. To inactivate the motor cortex, we injected the GABA agonist muscimol (800nl, 4.5mg/Kg) intradermally. On the 5 days before the experiment, we stereotactically implanted a catheter (OD:0.61mm, ID:0.28mm) into the left motor cortex at a depth of 1.5 mm. Correct catheter position was verified at necropsy on tissue sections stained (Invitrogen, USA) for fluorescent Nissl visualization.

Tracing and immunohistochemistry

We performed retrograde nerve fiber bundle tracing by bilateral infusion of fastplus (2%, in 0.1M phosphate buffer and 2% dimethyl sulfoxide) in the L1-L2 spinal segment (Courtine, g. et al, spinal on spine control to restore stepping through indirect spinal intrinsic relay junctions after spinal cord injury (Recovery of subasplant control of compressing via induced proprioceptive relay behind spinal cord 2008), Nature Medicine 2008). A total of 1.2. mu.l of pressure injection was performed at 6 sites (depth 1.5 mm). To trace motor cortical axons, we injected the antegrade tracer BDA 10,000 (10% in 0.01M PBS) into the left motor cortexCovering 6 sites of the hindlimb area (coordinate centers located-1 mm antero-posteriorly and-1.75 mm medially from the forehalogen, depth 1.5 mm). After 18 days, rats were treated with a solution containing 100,000IU/L heparin and 0.25% NaNO₂And then with 4% phosphate buffered paraformaldehyde containing 5% sucrose at pH 7.4. For cfos experiments, rats were perfused 60min after cessation of a 45min session of continuous movement (r.g. lovely, r.j.gregor, r.r.roy, v.r.edgerton, effect of training on recovery of full weight stepping in adult spinal cats (Effects of training on the recovery of full-weight-bearing in the same spot), Experimental neurology 92,421(May, 1986)). During the above-ground guided exercise, the exercise was performed on the ground for intact rats and ground-trained rats, as well as for treadmill-trained and untrained rats, in order to ensure that there was stepping in all animals. Brains, brainstems and spinal cords were dissected out, post-fixed overnight, and transferred to 30% phosphate buffered sucrose for cryopreservation. After 4 days, the tissues were embedded and sectioned at 40- μm thickness on a cryostat.

For immunohistochemical experiments, sections were incubated in serum containing rabbit anti-cfos (1:2000, Santa Cruz Biotechnologies, USA), anti-GFAP (1:1000, Dako, USA) or anti-5 HT (1:5000, Sigma Aldrich, Germany) or mouse anti-synaptophysin (1:1000, Millipore, USA) antibodies. Using Alexa488 or 555 labeled secondary antibody visualizes the immune response. BDA-labeled fibers were detected using streptavidin-horseradish peroxidase (1:200) in 0.1M PBS-Triton (1%). For signal amplification of pyrazinamide signal, anthocyanin 3 was used at a dilution of 1:100 for 1 min.

Neuromorphic evaluation

Fastplus and cfos positive neurons were counted on 5 evenly spaced sections 1.2mm apart and centered at the T8-T9 junction using image analysis software. Fiber density was measured using a stack of 5 confocal images of each region of each rat taken at standard imaging settings and analyzed using user-written scripts according to previously described methods (l.t. alto et al, chemo-guided stimulation of axonal regeneration and synapse formation following spinal cord injury (Chemotropic disorders axonal regeneration and synapse formation after spinal cord injury), Nat Neurosci12,1106(Sep, 2009)). The confocal output image is divided into square target Regions (ROIs) and the density within each ROI is calculated as the ratio of tracing fibres (amount of pixels) per ROI region. And carrying out color filtering and binarization operation on the file by using the intensity threshold value. The threshold was set empirically and maintained between different slices, animals and groups. A comparison of the computer and manual counts for the CST markers in T8-T9 shows no difference between the two methods. Manual fiber counting was performed on spinal cord sections covered with 5 vertical lines. The fibers in the gray matter that crossed these lines were marked and all crossed fibers on 3 sections of each rat were added to obtain cumulative counts. Manual and computer counting were performed without knowledge. Image acquisition is performed using a confocal laser scanning microscope and LASAF interface, and the image stack is processed off-line.

