Rehab Bioengineering - Sbobinazze PDF

REHABILITATION ENGINEERING: is systematic application of engineering sciences to design, develop, adapt, test, evaluate, apply, and distribute technological solutions to problems confronted by individuals with disabilities. The aim is to help people with disabilities to regain lost cognitive, sensor...

REHABILITATION ENGINEERING: is systematic application of engineering sciences to design, develop, adapt, test, evaluate, apply, and distribute technological solutions to problems confronted by individuals with disabilities. The aim is to help people with disabilities to regain lost cognitive, sensory and/or motor functions. POSTURAL CONTROL (slide Posture-f2024final) Parkinson Disease ​ Major Symptoms: ​ Tremor ​ Rigidity ​ Gait failure ​ Postural Instability Quiz – posture: which one is the healthy subject and which one is affected by Parkinson? The major symptoms of Parkinson Disease are tremor, rigidity, gait failure and postural instability. We make a distinction between data of three subjects to see if they are healthy or impaired with PD. stabilogram: ML / AL componentof the CoP plotted vs time statokinesiogram: AP vs ML (NO TIME!!) This is a healthy subject, because: ​ looking at the statokinesigram, the graph that represents the center of pressure coordinates in the Antero-Posterior direction vs the CoP coordinates in the Medio-Lateral direction, we can see that the trajectory can be approximated more by an ellipse, symptom that we oscillate more in the antero-posterior direction than in the ML one, but there is also a contribution of the ML direction, so the ellipse is a little bit rotated towards left. ​ Looking at the histogram, we see two peaks around 0 in the AP one, that is due to the already mentioned oscillation around this direction, caused by gravity that causes us to move forward and backward and also the ground force. ​ By the stabilogram, graph on the left, we see that we have both contribution from ML and AP diection. This subject has PD, because: siamo sicuri di questo?? ​ We can see from the stabilogram that the effect of AP is even smaller than the ML one, and this is due to rigidity. ​ We can see that in both directions, on the histogram, we have a gaussian distribution that means noise added to the normal movement. ​ From the statokinesigram we see that the trajectory can be approximated more by a circle, so there is not the prevalence of the AP direction. This subject has PD because: ​ On the stabilogram we see too many oscillations. ​ On the histogram we see too many peaks, too distributed. ​ On the statokinesigram we see that the ellipse is straight, so too much oriented in the AP direction while should be some degrees of rotation towards ML direction. Postural control The postural control is the ability to control the position of our body against the forces of gravity is and fundamental for every action. Every movement that an individual makes consists of: A postural component -which stabilizes the body. The first movement -which is related to a particular goal. The postural control has two functional objectives: postural balance/stability to be capable of controlling the CoP postural orientation to have the body segments well aligned Example. Subject with camptocormia: ​ involuntary forward flexion of the hips, so CM is not in the same position of an healthy subject. ​ Have excellent control over their center of mass (postural equilibrium/balance/stability). Other patients: ​ Excellent postural orientation (in terms of alignment and multisensory orientation to the external environment) ​ Still be at risk of falling due to poor postural control. So, the postural balance and p. orientation are two distinct aspects of postural control, but despite this example they can be interdependent, indeed flexed postural orientation of the legs and trunk compromises the ability to recover equilibrium in response to perturbations (postural balance problem) and often co-exist in pathologies like Parkinson. The limits of stability are not fixed, but depending on: ​ Person (age, disease…) ​ Environment (distraction etcc): this is a problem for the engineering because the experiment and measurement are done in laboratory and the result could be very different than the real environment. ​ Task : same position, but could be different muscle activation:  here we have the same position but different tasks to perform. Stability depends on these!! We can distinguish two types of perturbations: ​ Internal Disturbance: when self-initiating a movement (e.g. to pick up or lift an object) ​ External Disturbance when a disturbance from the outside accidentally destabilizesthe body (e.g. a person bumps into the crowd) Two mechanisms of postural adjustments ​ Compensatory postural adjustments (CPA) o​ They occur after a disturbance o​ They are activated by feedback sensory signals - feedback postural control o​ They are used to reorganize posture and maintain balance. ​ Anticipatory Postural adjustments (APA) Can be of two types: o​ Muscle activity that accompanies and maintains a fully operational situation (aAPA –a accompagnatory) o​ Muscle activity that occurs before the onset of disruption (pAPA - preparatory) They serve to minimize the displacement of CoM before the perturbation (Bouisset and Zattara, 1987; Aruin and Latash, 1995) Feedforward vs. feedback control: ​ Control feedback: comparison btw the desire state and the one that is measured (Present state). ​ Pro: simple. ​ Cons: stability problem due to the delay pf sensory feedback.  ​ We use to stay sit. ​ Learning feedback or feedforward: Inverse model of our body that can predict the way the body responds through experience. This allow to better adapt the way in which our body behave or adapt to the environment. Indeed, we have delay from reaction and noise from the environment. ​ Pro: no stability problem, because is based on anticipatory feedback. ​ Cons: Complex because it needs a internal model and learning process. ​ We use to stand (stare in piedi). The variables of interest Center of pressure (CoP): ​ Is the point in which the resulting of the grounding force is applied. ​ u.d.m [mm] Center of mass (CoM): ​ Is a hypothetical point where the entire mass of an object may be assumed to be concentrated. ​ u.d.m [mm] Gravity force ​ Applied to the CoM Ground reaction force ​ Applied to the CoP. Why we use model? Because it is possible to obtain information of greater readability and clinical utility. Indeed, starting from the models, we can extract a certain number of parameters that make the information obtained from the measurement more compact and readable. PROB: There is a trade off btw the ease of use and readability of the model and its adherence to the reality, indeed, more is similar to the reality and more the model become complex. We have to keep in mind that there is differences btw model and reality. Stabilizations strategies: To avoid that the difference CoM-CoP increases the natural tendency to fall, we can use mechanisms of control: 1.​ Ankle strategy: to act on the CoP modulating the Activation of the ankle muscles – direct and fast mechanism. 2.​ Hip strategy: To act on the CoM modulating the Activation of the hip muscles – indirect and slow mechanism. It uses also the ankle muscle 3.​ Augmenting the base of support by doing a step. Muscles responses to forward and backward sway perturbations Activation of muscle during these strategies: EMGs for 6 muscles arranged in antagonist pairs. Note: Here is the envelope, that for definition is positive, so the negative is for graphical reason, as for backward are activated the antagonist muscle (anterior chain) that are active for forward sway (posterior chain). 1.​ Normal surface: Ankle strategy In Forward Sway: (we have to go back) The first to activate is the medial gastrocnemius, then the hamstrings and finally the lumbar paraspinal muscles. 🡺​ Use of the Posterior chain In Backward Sway: (we have to go forward) The first is the Tibialis anterior (antagonist of gastro), then the Quad, i.e. rectus femoris (a. of Harm) and finally Rectus abdominis (a. of Para). 🡺​ Use of the Anterior chain ❖​ Activation is from distal to proximal (vicino al centro del corpo). 2.​ Short surface: Hip strategy In Forward Sway: (we have to go back) The first to activate is the Rectus abdominis and then the rectus femoris. The lower (distal) part of the body is almost not used. 🡺​ Use of the Anterior chain In Backward Sway: (we have to go forward) The first is the lumbar paraspinal muscles and then the hamstrings. 🡺​ Use of the Posterior chain ​ Activation is from proximal to distal. What can happen with older subject or affected by pathologies? ​ There is a delay in the activation. (see time bar in the fig) ​ Subjects affected from diseases (like Parkinson) that has rigidity symptoms has problem of coactivation (or called cocontraction) Upright standing The analysis of a motor act consisting of standing in a relaxed position, in quiet conditions without perturbation. Postural Sway during Upright Standing Subject standing on the force platform. CoM and CoP oscillate with different amplitude and frequency bandwidths: ​ CoM = 0.5 Hz ​ Cop = 4 Hz. 🡺​ CoM oscillation correspond to a real movement of the body mass. 🡺​ CoP oscillations do not represent any movement because are proportional to the ankle torque. Why there are that oscillations? In the model (Morasso Schiepatti 1999) the human body is represented as an inverted pendulum pivoting at the ankle. This model assumes that balance is maintained by controlling the torque around the ankle joint. The gravitational force acting on the center of mass creates a torque that must be countered by ankle torque to stabilize posture. For an inverted pendulum (see Fig. 1) the system equation is: ​ ​ θ: is the sway angle ​ m and Ip: are the mass and moment of inertia of the body (minus the feet) ​ h: is the distance of the COM from the ankle ​ τ𝑎𝑛𝑘𝑙𝑒 is the total ankle-torque ​ g is the acceleration of gravity ​ z stands for the set of external or internal disturbances (such as respiration) that perturb the standing posture  ​ Ip​θ¨ is the angular moment (comes from the application of the second law of Newton for a rotational system, that stabilizes that the sum of the moment or torques applied is equal to the moment of inertia multiplied for the angular acceleration). The variation of the angular moment has to be balanced from torques (right of the equation) to maintain the equilibrium. ​ Mgh sin(θ) is a gravitational torque For small angles, sin(θ)≈θ, so the gravitational torque simplifies to: Equation 1. The ankle-torque must also satisfy an equilibrium equation for the foot: 🡺​ ​ fv: is the vertical component of the ground reaction force ​ u: is the COP position If we take in account that in quite standing, then this equation tells us that variations of muscle torque are immediately and linearly translated into variations of the COP position. The two equations can then be combined into a single dynamic sway equation that relates the controlled variable y and the control variable u: ​ Final equation: ​ y represents the CoM position, so ÿ represents the CoM acceleration. ​ u represents the position of the CoP. In other words, the CoM-CoP difference is bound to be approximately proportional to the acceleration of the CoM for purely mechanical reasons (is fixed with the anatomy, indeed it depends on the gravity force and h), because the postural noise z’ is small in quiet standing. So can be written as: Equation 2. Conclusion, and what we are interested in: 1.​ Equation 2 explain that instabilty derives from the fact that a positive difference CoM-CoP positively accelerates the CoM, viceversa a negative difference leads to a backward acceleration. 2.​ From Equation 1. Critical Level of Combined Stiffness: to counteract the gravitational torque and stabilize posture, the body must generate an ankle torque that matches this critical level of stiffness. In terms of a stability criterion, the critical stiffness needed to stabilize this system is given by: This criterion indicates the threshold of ankle stiffness necessary to prevent a fall due to gravitational torque. Stabilization mechanism 1.​ Physical mechanism ​ it is linked to muscular stiffness  ​ Implicit feedback control ​ Acts instantly without delay 2.​ Reactive mechanism ​ It is determined by different types of reflexes acting independently ​ Closed-chain or feedback control ​ Acts with delay due to propagation of the nerve signal 3.​ Anticipatory mechanism ​ It has an integrative nature. ​ Open-chain or feedforward control ​ Based on internal model of sensory fusion and dynamics prediction. ​ It can cancel delay of transduction and propagation of the sensory information. These three mechanisms work in synergy as alone they cannot guarantee the postural stability: ×​ Experimental evidence show that the muscle and tendons stiffness give a contribute only of 60-90% of the critical value (Kcritical), so alone it is insufficient to stabilize the posture. The feedback control for the postural stabilization includes the interaction of visual, vestibular, proprioceptive and tactile afferences. This mechanism is insufficient to stabilize the posture as: ×​ The feedback control with the delay of the spine reflexes is instable ×​ The afferent channels do not provide info about the CoM ×​ Beside the tactile channel the other are below the threshold, i.e. they do not provide sufficient information for postural stabilization. So there in this context a feedforward control help to stabilize the posture, before an act happen (as in the act of caching a ball). The compensation is achieved by a mix of anticipatory feedforward control and muscle stiffness, the Feedback control intervene only in sudden & gross disturbance. Stiffening-up (coactivation) is a strategy for reducing insufficient feedforward control, but this strategy is functional but inefficient. It is biomechanical inefficient as it increases energy consumption, reduce agility and efficiency of movement and unnecessary strain on muscles and joints. It is cybernetically (cybernetics refers to the study of control systems) inefficient as: ​ Suboptimal Use of Control Resources: ​ When the coactivation is used, the system relies on stiffness rather than precise feedback or feedforward control, which is a less effective way to handle disturbances. ​ It essentially sacrifices precision and adaptability for brute force stabilization. ​ Limited Sensory Feedback Sensitivity: 1.​ Coactivation reduces the sensitivity of proprioceptive sensors (e.g., muscle spindles). High stiffness dampens the small changes in muscle length that these sensors detect, reducing the quality of feedback about joint position or external perturbations.  ​ Infettive Long-Term Strategy: ​ Instead of improving the predictive (feedforward) mechanisms for anticipating disturbances, this strategy compensates for deficiencies by "overloading" the system. It addresses the symptom (instability) rather than improving the root cause (better predictive control). Anticipatory feedforward control is a form of APA, it applies in general to all kinds of self- generate disturbances that cannot be compensate by reflex mechanisms but require an integrative model. Evidences of anticipatory nature of the control: 1.​ From postural control: ​ The CoP is the control variable, and CoM is the controlled one, i.e the body acts changing the CoP to control the CoM. Example: As you begin to lean forward, your postural system shifts your CoP toward your front feet before your CoM moves too far from your center of gravity. This anticipatory movement helps prevent an imbalance. ​ Claearly, CoM and CoP are in phase and CoM can be obtained applying a low pass filter on CoP. ​ To satisfy the first point the CoP moves before the CoM (they are in phase, but the CoP gives the ‘command’), so the anticipatory nature is clear. 2.​ From electromyography ​ Lookink at tge CoP and the EMG activity it is possible to notice that the muscle activation precedes the CoP, so the muscle activity has an anticipatory and not reflexed nature. STRUMENTATION FOR POSTURE MEASUREMENTS (Slide postura_strumentazione2024) How to perform a postural analysis? Force platforms/dynamometric platforms It measures displacement of the CoP, but also the ground reaction force, the momentum. What sensors they use? Force Platforms Purpose: measure the ground reaction force F. What is measured? In a fixed reference system, it measures the forces exerted by a subject during the execution of specific motor task. Technology The measurements can be transformed with two alternative transduction systems: ​ Piezoresistive (strain gauge load cells) which measure the variation in the electrical resistance of the sensor caused by compression or traction. (there are also piezo capacitive sensors which variate their capacitance) ​ Piezoelectric: measures of electrical polarization due to mechanical deformation of materials. Example ​ Three sensors are enough Dynamometric platform Purpose: A broader term that refers to devices capable of measuring forces, moments, or torque. For both… The information on the force vector is obtained and its point of application (CoP) is determined. We measure the vertical Force and Momentum and then we calculate the CoP. So, for the calculation of the CoP coordinates, it is sufficient to calculate three components the other are overabundant for this task, but it allows to obtain other information. What do you have to consider when buy a Force platform? ​ What you must calculate ​ Size (for neurological examination is better to have large, because in a small surface I oscillate more) ​ Signal: if amplify directly on board and not when we got on our pc because if I amplify after transmission, I will amplify also the noise of the transmission. ​ Software need to be certified with the device, so if is not giving with board and I have to use with patients I have to do the certification and this means higher cost. 2.​ Linearity/histerisis 3.​ Resolution (N,Nm,mm) (For posturography: spatial resolution of the CoP is 0.1-0.2 mm) 4.​ Frequency bandwith (Hz) (For post: it must include 0 Hz and exceed 5 Hz) Processing of the signal Can be done in: 5.​ Time domain 6.​ Frequency domain 7.​ Time-frequency domain Considering Time domain Stabilogram (SBG) has the time information, has the coordinate of the CoP through time. In healthy people we expect that the angle oscillation around the Y axis is higher than X. Statokinesiogram (SKG) is the displacement of the CoP in the plane (x,y). The abscissa is considered positive in the right of the subject, while the ordinates are positively oriented toward the anterior direction. Derivation of parameters Global parameter that could be extracted from the Cop, that can give you information if the movement is ‘abnormal’: 8.​ Sway path 9.​ Sway area 10.​Mean amplitude 11.​Mean frequency 12.​Confidence Ellipse Beyond the randomness of the randomness of Statokinesigram there is a regular structure characterized by: ​ Regular cluster ​ Quick shift from one cluster to another So it is important to define another parameter: 13.​Sway Density 🡺​ for each time instant, the number of consecutive samples of the CoP which fall inside a circle with a given radius (typically 2.5, 3 mm) multiplied by the sampling time. 🡺​ Sway density parameters: ▪​ Mean time between peaks (MT): cannot be affected by pathological condition, it depends only on biomechanical parameters. (MT = 600+-40 ms) ▪​ Mean value of the peaks (MP) ▪​ Mean distance between peaks (MD) With Parkinson we have increase in amplitude and decrease in distance between peaks. Effects of alcohol on stability Even when Alcoholic rehabilitate, they cannot restore completely the balance. Athletes injured and not P.S. If you compare populations you have to be sure that they are comparable, same age same sex and same characteristics. The number of controls they must be the same of injured. The non-injured players have more control on their balance than the controls. The sway area of injured players increases wrt controls and non-injured. ✔​ More quantitative values than insoles ✔​ Good for dynamics (sport and walking) as it measures F and M. ✔​ Sufficient for standing balance measures as we can calculate Xcop and Y cop ✔​ ×​ We measure the oscillation of the floor ×​ The heartbreak, especially in static condition, can influence the signal measured as it caused micro movement in the body. But it is possible through the averaging w.r.t heartbeat reduce this noise. Baropodometry platforms The pressure exerted by the feet of a subject can be detected by a platform, that through their sensors, allow the detection of the map of plantar pressures. In the color maps, the red is related to more pressure. Sensorized insoles: like baropodometry platforms, but we have insoles (solette) ✔​ Compared to the baropodometry platforms they are wearable. ✔​ Low cost ✔​ Applicability in wide range of situations ✔​ Flexibility, reliability and durability ✔​ Resistance response and applied force have linear relation in bi-logarithmic scale. ✔​ Measurable pressure range: 10kPa- 1MPa (during pitch up to 400 Kpa) ×​ Lose information related to the absolute reference. ×​ Sensitive to temperature changes: operating 30-35°. ×​ Not sufficient to give exact measurement of force/pressure. More qualitative than quantitative. ×​ Electromechanical characteristics of resistive sensors depend on: o​ Conductor geometry o​ Conductive polymer o​ Possible