Statistics of

All data are reported as mean ± s.e.m. Statistical evaluation was performed using one-way or two-way ANOVA, repeated measures ANOVA, or non-parametric Wilcoxon test. The post Kruskall-Wallis test was applied where appropriate. Adult rats received a left lateral hemisection at T7and a right lateral hemisection at T10. This SCI interrupts all direct supraspinal pathways, but leaves a central gap of intact tissue. However, the lesions caused complete loss of hind limb function with no signs of recovery after 2 months post injury. Similarly, humans with clinically complete SCI often show maintenance of ligation by lesions (b.a. kakulas, a neurological review of Spinal Cord injuries in humans emphasizing particular features (a review of the neuropathology of human Spinal Cord with clinical on specific features), J Spinal rod Med 22,119(Summer, 1999)). Thus, this experimental pathology reproduces the key anatomical and functional features of human SCI, while providing well-controlled conditions to investigate the mechanisms underlying Recovery (Courtine, g. et al, Recovery of supraspinal control of stepping through indirect spinal intrinsic relay connections after spinal cord injury, Nature media 14,69(Jan, 2008)).

To convert the lumbosacral circuits from a latent to a highly functional state (p. musienko, j. heutschi, L. friedli, r. v. den branch, g. court, Multi-system neurorehabilitation strategy for restoring motor function after severe spinal cord injury (Multi-system neuroregenerative stimuli following), Experimental neurology, (Sep 7,2011)), we apply a tonic (40Hz) epidural electrical stimulation (count, g. et al), after loss of brain input, the non-functional spinal circuits are converted to a functional state (Transformation of non-functional spinal functional circuits), and administer a natural therapeutic system (HT, 365) neural stimulation (HT 2009), n. et al), and the spinal column is used to restore motor function after severe spinal cord injury, and the spinal cord is used to restore motor function after loss of brain input, and the spinal cord circuits are converted to a functional state (t. HT, g. et al), and the spinal cord circuits are administered to a functional state (HT, HT 5, HT, r. tm. fig. the spinal cord is used to restore motor function of the spinal system to restore spinal system after loss of spinal system, and to restore the spinal system to restore function of spinal system, e.g. spinal system to restore the spinal_1A/7、5HT_2A/CAnd D₁Customized mixtures of receptor agonists (p. musienko et al, control of specific motor behaviors through multidimensional monoamine energy modulation of the spinal cord circuit (Controlling specific local nutritional organisms with systemic modulators of spinal circulation), jneuroci 31,9264(Jun 22,2011)). By increasing the overall level of spinal excitability, this electrochemical spinal nerve prosthesis enables sensory information to become a source of control for stepping (Courtine, g., etc., conversion of non-functional spinal circuits into a functional state after loss of brain input (Transformation of non-functional spinal circuits), Nat Neurosci12,1333 (Oct, 2009); p.musienko, j.heutschi, l.friedli, r.v. den branch, g.courtine, Multi-system nerve rehabilitation strategies for restoring motor function after severe spinal cord injury (Multi-system nerve stimulation) to bone marrow surgery, experiment, (Sep 7,20, Sep. 20, etc.)11)). This intervention promotes coordinated bipedal stepping on the treadmill as early as 7 days after the injury, albeit involuntary.

These stepping movements are initiated by moving treadmill belts (Courtine, g., et al, Transformation of non-functional spinal loops into a functional state after loss of brain input, Nat neural circuits of the heart, 12,1333(Oct,2009)), suggesting that rats cannot autonomously initiate hindlimb movements on the ground. To confirm the absence of control over the spine, we applied an electrochemical neuroprosthesis and placed the same rat biped in a robotic postural interface that provided adjustable vertical and lateral torso support, but did not promote motion in any direction. All rats (n ═ 27) were unable to initiate hindlimb movements in the ground 7 days after injury (p < 0.001).

Then, we designed a multi-system neuroprostheological training program that covers two objectives. First, we aimed to improve the functionality of the lumbosacral circuit by treadmill-based training enabled by electrochemical neuroprostheses (Courtine, g. et al, spinal cord circuits that are not functional after loss of brain input are transformed into a functional state (Transformation of non-functional spinal circuits after the loss of brain input), Nat Neurosci12,1333 (Oct, 2009)). Second, we attempt to promote the recovery of spine-mediated movement; we utilized a robotic postural interface that not only energized but also forced rats to actively use their paralyzed hind limbs to move toward the target biped.