layer of elastomeric rubber superimposed o​ Rigidity/dimension of the support Technology & Principle of operation The sensors in the insoles measure a variation of an electrical property in response to a normal pressure applied on them Types of sensors: ​ Resistive sensors: as the pressure changes, the electrical resistance (conductance) varies. ​ Capacitive sensors: as the pressure changes, the capacity varies ​ Ink or polymer sensors: exploit the conductive capabilities of inks or polymers. The most used sensors are polymeric piezoresistive elements, based on the variation of the electrical resistance in the conductor-polymer path. This is translated into electrical voltage by appropriate circuits. ​ Types of sensorized insoles: ​ Discrete distribution of sensors: measure the vertical component of the ground reaction force in the areas of interest. ​ Sensor matrix: measures the CoP, distribution of plantar pressures, the vertical component of the ground reaction force. Possible visualizations: Gait analysis: Vertical component of the ground reaction force: Computation of CoP: the insoles allow obtaining the center of gravity of the pressure separtely for the two feet, and then the overall one, once the relative position of the two feet is know. The reconstruction of the CoP can be computed via software, starting from the acquired pressure map, or directly via harware, using apporpiate electrical circuits. Stereophotogrammetry The position of the CoM (!not CoP) is reconstructed starting from the movement of the individual body segment. IMU POSTURAL CONTROL (slide controllo posturale 2024) The postural control deficits mainly derive from deficits in the sensory information that will not allow ate the internal model to operate with precision and as a consequence there will be an higher use of stiffness and hip control. Importance of sensory afferences ‘In pathology it is important to recalibrate sensory information and limit damage created by sensory “silence”, to obtain a more significant postural Stability’ (Marigold et al., 2004) Vestibular system With the Vestibule we perceive linear acceleration, specifically the macule of the utricle allows to detect a. horizontal, while saccule a. vertical. The Semicircular canals encode, through the ampullary receptors, for angular acceleration. The vestibular system alone is not enough to know position of the body in space and control posture, it needs to integrate different sensory information related to: 1.Gravity With otoliths the Somato-sensory system decodes signal of gravity. Somatic graviceptors: ​ Kidneys: through the ligaments placed in the renal suspensory ligaments, through the renal nerves coming from the eleventh dorsal root. It is located in abdomen. ​ Receptors in vena cava ligaments: receptors stimulated by circulating blood, through the vague nerve and the phrenic nerve. It is located in the upper part of the trunk. 2.Support on the ground a. Proprioceptive information, from: Two from the kinesthetics system: ​ Osteo-muscle-tendon truss structure (struttura reticolare) ​ With organs of Golgi (proprioceptive sensor receptors located at the junctions of muscles and tendos that monitors the muscle tension) and neuromuscular spindles (control the length of the muscles) From somato-sensory system: ​ Superficial (Merkel and Meissner) and deep (Golgi-Mazzoni) pressure receptors. b. Somatosensory cutaneous information: ​ From receptor on the skin that are essentials to perceive the contact with the ground and environment ​ From receptor like Merkel, Ruffini and Meissner that detect deformation and vibration on the skin to increase balance and orientation. It is referred, in particular, to the skin of the soles of the feet. Indeed, the vestibule-podalic information, integrated with the proprioceptive and somatosensory systems, allow the body to prevent imbalances through continuous postural adjustment and react quickly to disturbances, improving stability and safety during movement. 3.Body geometry 4.Position of the body wrt the environment 5.State of tension of the anti-gravity tonic musculature (postural tone) Rehabilitation Contact Hand-Orientating Response (C.H.O.R) In rehabilitation the tactile sensor as a lot of importance and it affect the improvement the body pattern and the sense of orientation of the body. The concept of C.H.O.R is referred to the role of the hand touching a surface, that gives sensorial information, support and increase the posture, balance and coordination. C.H.O.R is a mechanism that use the friction of the hand on the surface to help the body orientation wrt the midline, the central axis of the body. ​ A light touch of the hand (even with one finger tip) with a surface will help the patient to balance and maintain postural control. ​ Stabilization for selective movement as it helps to stabilize articulation like wrist, elbow and shoulder allowing more precise movement. ​ Improving of the movement that needs to cross the midline. E.g movement of the arm on the diagonal as you do for the shoulder. Clinical Measurement in 1986: different visual and surface conditions. LEZ 20241001 Movendo non vista Sensory Organization Test (SOT) The images shows a diagram of the Sensory Organization Test (SOT), a tool used to assess balance and postural stability by manipulating sensory inputs. Purpose: The test isolates sensory contributions to balance (visual, proprioceptive, and vestibular) by controlling or altering the stability of these systems. Application: This test is often used in clinical and research settings to evaluate individuals with balance disorders, vestibular or neurological conditions. (to diagnosing vestibular disorders, as in condition 6 the subject has to relay only on vestibular system, and so we can see if he has problem). Disturbance -Visual system is altered from the bandage on the eyes, also from the moving wall. -The Proprioceptive systems feedback is disrupted from the moving platform. (Normally, proprioception provides reliable information about joint and muscle position, but this is no longer accurate due to platform motion) - The vestibular system senses motion and spatial orientation using the inner ear's semicircular canals and otolith organs. It is independent of external factors like visual input (wall movement) or proprioceptive feedback (platform movement). Conditions 1.​ Eyes Open (Condition 1): o​ All sensory inputs (vision, vestibular, proprioceptive) are available. o​ Represents the easiest condition as the subject can rely on all systems. 2.​ Eyes Closed (Condition 2): o​ Visual input is removed, forcing reliance on proprioceptive and vestibular systems. 3.​ Moving Wall (Condition 3): o​ The visual surround (wall) moves, disrupting visual input accuracy. o​ The individual must rely more on proprioceptive and vestibular systems for balance. 4.​ Moving Platform (Condition 4): o​ Proprioceptive input is disturbed by platform motion. o​ Balance depends on vision and vestibular inputs. 5.​ Eyes Closed + Moving Platform (Condition 5): o​ Both vision and proprioception are unavailable or disturbed. o​ The vestibular system becomes the primary balance mechanism. 6.​ Moving Platform + Moving Wall (Condition 6): o​ Both visual and proprioceptive inputs are altered. o​ The vestibular system is the main input available for maintaining balance. The graph below indicates balance performance across the six conditions. o​ 0 represents the best performance (most stable). o​ 1 represents the worst performance (least stable). Trends: o​ Balance typically worsens as more sensory systems are disrupted. o​ Normal individuals show the best performance in Condition 1 and progressively more difficulty through Condition 6. Role of Sensory Systems: o​ Vision is crucial for balance when proprioception or vestibular input is compromised. o​ Without vision, the body relies on proprioception and vestibular systems. THE RESULTS: Sensory inputs for Steady State Balance. From the research on the examination of the sensory organization in quiet stance balance using dynamic posturograhy it comes out that: ​ All the 3 senses are used for steady state balance.  ​ The relative contribution of individual senses is flexible. Improving balance via supplemental feedback 1. Auditory feedback is able to reduce a oscillation of a person. IMU to calculate the COP and COM A possible schema of the study procedure and results The follow up is important because the person must not lose the restored abilities. Follow up is the real life, so if the test is something connected to the real-life situations maybe the person will be more and more skilled to perform the task 2.Vibrotactile feedback​ We have to control frequence and amplitude together to be more effective on the vibration controls. The vibration and its intensity is proportional to the oscillation of the subject. The Blue zone is referred to the non-correct position, where there is vibration. In the middle, there is a ‘silent zone’ that indicate the correct position where there is not vibration. Visual Feedback Commercial system to train people’s balance/postural ability Using vision instead of proprioception is not always a good idea, it depends on what we wanto to train. Indeed, vision is a good for the results but people does not learn. The experiment consist on puts the subject sitting on a force platform in front of the screen and he has to connect the point. This does not give information about postural control. GAIT ANALYSIS – Slide Gait Analysis 2024 Two different types of gait analysis can be performed: ​ Single step analysis 🡺​ Performed on a few steps ✔​ Usefull when it is necessary to relate kinetics, kinematics and electromyography for one to three steps. ✔​ It allows to identify cause-effect relationships ×​ It doesn’t describe the performance of the subject in a statistically acceptable way. P.S. Kinetics is related to the force that cause the movment (to study it the subject walks on a platform with both feet), kinematics describe the movment considering parameters like velocity, acceleration… ​ Statistical gait analysis 🡺​ Performed on entire walk lasting few minutes ✔​ Usefull when it is necessary to quantify the motor performance of a subject during a walk that has a functional meaning and consists of various steps. ✔​ It allows to identify cause-effect relationships The gait analysis until now has been done in gait analysis lab. ​ Usually, motion analysis lab is made by camera (IR light goes on marker that are on the top of the person and the reflections are captured by the camera 🡪 marker motion capture system). Gait analysis, in clinical, can be used for 2 different purposes: ​ We use gait analysis to answer questions about a specific subject, quantify a specific problem. ​ We can have homogeneous population we need specific protocols that depend on the population. EMG: ​ has always been done for analyzing muscle activity. ​ cannot be performed alone, it must be recorded with another signal 🡪 because we want to coordinate the movement we’re looking. ​ depends on skin conductance, quantity of fats etc. 🡪 so I cannot compare EMG signals across subjects. In gait analysis, the ISB (International Society of Biomechanics) convention regarding the coordinates of the reference system are: ​ Axis x directed as the subject motion during walking (parallel to the ground). ​ Axis y directed perpendicular to the ground and positive from bottom to up. ​ Axis z directed consequentially. The plane defined by the coordinates XY is the sagittal plane, YZ frontal, XZ horizontal/transversal. ! When we take a study, the control group, made of unimpaired people, needs to be with people of same age, sex, height and weight. Parameters that are considered during the gait analysis: Understanding both temporal and spatial parameters is essential for a comprehensive analysis of gait. This allows for better comparisons across individuals with varying walking speeds and provides insights into stability and mobility. Temporal Parameters 🡪 refer to the timing aspects of walking. Help in understanding how different individuals walk at different speeds. ​ Considering the gait, the duration of gait phases can be expressed in: - Percentage of the gait cycle for better comparisons across different walking speeds. - Seconds (Stance, swing and stride durations) Spatio – temporal parameters Expressed in meters: ​ The stride length indicates the interval between two successive initial contacts of the same foot with the ground ​ The step length is instead the interval between the initial contact of one foot and the initial contact of the other foot. Distance covered in a single step. ​ The step width is the distance between feet during walking (can indicate stability: a wider distance may indicate instability). Expressed in meters/second: ​ Walking Speed normal walking speed can vary between individuals. o​ Swing speed, speed stance average walking speed Expressed in Hertz: ​ Cadence [steps/minute] Normality Range 🡪 range of gait parameters observed in a healthy population. ​ Not the same for everyone (typically varies by age and sex). ​ A single normality range is often insufficient, as walking patterns differ significantly. Ex: Older adults and children have distinct walking patterns despite similarities. Gait phases During the entire gait cycle (100%), we have the stance (60%) and the swing phase (40%) (we’re interested in the percentage 🡪 it’s all expressed as % of the gait cycle because we want to see if moving slower the percentages are closed to the ones of normal subjects). Stance phases The stance is the phase where the foot is in contact with the ground and has a: ​ accepting response phase or loading response o​ Initial contact 1 o​ Load response 2 ​ two single support phases o​ Intermediate support 3 o​ Terminal support 4 o​ Pre-oscillation 5 The weight acceptance/loading response phase has the functions to absorb the shock of impact, stabilize the limb and preserve forward progression. The single limb support has the functions to support the entire body (with only that limb) and continue the progression. Limb advancement is when we advance the limb and we no longer need to support weight. Swing phase The swing is when the limb advances. The swing phase includes: ​ Initial oscillation 6 ​ Intermediate oscillation 7 ​ Terminal oscillation 8 P.S. Initial and terminal are when the limb accelerates (rectus femoris) and decelerates (hamstrings). The duration of the double support phase decreases as the cadence increases, that is when we start running. The 8 phases of gait cycle can be reduced to 4: heel contact, foot contact, heel raise and swing. This is done for two motivations: ​ Clinical motivation: It has been shown that the majority of pathologies observable by gait analysis can be quantified/monitored effectively even with the four phases of the walk described in the previous slide. We can have equine foot in case of reduced heel contact, lack of heel support phase, spasticity in case of reduced heel raise and limited knee excursion in case of reduced swing. ​ Technical motivation: Quantifying the steps means measuring something through a sensor. To perform an analysis of the step in 8 phases we would need too many sensors, so a procedure inapplicable on pathological subjects and children. Center of Mass It is defined as a single point of mass equal to the entire mass of the body and placed in the center of gravity of the body. It moves when we move a body part. In gait analysis, it is conventionally assumed to be located in front of the second sacral vertebra in the center with respect to the two hip joints. 🡺​ Humans tend to choose a way of walking that minimizes energy consumption. In vertical, CoM moves rhythmically up and down. The lowest point is when both feet are resting on the ground, the highest point in the midstance phase. The average value of the vertical displacement is about 5 cm. In lateral, the weight is transferred from one leg to the other, so the pelvis swings in the frontal plane. CoM oscillations from one point to another have an average peak to peak amplitude of 5 cm. Lateral limits are reached in the midstance phase. 🡺​ The movement done in order to minimize energy consumption is a straight line, in which you go up how much as you go down. To measure energy consumption, it can be looked at the oxygen consumption. From the oscillations of the center of gravity, are extracted parameters like: ​ Velocity ​ cadence ​ step length Often in pathological locomotion there are alterations in the movement of the center of gravity greater than any visible alteration in segmental movements. So, you can have a first characterization of the normal and pathological path based exclusively on monitoring the movements of this last point: the advantages are the simplicity of the measurements to be carried out and low costs, but it is obtained limited information. Analysis of a subject from a paper There is no symmetry between the two legs, and this can be seen by looking at the difference between the parameters of the two legs. She stays more on the stance than on the swing phase. Double support phase is higher when there is the right foot in 1° position. Cadence is lower. It’s not just that she walks slow but also the steps are really shorter (more than half shorter than the control group). She stays in the swing phase less and in that case she moves really fast. 🡪 step width (gambe larghe) is large 🡪 in order to increase the base to have more stability. 🡪 person is walking with smaller step, with asymmetry and with feet far. The subject could be affected by stroke. Average speed is computed in general in the center of mass. Kinematics It is the study of the movement of a body, without reference to the forces that caused it (trajectory, speed, acceleration). It is based, for motion analysis, on the computation of angular quantities (flexion-extension, abduction-adduction, intra-extrarotation), angular velocities and accelerations. A stick diagram is made of sticks, each one representing a joint segment. This diagram shows how the position of the articular segments evolves over time. As a result, it also allows the estimation of instantaneous joint angles. 🡺​ The movements of the ankle during the gait are plantarflexion and dorsiflexion. 🡺​ The knee, during gait, is subjected to flexion and extension. ×​ We can have two anomalies that are knee varus (externally rotated) and valgus (internally rotated). ​ The pelvis has high and low obliquity, anterior and retro tilt, internal and external rotation.  Tilt Obliquity Rotation Analysis of a subject from a paper We have on the left the frontal plane, on the center the sagittal plane and on the right the transversal plane. Because of variability we plot everything on the same scale 🡪 so we use %. Right vs. Left 🡪 in gait analysis, we separate data for the right and left sides because there are often minor differences between them. Normally, right and left steps are very similar unless a pathology is present, which can cause significant differences. Common Issues: - Ankle Problems - Stroke Patients 🡪 often experience ankle stiffness, making flexion difficult. Additionally, hip movement may also be affected. *devi saper commentare I grafici lei all’esame li chiede Kinetics It is the study of the forces that determine the motion of a body: external forces that cause joint motion highlighted by joint moments and powers. If we neglect the contribution of the wind, the friction, etc., the two external forces to be considered in the analysis are: Gravity force (BW): Applied in the center of mass and intensity equal to body mass*g with g=9,81 m/s^2. Ground reaction force (GRF): During the walk the feet exert a force on the ground and this exerts an equal and opposite reaction force. The vectogram is a graph that represents direction and intensity of the GRF during walk, as a function of the sagittal and frontal coordinates. GRF can be measured by force platforms. The force can be decomposed on the 3 Cartesian axes: ​ Front/rear: propulsion/stop ​ Medium/lateral: balance and oscillations ​ Vertical: body support Vertical GRF We have a M curve: at the maximum point, the GRF reaches 120% of the body weight, and this point corresponds to the double support phase; at the lowest point, instead, the GRF is equal to 80% of the body weight, and this point corresponds to the single support phase. Anterior-posterior GRF There is first a braking component (midstance) and then a "propulsive" one of the complex. The AP component can be approximated with a sinusoid of amplitude 25% of the BW. In a complete step cycle the sum of the two components is 0. Center of gravity vertical displacement and acceleration during gait cycle During walking, the center of gravity of the body has vertical displacement and acceleration with opposite sign throughout the cycle. Moving up, acceleration decreases. Moving down, acceleration increases.. Model of the human body The human body is modelled as a single mass on a spring and are considered the forces acting on it. From Newton’s second law, F=m*a. In case of positive acceleration, we have that the GRF is higher than the body weight. During the phase of the double support the center of mass reaches the closest point to the ground. In case of negative acceleration, we have the body weight higher than the GRF. During the single support phase the center of mass is at the farthest point from the ground. Normal subject vs subject with unbalanced limb in anterior-posterior GRF The point of application of