Rats (n 10) were trained with a combination of the two paradigms for 30min each day starting 7-8 days after injury. The first effort of self-stepping occurred 2-3 weeks after training (p < 0.01). With return to voluntary movements, we gradually increased the relative duration of above-ground training. All rats were able to initiate and sustain full-weight biped motion for a longer period of time 5-6 weeks after injury, but only during the electrochemically-energized kinetic state. Kinematic analysis revealed that the above-ground trained rats deployed a similar control strategy to that of intact animals to generate motion. To measure recovery, we adapted a clinically standardized 6-minute walk test (G.H. Guyatt et al, 6-minute walk: a new measure of motor capacity in patients with chronic heart failure (The 6-minute walk: a newmeasure of extrinsic capacity in patients with chronic heart failure), CanMed Assoc J132,919 (Apr 15,1985)) to bipedal rats. Above ground trained animals with paralytic SCI cover distances up to 21m within 3 min.

Next, we tested whether stepping training confined to a stepper under electrochemically energized conditions also promoted recovery of voluntary movements (n ═ 7 rats). Although repeated trials were performed in 4-8 sessions at 9 weeks after injury, this automated stepping training failed to reestablish ground motion (p < 0.001). Furthermore, treadmill trained rats were unable to maintain robot-initiated ground movement.

To further enhance the supraspinal contribution, we introduced two conditions that require self-mediated fine-tuning of gait, stair and obstacle (t.drew, j.e.andujar, k.lajoie, s.yakovenko, Cortical mechanisms involved in the coordination of visual movements during precise walking (Cortical mechanisms in visual motor coordination during walking), Brain ResRev 57,199(Jan, 2008)). After another 2-3 weeks, the ground-trained rats (pre-treadmill warm-up for at least 10min before robot training) were able to sprint both feet up the stairs and avoid the obstacle. To accomplish these paradigms, animals showed task-specific adjustments of hindlimb locomotion.

Anatomical examination highlights the extensive remodeling projected on and within the spine in rats regaining voluntary movement. We first performed retrograde nerve fiber bundle tracing from the L1-L2 motor center. We found that the number of labeled neurons in the medial and ventral layers of the T8-T9 segment was significantly increased (p) compared to untrained animals in both ground-trained and treadmill-trained rats (p)<0.01). Analysis of the activity-dependent marker cfos after continuous aboveground locomotion confirmed that the labeled neurons were active during walking. In rats trained on the groundCfos in regions enriched in neurons retrolabeled from the L1-L2 motor centers compared to all other groups^onThe number of nuclei is larger (p)<0.05). Thus, thoracic neurons may play a critical role in restoring autonomic movement (school, g., et al, spinal control of stepping through indirect spinal intrinsic relay connections after spinal cord injury (Recovery of superior control of stepping vitamin index association back) Nature media 14,69(Jan, 2008); F.M.Bareyre et al, spinal cord injury in adult rats spontaneously forms a new internal circuit (The input spinal cord complex for a new inter-spinal circuit), Nature neural 7,269 (2004); K.C.Cowley, E.Zaporonzhends, B.J.J.T., Mariots. spinal nerve signal for neonatal rat, Marulosonal nerve propagation of spinal sensory command for spinal cord injury (Journal of spinal chord 32)). To address this hypothesis, we ablated T8-T9 neurons by infusion of the axon-retaining excitotoxin N-methyl-D-aspartate (NMDA) (Courtine, g., et al, Recovery of subasplanchnic control of beating vitamin index repair after spinal cord injury, Nature media 14,69(Jan,2008)) to restore control over the spine of stepping by indirect spinal cord-intrinsic relay junctions following spinal cord injury. While the functionality of the lumbosacral circuit is intact, NMDA infusion abolishes regained locomotion (p) by regaining locomotion<0.01). Likewise, above-ground trained rats lose voluntary control of locomotion (p) following complete interruption of input to the spines of the T8-T9 neurons<0.01)。