GRF, so the CoP, changes its position in the foot during gait. In the figure below ce can see the progression of pressure, monitored by using baropodometric measures, starting from the loading response, passing through the midstance, terminal stance and finally pre-swing. After the pre-swing the foot is not in contact with the ground anymore. This is the reason why on the x axis, in this case, 100% doesn’t correspond to the gait cycle, but to the stance phase. If we wanted to represent the whole gait cycle on the x axis, then we would have had the curve in the graph stopping at 60% and the remaining blocks would have just been empty. As we can see from the curve though, we can say that at the beginning the force is more medial, then it goes more distal and then medial again. The red curve represents what happens in case of a normal foot, the green one what happens in case of a flatter one and the blue what happens in case of a more curved one. Joint moment The joint moment measures the tendency of a joint to rotate around its axes (flexion-extension, abduction-adduction, intra-extra rotation) in response to applied forces. The force has an arm ‘d’ with respect to the ankle and as we can see there’s the generation of a momentum. The displacements of each segment are determined by the forces that produce the angular rotations of the joints (rotoidal). Each joint is subjected to the action of different forces partly generated by the muscles, tendons and partly due to the loads and resistances applied on the body. The most correct way to understand the mechanisms of force production during posture or movement is to schematize body segments as systems of levers. In fact, the easiest way to understand what happens with moments is to refer to the force that generated them and to the moment arms. This implies schematizing the body as a system subjected to a system of forces and related arms. During the support the main function of the muscles is represented by the stabilization of the joints, while the weight of the body advances on the limb on stance. ​ Whenever the GRF vector does not cross the centre of the joint, a rotational force is generated that determines the joint movement and is defined as a moment. When the arm of the GRF is 0 (so when its vector passes through the joint), then there is no momentum. Otherwise, when there’s a moment arm, we have a momentum, and the force generates the rotation of the joint. Apart from the effect of external forces, we have the action of muscle and tendons that instead try to stabilize the joint. WHAT WE REPRESENT ON A GAIT ANALYSIS REPORT ARE THE INTERNAL MOMENTS! Of course, internal and external moments (the ones generated by external forces) are related, but in order to compute the internal momentum we need to use inverse dynamics. During the response to the load, the GRF vector is located posterior to the flexed knee, creating a flexor moment. To stabilize the knee, an extensor response is required: the action of the quadriceps. Joint moments describe the net sum of all internal moments delivered by all internal structures around a joint. Typically, joint moments are delivered by muscles and, toward the end range of motion, by ligamentous or bony tissue. Each of these structures generate an internal force across the joint, the effect of which is given by the multiplication of this force with its moment arm. Where internal muscle, ligament, and bone generated forces are difficult to measure in vivo, their summed effect can be calculated by inverse dynamics. This calculation takes into account the measured segment motions, ground reaction forces, and the (estimated) segment masses and inertia. The net joint moments during gait give an indication of the minimum force level that muscles (or ligaments) produce at any instant during the gait cycle in that they do not take into account for instance any co-contraction of the muscles. It is important to note that while joint moments in gait analysis are in generally defined as the internal moments (as delivered by internal structures), literature has not always been consistent and depending on the time period and discipline might report external moments (as delivered by the ground reaction force over a joint). Conventions: 1.​ The net internal moment is shown including all the component agents: muscles, tendons, ligaments, meniscus, capsules, etc. Thus, the action of agonist and antagonist muscles can be simultaneous (if I have a sharp flexor moment it is not said that only the mm flexors are active) 2.​ Convention for signs: ​ Sagittal plane: extensor moment >0, flexor < 0 ​ Frontal plane: abductor moment > 0, adductor < 0 ​ Full transverse: external rotary moment > 0, internal < 0 In kinematics, flexion is considered positive. ​ The discussion of normal kinetic patterns, so what we really look at in graphics, is focused on: ​ Ankle, hip and knee moments in internal sagittal plane ​ Knee and hip moments in frontal plane This is because transversal joint moment curves vary without showing a clearly interpretable pattern. Power It is given by derivation calculations, that consider a ΔE variation of energy (or work ΔW) in a time interval ΔT. So, the average power is the ratio P=ΔE/ΔT or P=ΔW/ΔT. Power identifies the speed with which the work is done, by considering the mathematical operation, that leads us to the multiplication of the momentum τ and the angular speed ω. It is also related to the ability to generate force over time. Power can be generated or absorbed: ​ Generation is positive. POWER IS GENERATED WHEN THE MOMENT AND THE MOTION (for instance when we flex the knee and go forward. We consider the movement, not the angle) ARE IN THE SAME DIRECTION. In this situation then, we observe an increase in speed in joint’s motion (flexion/extension). ​ Consumption is negative. POWER CONSUMPTION OCCURS WHEN MOMENTUM AND MOVEMENT ARE IN THE OPPOSITE DIRECTION. In this case we have a decrease of speed in the joint’s movement. BY LOOKING AT WHETHER THE POWER IS CORRECTLY ABSORBED OR GENERATED, WE CAN ALREADY SAY IF WE HAVE AN EFFICIENT WALKING. In general, if we plot the speed and the generated power related to the ankle, we find out that there’s a linear relation. This allows us to better understand the importance of the ankle as the joint from which the gait starts, because everything is related to its ability to enable us in the propulsion. This is the reason why usually, when looking at powers and momentum, we mainly pay attention to the ankle’s ones. In kinetics, ankle moment is considered as plantar or dorsal. Knee moment is considered as flexion and extension. Hip moment can be flexion-extension or abduction-adduction. ​ ​ Analysis of a subject from a paper The blue line represents 3 repetitions from the subject whose gait we’re analysing. Why 3 repetitions? They are the recordings related to force platforms, but we don’t have many force platforms in a lab. Typically, there are 3 or 4 platforms at the centre of the room and for each of them we consider the step done on the centre of it, since in order to get a reliable measure, we need the foot to be completely on the platform. If it’s half out, then the measure is not good. The grey line instead represents the reference, given by the average of recoding obtained from subject belonging to the control group (same age and same sex!). By looking at the power of right hip we can see that there is an important delay and also at the level of the right ankle. Moreover, in this case we can also observe that there’s an absorption when there should be a generation and that the curve is almost flat at the beginning and the power generated is very low. Regarding moments, the part in which we expect having a force in abduction-adduction of hip is almost flat, meaning that there is no modulation (it starts then get flat and then decrease). Looking at the power of the right ankle we observe a delay and moreover the power generated is lower than the control. The left ankle power is almost flat, indicating almost no absorption and no generation. This indicates an important impairment. In general, we can also observe that we have the blue lines until the end of the stance phase, because of course, because we have the information recorded from the force platform only until that moment. Hip power is characterized by very small variations and the momentum doesn’t really correspond to the usual one. IMPORTANT OBSERVATION: for what concerns kinematics of the left knee we can see that we have a sort hyperextension, which is pathological. When looking at the kinetics, it means that we observe at the internal moments applied by muscle and tendons in order to produce such hyperextension. If the kinematics was zero, then also the momentum should have been, because 0 kinematics means that there’s no movement. Let’s consider data acquired through force platform now: The vertical forces should have the so-called M shape, while the antero-posterior one should present symmetry. What we can observe in this case is that the 2 maxima of the M are not on the same level of the control group. For the antero-posterior force, the right presents a huge delay, and the peaks don’t reach the same level of the controls. In the left case, not only there’s a delay, but the second peak (so the one related to progression) is also almost absent. The most impaired side is the left one! 🡺​ We can notice that when we deal with walking, the pattern of the so called not impaired side has an impaired pattern too. Remember: FLECTION IS POSITIVE WHEN WE LOOK AT THE KINEMATICS, WHILE WHEN WE LOOK AT THE KINETICS, THE CORRESPONDING MOMENTUM IS NEGATIVE! Muscle activity during walking We’re going to deal only with surface muscle recording. We have 2 electrodes (+1 if the ground is separate) that record a voltage difference at the level of the skin. Usually, we apply gel on the electrode to adapt the impedance. The recorded signal is affected by noise, so we need to use electrodes that have amplifiers on top of themselves, so that we don’t amplify the transmission noise (in order not to reduce the SNR). This is specifically important because signal coming from muscles is really low (in the order of microvolts). In general, when we take a muscle recording, we can look at: -​ Muscles in the frequency domain -​ Muscles in the time domain -​ Muscles activation -​ Amplitude (it’s more difficult to be compared across subjects). We cannot directly compare the measurement coming from muscles of different subjects, because of skin and fatty tissue. In most extreme cases, in which the exam is conducted in a very serious way, the skin can be cleaned using sandpaper and the subject must shave. Of course, skin and the amount of fatty tissue differ from subject to subject. To make the comparison usually people normalize by the maximum voluntary contraction. However, we may end up saying that, for instance, the percentage of maximum voluntary contraction of a subject is bigger than the one of another, but the maximum voluntary contraction of the second subject was bigger than the one of the first. We may also compare by considering the standard deviation at rest, but also that one has drawbacks. In general, we can find different methods, but comparing amplitude remains a problem. For sure the most reliable we look to when considering walking are the specific time instants at which different muscles are active. Muscles, indeed, shouldn’t be active for the entire duration of the gait cycle. If they’re active in the wrong moments we may have co-activation of muscles that are antagonist which is not desired. The pictures on the slide above are different representations of such activation. Specifically, in the bottom picture we have the representation of the envelope. How do we compute the envelope? -​ We start from the raw signal, which must have zero mean. -​ The signal is then band passed with a filter whose band is between 30 and 500 Hz. We do so because below 30 Hz we have noise related to the movement of electrodes and of the subject itself. -​ The signal is sampled at 1 or 2KHz usually (applying the Nyquist theorem). -​ The signal is rectified, meaning that we consider only absolute values. -​ We obtain the envelope by applying a low pass filter with a cut-off frequency in the range btw 3 and 10 Hz (usually 6). How does it come that we use a cut-off frequency which is so low if by band passing the raw signal at the begging we previously cut them off? BECAUSE RECTIFICATION CHANGES THE ENERGY CONTENT OF THE SIGNAL! (think about the sine. If we have a sine and we rectify it, we obtain the double of the frequency). During heel strike (initial contact) the gluteus maximum is active and also the tibialis anterior, they will then be switched off when we start the loading response, during which instead we have the activation of the triceps surae and the quadriceps (that are then switched off until the pre-swing phase). Through loading response, midstance and terminal stance also the calf (polpaccio) muscles (soleus and gastrocnemius will be active). During pre-swing phase also the iliopsoas will be active, to be switched off during the middle swing. In the terminal swing instead, we’ll have the hamstrings active and they’ll remain active until the beginning of the heel strike. Example of normal walking In the picture above we have the representation of a real recording, with a pretty quite old system. We can see different channels. We can recognize the raw EMG signals (represented in yellow). The first starting from the bottom is the EMG of the right gastrocnemius, then right tibialis, left gastrocnemius and left tibialis. What are the yellow bars? They’re the result of a statistical detector of intervals activation. The blue lines represent the knee flection, taken by using an electro goniometer. Green lines instead, give us information about the 4 phases of the gait cycle and they come from a system that is called BASOGRAPHY that gives the information about the phase of the gait cycle. There are switches that are placed under the foot. They measure which part of the foot touches the ground and divide the measurement in 4 phases. There are different technologies to realize the switches. For instance, they can be more discrete or more continuous and so on. The information about which part of the foot touches the ground, gives us the 4 phases of the gait cycle. To know if walking is good we need to look at the different phases of the gait cycle and see if the tibialis and the gastrocnemius-soleus are active during the correct intervals. In general we could say that we’re not able to discuss the efficiency of the gait only by looking at electromyography. This is not completely true though, because there are some problems that can be spotted also starting from it. In this specific case, we can see that there’s co-activation. When gastrocnemius-soleus is active the tibialis has to be silent and viceversa. In this case, the tibialis is silent when there’s the important activation of the gastrocnemius, but the gastrocnemius is also active when the tibialis is. This is called co-activation and such co-activation is pathological (because they’re antagonist). It could be a sign of rigidity or of others neurological problems. When they are always co-activated we have an important rigidity. Gait disorders In the picture we have 3 plots representing the joint angles of hip, knee and ankle against the gait cycle, in the sagittal plane (in which we need to remember that flection is positive). It’ s possible to notice the hyperextension of the knee (genu recurvatum). This leads to an incorrect dorsi-plantar flection and regarding the hip, in a normal condition it was symmetric, while here it is extended. Also, the timing is not perfect. 🡺​ The pathology is poliomyelitis. Given the paralysis of the quadriceps, in absence of knee extensors, we have to prevent the external flection moment, so we always stretch more and more until the posterior capsule also stretch leading to the hyperextension of the knee, which causes problems. In the pictures above the dashed lines represent normality, while the continuous line comes from the recording of our subject. In the graph on the left top part, we have the representation of ankle angle against the gait cycle. In it we can observe a difficulty in dorsi-flection of the ankle. In the left bottom graph, we have the plot of the ankle power against the gait cycle. We can notice that there’s a huge peak of absorption at the beginning and there isn’t the peak that we would expect for what concerns the power generation. So, there is not a proper ankle dorsi-plantar flexion, and this leads to the inability to generate the power. In this case we’re looking at recordings of the left leg of a six-year-old girl having spastic hemiplegia, affecting the left side. Due to the hemiparesis, we expect spasticity (rigidity) to be possibly present. If the problem is at the ankle we expect spasticity at the level of the gastrocnemius, which it is the muscle that allows the ankle movement. In the right bottom graph, we have the representation of the gastrocnemius EMG. The ankle was more plantarflexed than normal throughout the gait cycle: following initial contact in plantarflexion, the ankle became increasingly dorsiflexed through most of the stance phase, and this is due to the fact that the gastrocnemius was firing from the beginning, even if it shouldn’t (the tibialis should have done it). Central Pattern Generator (CPG) They are biological neural circuits that produce rhythmic outputs in the absence of rhythmic input. CPG, networks responsible for locomotion, are distributed throughout the lower thoracic and lumbar regions of the spinal cord. The concept of CPG is very common in the rehabilitation field. Specifically, we’ll discuss the example of restoring walking in spinal cord injury subjects. The idea is that these patterns generating automatic behaviour are in our spinal cord. Why is this important in the rehabilitation field? For 2 reasons. When we’re on a moving treadmill, if we don’t want to fall down, we’re forced to move and this is not necessarily related to a voluntary intention. In the same way, when we walk over ground, we need to decide to start walking and then as time passes the movement can become rhythmic. This is one of the reasons why when we walk over ground some patterns are different with respect to when we walk on a treadmill. The other reason is that walking over ground is more complex, since we need to bring our weight forward. In the study conducted on people with spinal cord injuries, they considered people that were not able to move without proper assistance and that underwent pharmacological stimulation and rehabilitative exercises in order to walk again. Of course, the approach was studied first in mice. They created artificially the spinal lesion in such mice, then they carried out the pharmacological and stimulation intervention at the level of the spinal cord (a robot was also created to actually train them) and in the end they noticed that the mice were able to walk on the treadmill. So, they were able to activate the CPG. They were not able to make the mice walk independently for a long time though, but then they eventually succeeded also with it. The key point is that the parts that really must start from our brain and the automatic one are 2 important components. Motor primitives and muscle synergies Rich repertoires of complex behaviours are created from the flexible combination of a small set of modules. A locomotor module is a functional unit implemented in a neuronal network that generates a specific motor output by imposing a spatiotemporal structure to muscle activations. Each module involves a basic activation pattern (temporal structure) with variable weights of distribution (spatial structure) to different muscles. The idea is that we can decompose things that have a lot of signals, a lot of degrees of freedom in a subset of combinations that simplifies the evaluation of the problem. If we think about human body, we can for sure state that we are redundant, meaning that we can activate our muscles in a lot of different ways to execute the same action. How does our central nervous system deal with this redundancy? Modular Control Hypothesis One hypothesis is that the CNS doesn't control individual muscles independently but instead organizes them into modules or synergies. These synergies are groups of muscles that are activated together as functional units to simplify control. This hypothesis implies: ​ Dimensionality reduction: The CNS reduces the complexity of controlling many muscles by grouping them into manageable units. ​ Combination and scaling: By combining and scaling these modules, the CNS generates the necessary diversity of movements. Modules are characterized by: 1.​ Temporal activation patterns: When and for how long a module is activated during a task. 2.