We labeled projections from the left hind limb motor cortex by infusion of biotinylated dextran samine (BDA). Bilateral interruption of dorsal column at T7, which was half-cut, left only a small amount (1-2%) of axons in the right dorsal cord (c.brosamle, m.e. schwab, Cells of origin, processes and termination patterns of ventral non-crossing components of the cortical spinal cord fascicles of mature rats (Cells of origin, nerve, and termination patterns of the vetebral, uncrossed component of the maturated spinal cord tract), J Comp Neurol 386,293(Sep 22,1997)) cortical spinal cord fascicles (CST). Therefore, untrained rats showed rare CST labeling in segments T8-T9. Training confined to stepper does not promote significant changes in the projected intensity of the thoracic CST. In contrast, we found 45 ± 7% reconstitution of the pre-lesional bilateral fiber density in above ground trained rats. These CST axons are exclusively branched off from the dorsal-right cord and profoundly innervate the right side of the T8-T9 segment and more unexpectedly their left gray matter (e.s. rosenzweig et al, Extensive spontaneous plasticity of corticospinal projections after primate spinal cord injury, Nat Neurosci 13,1505(Dec, 2010)). We detected a number of CST fibers extending from gray matter at the T7 lesion into the right dorsal cord. These indicate that regenerating ectopic fibers (O.Steward, B.ZHEN, M.Tessier-Lavigne, false reactivation: differentiating regenerated from residual axons in the injured central nervous system (falserrecovery: distinguishing regenerated from residual axons in the injected central nervous system), J Comp Neurol 459,1(Apr 21,2003)), cause a nearly two-fold increase in CST axon density of the T8-T9 dorsal cord (p < 0.001). Chest CST fibers were half cut around T7 by the right dorsal cord, branched into gray matter, and re-crossed over the midline. These fibers developed large axonal structures with club-head-like bulges, indicating growth of tree-like ends. Confocal microscopy confirmed that breast CST fibers carry synaptic elements as they co-localize with synaptophysin. These fibers establish contact with relay neurons that are retrolabeled from the L1-L2 centers of motion.

Remodeling of the motor cortex axon projection is not limited to the remaining tissue bridge. Quantification of CST fibers at T4-T5 above the injury revealed a significant bilateral increase in axon density (p <0.01) in the above-ground trained rats compared to untrained, treadmill-trained and intact rats. We found a nearly 4-fold increase in the density of cortical projections in the various brainstem motor regions, including the left and right atrial nuclei (p <0.01), the overall reticular architecture (p <0.001), and the parapyramidal regions (p < 0.01). These regions contain reticular spinal cord neurons and spinal cord-projected serotonergic (5HT) neurons, both of which contribute to the initiation and maintenance of movement (M.hagglund, L.Borgius, K.J.Dougherty, O.Kiehn, Activation of groups of excitatory neurons in the mammalian spinal or hindbrain evokes movement (Activation of groups of excitation neurons in the mammalian spinal cord or hindbrain tissue uptake), Nat Neurosci 13,246(Feb, 2010); Stimulation of parapyramidal regions of J.Liu, L.M.Jordan, neonatal rat brain stem produces movement-like activity involving spinal5-HT 2 and 5-2A receptors (Stimulation of the paravertebral region of the spinal5-HT7and 5-3-neural receptor HT 54, HT 5-calcium receptor). Thus, by training, it is possible to reorganize the descending 5HT fibers. We found near complete layer-specific recovery of T8-T9 serum innervation in above-ground trained rats, as opposed to 5HT fiber depletion in untrained and treadmill-trained animals (p < 0.05).

Taken together, these analyses demonstrate that automated treadmill-limited training cannot mediate anatomical changes in the descending branch pathway, whereas active training in a highly functional state promotes multi-level plasticity in the cortical and brainstem-derived axonal system.