​ Weight distribution vectors: The contribution of each muscle to the module. Statistical Decomposition vs. Physiological Reality The debate centres on whether the modular organization observed in studies reflects: ​ Physiological control mechanisms: Actual neural strategies used by the CNS. ​ Statistical abstractions: Artifacts of the decomposition methods applied (e.g., non-negative matrix factorization, principal component analysis). Evidence for Physiological Control: ​ Consistency across tasks: Identified synergies often remain consistent across different tasks or individuals. ​ Neurological and biomechanical studies: Neural recordings and electromyography (EMG) data sometimes show patterns that align with synergy-based control. Evidence for Statistical Abstraction: ​ Dependence on methodology: The identified synergies can vary depending on the computational technique used. ​ Alternative explanations: What appears as modular control could emerge from biomechanical constraints or optimization principles rather than explicit CNS control. Two Main Challenges A. Direct Evidence To definitively prove that synergies are a fundamental control strategy, more direct evidence is needed, such as: ​ Neural correlates of synergy control (e.g., specific patterns of neural activity corresponding to synergies). ​ Disruption experiments showing that breaking down a synergy disrupts movement. B. Coordination vs. Control It's possible that: ​ The CNS controls higher-order features (e.g., task goals or joint-level forces), and synergies are emergent properties of this control. ​ Synergies are abstractions that simplify our understanding but do not fully capture how the CNS organizes motor commands. Future Directions ​ Machine learning: Advances in machine learning can improve synergy identification and interpretation, potentially revealing insights into their neural underpinnings. ​ Neurophysiological studies: Combining neural data (e.g., from motor cortex or spinal cord) with biomechanical and computational analyses could clarify the CNS's role in synergy control. ​ Task variability: Investigating synergies across a broader range of tasks, conditions, and individuals could help disentangle physiological mechanisms from statistical artifacts. Elevation angles: looking at the sagittal plane are the angles that foot, shank(stinco) and thigh(coscia) makes with respect to the vertical line. If we plot them in a 3D space, in which each one of the 3 dimensions is represented by one of these angles we find out that for every adult and healthy subject the 3 angles draw the shape highlighted in figure. We move from the toe off to the heel strike to the toe off in a counterclockwise way, always describing this shape. This is called ‘low/planar covariation’, indicating that these angles covariate so that they’re staying on a plane, meaning that the real dimensionality is 2, not 3! PLANAR COVARIATION TELLS US IF THE SUBJECT HAS A GOOD COORDINATION. A study was conducted in order to understand if planar covariation also holds for newborns or not. It actually doesn’t, because newborns are more unstable. Results shown on the right, demonstrated that planar covariation is something that we build through time. Toddler first steps are characterized by a path that is not on the plane. Seven weeks after the situation is already better. 13 months after a planar covariation is almost observable. DO WE CHANGE PATTERNS OR HOW THEY ARE ACTIVATED? Newborns have only 2 patterns. The toddler (bambino) already has 4 patterns, but they’re activated in a different way. Through time, the coordination of the 4 patterns improves though. In adults there are 4 patterns, that in this figure are indicated with different colours. The 4 patterns are combined to give the overall walking pattern. In the middle we have the sticky diagram containing the information about which pattern is more active during the different phases. Key point: we don’t modify the basic patterns, so the 2 that we already have, but we modify how we activate them. Moreover, through life, we learn 2 more patterns. Muscle synergy to investigate motor coordination Muscle synergies have been described as a neural strategy to simplify the control of redundant motor actuator leading human movement. They have been used to investigate motor coordination in healthy subjects and patients with neurological disorders. We can use a lot of muscles when executing an action and we can apply decomposition algorithms (for instance PCA). In the case we consider here they used as decomposition algorithm the non-negative matrix factorization. The principle is that, starting from high dimensionality we get a lower dimensionality, but such low dimensionality is based on variance, so with the basic modules obtained through dimensionality reduction, we can reconstruct the 95-99% of the muscle activation. So, we get the group of muscles that we want to analyse, and we apply non-negative matrix factorization technique. What we get is that we can reduce the large number of muscles to a lower number of blocks that are called ‘muscle synergies’. Each module is characterized by muscles that contribute more to that synergy and less to others and they have a temporal profile in which they activate. Different muscles are represented in different colours and the height indicates how much each muscle contributes to a specific synergy. For instance, in this case, the first synergy will be characterized by muscle m2, while the second one by m1 and m3. The 2 synergies are active in a certain moment of the action and the in the figure we can see the activation profile. By multiplying each synergy by the corresponding activation profile and combining them we can reconstruct the activity of single muscles. In this example we have 4 muscles that can be described with just 2 synergies. So, starting from muscle activity, for each segment we connect each muscle with a specific weight to the activation of that segment on the spinal cord. On the side we have the classical view of a muscle synergy. We have the synergy weights and the activation profile. We have 5 muscles, and the decomposition shows that we’re not controlling independently the 5 muscles but only 3 synergies (3 groups made by such muscles). In the first synergy, as we can see from the synergy weights column, we have the contribution of the first, the fourth and the fifth. In the second synergy, the contribution of second and the fourth. In the third synergy we have the contribution of the first, the second and the fifth. From the basic pattern column, we can see that one synergy is active at the beginning, one in the middle and one in the end. By combining them we get the single muscles’ activations that are visible in the third column. Usually though, we proceed in the other way around, meaning that we start from muscles in order to obtain the subsets. Muscle synergies provide an important information about how muscles coordinate. They found out that for both lower and upper limbs of stroke survivors there are merged synergies, meaning that they cannot be controlled independently and fragmented synergies, meaning synergies that cannot be activated together as it happens normally. Usually, when we analyse muscle synergies, we look at the variability accounted for (as for the principal component analysis). On the right, we have an example of representation (black bars represent the control, dark grey ones the non-paretic leg of stroke survivor and the light grey one the paretic leg). *possible question: how many muscles are active together? They set a threshold, since they observed that with 4 synergies, they’d get almost 100%, meaning that they could really visualize it with 4 synergies. In the photo below we can understand better what these 4 synergies are. There are 8 muscles, since if we analyse gait, it is not worth doing it with less than 8 muscles (typically we should use from 8 to 12 for each leg). On column A we have the representation of the weights. From it, we can see for instance, that the third synergy is dominated by the tibialis, and we also have a little contribution of the rectus femoris. We know that the tibialis is active at the beginning of the gait cycle and this synergy is indeed active at the beginning of the gait cycle. In the second synergy, the active muscles are soleus and medial gastrocnemius. In the first, the active muscles are the vastus medialis, the rectus femoris and the gluteus. In the last synergy instead, the active muscles are hamstrings. 🡺​ So, the first column gives us information about how synergies are made and the second and third columns give us an information about when they are active.Specifically, in the second column we also have information related to when the subject walks at his/her preferred speed. In the third column the different coloured lines represent people walking with a continuously increased speed. As we can see, the module doesn’t change, meaning that the synergies are always the same (independently on the speed). What changes when we go faster is the activation profile. So: THE BASIC MODULES DON’T CHANGE, BUT THE ACTIVATION PROFILES DO WHEN WE CHANGE OUR WAY OF WALKING. Intervento Danilo Muscle synergies in swimming Muscles synergy refers to the activation of different muscles together at the same time with only one neural command from CNS, so they are activated together as functional units to simplify and stabilize movement control. From muscles synergies we obtain two components: weights and activation profiles. ​ Weights are numerical values that represent the contribution of each individual muscle within a particular synergy. ​ Activation profiles are time-dependent signals that describe when and how a muscle synergy is activated during a movement or task. Four swimming styles analyzed both individually and the medley event, that is when an athlete does all styles together in only one race. This was done to see if there were differences between commands of CNS in the two different cases. They used 12 EMG probes on the muscles of interest (8 upper body 8 lower body). They applied the probes to take EMG signals and tried to characterize each style analyzing the stroke in different phases. Butterfly backstroke and freestyle were segmented in different phases using the arm, while breaststroke was segmented using leg. 3 phases for each style: ​ pull phase where the arm is along the body and there is the recruitment, and the athlete brings the arm near the body. ​ push phase where the athlete develops the strength and makes him go through the sprint. ​ recovery phase where the arm is above the water. They obtained 5 synergies for the arm characterized ones and 4 for breaststroke (rana). They were then reordered looking at butterfly activation profile to describe the temporal activation of the motor command. For the medley they performed the algorithm on the concatenated matrix of the 4 styles. Here we have 7 synergies, but the seven one is specific only for breaststroke because involves only the tibialis muscle. The other 6 synergies are in common between the different styles but with different activation profiles: the CNS command is the same but is active in a different time depending on the swimming style. 