In contrast to primates, rodent motor cortex is not essential for generating movement (g. courtine et al, Can experiments in non-human primates accelerate the transformation of treatments for spinal cord injury in humans. Therefore, we attempted to demonstrate that training-induced remodeling of the motor cortex projection does contribute to controlling voluntary movements. First, we implanted a stimulating epidural electrode above the left motor cortex to verify that reorganization of the neuronal pathways reestablished connectivity across the lesion. Prior to SCI, application of a series of low intensity (0.7-1.5mA) electrical stimuli evokes a large response in the left tibialis anterior. SCI permanently abolished such response in untrained rats (p < 0.001). In contrast, the above-ground trained rats regained responses below the lesions, averaging about 10% of their pre-lesion amplitude (p < 0.001). These responses were delayed by 12 ± 3ms (p <0.01), indicating that a large number of synaptic relays were necessary to send the supraspinal stroke to the hindlimb motor pool. The magnitude of the response is significantly increased in the electrochemically-enabled motor state (p <0.01), indicating an improvement in command transmission on the spine (k.c. cowley, e.zaporozhets, b.j. schmidt, intrinsic spinal neurons in the spinal cord in neonatal rat spinal cord sufficient for the myelinated propagation of motor command signals (pro-cervical area administration for the cortical communication of the local communication signal in the cervical spinal cord), the journal of physiology 586,1623(Mar 15,2008)). Second, we implanted the micro-lead array near CST neurons projected at segments T8-T9 in an above ground trained rat (n ═ 3) and recorded the neuronal modulation during voluntary movements. We found that the modulation pattern of a large number of neurons (n-17/24 neurons) is clearly related to gait initiation, sustained motion and corrective motion (p < 0.05). A significant number of motor cortical neurons (36%) showed a dramatic increase in firing rate before any significant movement or motor-related muscle activity occurred. In contrast, the firing rate of moving cortical neurons during involuntary movement was significantly reduced compared to quiescence (p < 0.05). Third, we inactivated the left motor cortex using microinjection of the GABA agonist muscimol. While not impairing the functionality of the lumbosacral loop, muscimol immediately inhibits autonomous hindlimb movement (p < 0.01).

Up to now, functional recovery after SCI has been explained as the need to promote long-distance regeneration of the severed fibers to their original target (L.T. alto et al, chemo-guided promotion of axonal regeneration and synapse formation after spinal cord injury (Chemotropic differentiation and synapse formation) Nat neurosci12,1106(Sep, 2009); F.Sun et al, Sustained axonal regeneration due to co-deletion of PTEN and SOCS3 (Sustanated axon regeneration by co-deletion of PTEN and SOCS3), Nature, (Nov 6,2011)). There is no question that nerve regeneration is essential after nearly complete SCI. However, a more direct approach may be to take full advantage of the significant ability of the remaining neuronal system to reorganize by using a usage-dependent mechanism (A. Wernig, S. Muller, using weighted Laufband motion to support improved walking in people with severe spinal cord injury (Laufband and coordination with body weight super exercise in patients with severe spinal cord injury), Paraplegia30,229(Apr,1992), S.Harkema et al, the Effect of epidural stimulation of the lumbosacral spinal cord on voluntary movement, stance and assisted walking after a complete motor amputation-case study (examination of the spinal cord tissue fusion on the spinal cord fusion movement, and simulation of the spinal cord, surgery around the spinal cord, network 2008, sports et al, branch 6332, branch of sports, network 2008. license 12. Branch.32, Branch et al). Here, our established training conditions not only enable but also force the brain to build multiple new brainstem and intra-spinous relays to regain quantitative and qualitative access to the electrochemically energized lumbosacral loop. There is increasing evidence that active Training with appropriate sensory cues is clearly superior to passive robot-guided rehabilitation to improve stepping capacity in humans (a. wernig, s. muller, using weighted Laufband exercise support improved walking in people with severe spinal cord injury (Laufband coordination with body weight weighted exercise in people with severe spinal cord injury), Paraplegia30,229(ap, 1992), s.harkema et al, Effect of epidural stimulation of the lumbosacral cord on autonomous exercise, standing and assisted stepping after exercise complete Paraplegia case study (Effect of anatomical simulation of the spinal cord exercise, Training, exercise assisting strategy), and learning of spinal cord after exercise complete Training (2008. 57,241. branch. 3. exercise machine Training) as-a-novel fibrous step trailing after a complex mineral aggregate in mineral matrices of motor learning), J Neurosci 26,10564(Oct 11,2006); wernig, "invalidity" of automatic sports training, "achives of physical media and rehabilitation 86,2385(Dec, 2005); wirz et al, Effectiveness of automatic motor training in patients with long-term incomplete spinal cord injury (automatic of automatic motor training in patients with chronic incomplete spinal cord injury: organic instruments of physical media and rehabilitation 86,672(Apr, 2005); P.Musenko, R.van den Brand, O.Magezetondorfer, A.Larmac, G.Courtine, Combined Electrical and pharmacological neural interface to regaining motor function after spinal cord injury), IEEE.d. 56,2707 (Experimental 56,2707). Likewise, automated confined treadmill training without cortical neuron participation promotes sub-pathological plasticity, but does not promote remodeling of the descending branch pathway. Treadmill trained rats did not regain spine-mediated locomotion.

In view of the foregoing description and examples, the present invention introduces a new training paradigm that encourages active participation by subjects and triggers a cortex-dependent activity-based process that restores autonomous control of delicate complex motor movements after SCI leading to long-term paralysis.

These results demonstrate the ability of the intraspinal circuits to bypass the pathology (Courtine, g., et al, control on the spine to restore stepping through the indirect spinal intrinsic relay connections following spinal cord injury (recovery of spinal adjuvant regenerative feedback after spinal cord injury), Nature media 14,69(Jan, 2008); f.m. bareyre et al, spinal cord injured in adult rats spontaneously develop new intraspinal circuits (treated spinal adjuvant circuits for a new intraspinal circuit in adducts), Nat Neurosci 7,269(Mar,2004)), and extend their treatment to functional recovery after the paralytic potential of sciatic. Training the ability to promote this extended plasticity and recovery in a highly functional state may lead to new interventions that can improve function in humans with a variety of neuromotor impairments (s. harkema et al, the Effect of epidural stimulation of the lumbosacral Spinal Cord on voluntary locomotion, stance and adjuvant stepping after motor complete paraplegia; case study (Effect of anatomical simulation of the lumbosacral Spinal Cord on volitional locomotion, standing, and assisted stepping; lancet study of a. cassette study; lancet377,1938(Jun 4,2011); b. a. kakulas, a particular characteristic-emphasized neurology of human Spinal Cord injury (a review of the neurological of the same kind of Spinal Cord injury, surgery of Spinal Cord injury, Spinal Cord recovery model of Parkinson's disease, Spinal Cord stimulation of Spinal Cord, Parkinson's model of Spinal Cord stimulation, Parkinson's disease, surgery, Parkinson's disease, Spinal Cord stimulation of Spinal Cord injury, Parkinson's disease, Spinal Cord stimulation, Parkinson's disease, model of Spinal Cord, r. 22,119, Spinal Cord stimulation, Parkinson's model, p. r. sub. model, Parkinson's. 3. model, Spinal Cord injury, model, 3. b. sub. 1, Spinal Cord injury, animal model, Spinal Cord injury, Spinal, science 323,1578(Mar 20,2009)).

Claims (26)

1. An apparatus for restoring voluntary control of locomotion in a subject suffering from neuromotor impairment, the apparatus comprising a multidirectional trunk support system and a device for epidural electrical stimulation, wherein the multidirectional trunk support system comprises:

a. a robotic interface that actuates an end effector with a degree of freedom n;

b. a mechanism integrated in or attached to the robotic interface to provide compliant/elastic or viscoelastic behavior in the degrees of freedom at an end effector of the robot;

c. a sensor for measuring the movement of the end effector resulting solely from such compliance; or a sensor for measuring the force generated by such compliant movement;

d. an interface connected to the subject using the apparatus to transfer any twist in the degrees of freedom to the subject.

2. The device of claim 1, wherein the device provides support for the subject against gravity.

3. The apparatus of claim 1 or 2, wherein the multidirectional torso support system comprises:

i. a multidirectional elastic decoupling system; three motor-driven linear actuation modules having horizontal, orthogonal axes X and Y along an X, Y, Z cartesian coordinate system and a vertical axis Z, each defining 4 degrees of freedom, and one motor-driven rotary actuation module about the vertical axis Z; wherein the linear actuation modules are simultaneously decoupled by a suspension system having compliant elements oriented in each of the 4 degrees of freedom;

a passive parallel Delta motion system to prevent tilt.

4. The apparatus of claim 3, wherein the sensor is a position sensor or a force sensor.

5. The device of claim 3, wherein a computer communicates with the module and acquires information from the sensor.

6. The apparatus of claim 5, wherein the computer exchanges information with a second computer running the user interface.

7. The device of claim 4, wherein a computer is in communication with the module and acquires information from the sensor.

8. The apparatus of claim 7, wherein the computer exchanges information with a second computer running the user interface.

9. The apparatus of claim 3, wherein the motor-driven actuation modules provide constant force modes independent of each other.

10. The apparatus of claim 4, wherein the motor-driven actuation modules provide constant force modes independent of each other.

11. The apparatus of claim 5, wherein the motor-driven actuation modules provide constant force modes independent of each other.

12. The apparatus of claim 7, wherein the motor-driven actuation modules provide constant force modes independent of each other.

13. An apparatus according to claim 3, wherein the motor-driven linear actuation modules along the horizontal, orthogonal axes X and Y and the motor-driven rotational actuation module about the vertical axis Z provide a transparent mode and the motor-driven linear actuation modules along the vertical axis Z provide a constant force mode.

14. An apparatus according to claim 4, wherein the motor-driven linear actuation modules along the horizontal, orthogonal axes X and Y and the motor-driven rotational actuation module about the vertical axis Z provide a transparent mode and the motor-driven linear actuation modules along the vertical axis Z provide a constant force mode.

15. An apparatus according to claim 5, wherein the motor-driven linear actuation modules along the horizontal, orthogonal axes X and Y and the motor-driven rotational actuation module about the vertical axis Z provide a transparent mode and the motor-driven linear actuation modules along the vertical axis Z provide a constant force mode.

16. An apparatus according to claim 7, wherein the motor-driven linear actuation modules along the horizontal, orthogonal axes X and Y and the motor-driven rotational actuation module about the vertical axis Z provide a transparent mode and the motor-driven linear actuation modules along the vertical axis Z provide a constant force mode.

17. A pharmaceutical composition comprising a combination of agonists of monoaminergic receptors for restoring locomotion in a subject suffering from neuromotor impairment in combination with the device of any one of claims 1-16.

18. The pharmaceutical composition of claim 17, wherein the combination is a combination of agonists for serotonergic, dopaminergic and adrenergic receptors.

19. The pharmaceutical composition of claim 18, wherein the combination is a combination of agonists for 5HT1A, 5HT2A/C, 5HT7, and DA 1-like receptors.

20. Use of a device according to any one of claims 1-16 in combination with an optionally administered pharmaceutical composition according to any one of claims 17-19 for the manufacture of a medicament for restoring voluntary control of locomotion in a subject suffering from a neuromotor impairment.

21. Use of treadmill exercise, the device of any one of claims 1-16, and optionally the pharmaceutical composition of any one of claims 17-19 in the preparation of a medicament for restoring voluntary control of locomotion in a subject suffering from neuromotor impairment.

22. The use according to claim 20 or 21, wherein the neuromotor impairment is partial or complete paralysis of a limb.

23. The use of claim 20 or 21, wherein the neuromotor impairment is selected from the group consisting of spinal cord injury and stroke consequences.

24. The use of claim 22, wherein the neuromotor impairment is selected from the group consisting of spinal cord injury and stroke consequences.

25. A method for operating the apparatus of any one of claims 3-16, the method comprising the steps of:

a. an evaluation mode, in which the device provides support against gravity in a spring-like or gravity-reduced state, using a motor-driven actuation module along the vertical axis Z;

b. an energized mode, wherein the device provides propulsive and/or postural assistance using constant speed forward motion with the motor driven actuation modules along the horizontal axis X, while the motor driven actuation modules along the vertical axis Z provide constant force vertical support at a percentage of body weight, and the motor driven actuation modules along the horizontal axis Y and the motor driven rotational actuation modules around the vertical axis Z provide rigid support in both lateral and rotational directions;

c. a training mode wherein the apparatus provides postural support against gravity with the motor-driven actuation module along the vertical axis Z, the motor-driven actuation module along the horizontal axis X is set to be transparent, the motor-driven rotation actuation module about the vertical axis Z is set to be rigid or transparent, and the motor-driven actuation module along the horizontal axis Y is set to be rigid or transparent.

26. The method of claim 25, wherein principal component analysis is performed on a gait cycle.