🡺​ The reason for the different number in the 2 cases was thought to be the fact that the real number for muscle synergies is 6 for butterfly, backstroke and freestyle and 7 for breaststroke, but we obtain a lower number by analyzing them separately because the CNS sends 2 different commands, activates 2 different muscle synergies at the same time resulting in a combination that is a merge of 2 different synergies from the medley resulting in only one. Why? Because the medley includes transitions between different styles (butterfly, backstroke, breaststroke, freestyle), each of which requires specific muscle combinations. This prompts the CNS to generate separate commands for each of these movements. To verify this, it was applied an algo that starting from the medley synergies combined them linearly to try to reconstruct the synergies of each single style. For example, the orange for butterfly has synergies 1 and 6 from the medley that combined form the first one of the butterfly. So, 2 synergies are active simultaneously from 2 different commands at the same time resulting in just 1 synergy for the single style analysis. To be more certain on that it was also computed a similarity coefficient to see how much the reconstructed synergies were similar to the original one, and the coefficients were very close to 1 so it was performed well. 🡺​ So, 7 synergies are needed to describe all the swimming styles, but they are combined if we look at the single style because the CNS sends different commands at same time to perform a very complex motor task like swimming (with respect for example to gait). Robots for gait re-education (from slide gait_2022_3 pag 19) The robots used for the gait re-education are the treadmill and the exoskeleton. The treadmill A method used to set up the rehabilitation procedure is Error Augmentation. We can use training forces that tend to reduce errors (like the robot helps to do the right thing) or to magnify them (robot induces to fail). For the improvement it is better the second one, because with the force one the patient puts less effort. Error reduction leads also to a worse performance than doing nothing. An example is the split-belt treadmill for the lower limb, that allow to reduce the asymmetry between legs (stroke survivor) because the belt for the impaired leg goes with a different speed, depending on the strategy adopted by the patient. It works increasing the assimetry and then going to normal speed the patients works better. Regarding mechanical structures, we can have of two types: Geo and Lokomat. They are expensive devices. Geo: The points of contact are the feet. The patients have lot of freedom because knee and hip are free, but they can’t be left alone. Can just improve the endurance by increasing the compensation strategy (the doctor teaches the pattern). Lokomat: The freedom is lost but you can really control the pattern because you control everything at the level of the joints. It is a wearable robot. Experimental set-up: subject’s data First column is age: they are stroke survivors, so the incidence generally after 65 years increases. D is the number of years from the disease. They were all chronic: this stage causes uncertainty when is defined because someone says it is after 6 months, someone else after 1 year, in general after 1 year the pattern is really stable and the recovery has happened. Then there is gender, ethology that can be ischemic or haemorrhagic stroke, weight, height, paretic side, scale WHS that says how much the patient is impaired, and if they work with a cane (?) or not. Training parameters Training time and walking speed improved. If during one session a patient stays more on the treadmill, he improves his endurance, but also goes faster. There are 3 most common clinical tests: ​ 10MWT is the ten meter walking test, where you walk 10 meters, and you are asked to walk with normal speed or maximum speed. [m/min] ​ 6minWT, you walk for 6 minutes, and they measure how long you walk. [m] ​ TUG, time up and go test, you are sit and you have to stand up and walk, it is more complex.[s] Three periods of evaluation are taken, that are before training, immediately after training and during the follow-up (after 3 or 6 months). ​ Results: ankle power vs gait speed. There is an improvement in most cases, and they are strongly related. We are in sagittal plane, i.e. flexion and extension. First column is the pelvis. Second column is the hip. Third one is the knee. Fourth one is the ankle. On the rows we have the 2 different legs (unaffected and affected side). The 2 lines are pre and post. Bottom row is impaired: we can see knee not flexed, ankle not dorsiplantar flexion, not correct extension movement of the hip. On the non-paretic side after the training there is more improvement, because people in the chronic stage don’t change their strategy but just adjust their compensation strategy. They walk more, little better, but not huge improvement in quality of walk (of life yes). Planar covariation for the two legs of stroke survivors Planar covariation for stroke survivors is still planar in the low. The motion/elevation angle is planar but on the paretic leg the drop that we expect is restricted with respect to the other side. Exoskeletons They can be used both for assistance and rehabilitation. We can have three types:  ​ Exoskeletons like Ekso are rigid structures designed to improve and re learn gait ability after an injury or disease. ​ Other are to assist and empower people that maybe have already normal gait ability and pattern of walking. ​ Other are exosuits that do not have rigid structures, everything is flexible, from sensors to part of the actuation. The motivation in their development comes from military, idea to build the super soldier. From this, they became a scientific technological progress and could also help people with disabilities and target people with difficulty in the normal motion pattern, for rehab or assistance. Generally, they have 2 joints actuated (hip and knee), the ankle in most cases is not actuated, it’s rigid or semi rigid. Then there are sensors at level of the joints (like encoders or goniometers) that measure motions at joint level. There are the motors, the electronic and control parts inside the backpack. On the left there is ReWalk. It has 2 models, one for rehabilitation and one personal model that can be used at home after a specific train. 1.​ ReWalk Rehabilitation 2.0 has a supervised use at Rehab Facilities ​ Universal sizing of ~5'2" to 6'3" in a single unit. It could be a problem because there is a range of height and weight that can be put in an exoskeleton but there are not universal ways because there could be people with sizes that would not fit well. ​ Rapid Exchange from patient to patient ​ Smoother Natural Gait Mode 2.​ Personal Model is for Home & Community Use. ​ Custom fit and programmed to individual ​ Must complete FDA required training program Indego is another example, a “Clinical Indego Kit” consists of 3 sizes (S, M, L), body weight up to 250 lbs, body height from 5’1” to 6’ 3”. Ekso: we can start walking using: ​ The First Step that is like a controller that can be used by the patient or by a therapist. ​ ActiveStep that is a button on the crutches. ​ ProStep that is an active modality, is a calibrated procedure that allows patients, when the exoskeleton starts walking, to control their residual motion. One limitation is that can be used only with crutches or the walker. Also, there is a click that releases noise and bores people a lot. The ankle and foot are really rigid, causing a non-natural way of walking. Problem: the noise is too loud as underline the diversity so border the patients. Three goals in the exoskeleton use: ​ Goal 1: Identification of Clinical Criteria for Safe and Efficient Use of Exoskeletons in Individuals with SCI. ​ Goal 2: Develop Training Strategies for Independent Over-ground Ambulation of Individuals with SCI in a Clinical Setting. ​ Goal 3: Develop Training Strategies to Allow Independent Ambulation of Individuals with SCI at Home and in the Community. Automatic evaluation of exoskeleton use proficiency 3 sensors (accelerometer/gyroscope inside IMUs) monitor trunk, foot and arm motion. Features are extracted from acceleration data. Model is built to quantify expertise of use.  Optimal values allow to normalize each step in order to became really proficient. On the top, you expect that if it was a good model, when the patient starts to use the exoskeleton through sessions, he will start from a value close to 0 and then he will go by increasing the expertise toward to 1. If it is not increasing for a specific subject, maybe the subject has some problem with the use of the device and needs to go to a different therapy to know how to use the device. We can measure walking efficiency using different devices by looking at oxygen consumption. In blue there is the oxygen consumption during weeks using Ekso, there is a decrease, so they became more proficient in using it and developed a more efficient way of walking. KAFO is referred to orthosis for the limb (knee ankle foot orthosis). It provides support at knee ankle foot level being passive, it improves walking stability but with really high oxygen consumption so energy expenditure and really less efficient gait. In both cases practice decreases consumption but for exoskeletons it is really better. Below is the time in which people walk, and number of steps performed through session. Sessions became longer and people did more steps during sessions. Assessment is done to: 1.​ Answer specific clinical questions about a population or a single subject. 2.​ Understand how the CNS adapts the control to a body that changes in a variable environment. 3.​ Understand the effects of a device and optimize its use, so improve the device itself. Different leg muscles are investigated because they are activated when walking with exoskeletons. For the first problem related to clinical questions to answer, we want to see if these devices are useful to train walking abilities for different people (one with complete SCI, one incomplete, one stroke survivor). ×​ It’s not good for people with injury from a long time that have fragile bones, even if the skeleton helps to compensate for the weight and not to load the weight on them. ✔​ For someone who had a recent injury and could not be able to walk anymore, exoskeletons could be good because everything makes them walking is going to be good for 3 reasons: ​ passive motion; ​ psychological reason because they stand up looking at a friend and can walk, even if with assistance, while without psychological motivation they couldn’t walk; ​ the position of our organs when we stand, that is not good. So, if a person with a complete spinal cord injury has a recent injury, exoskeletons can be really good; if instead he has had an injury for a long time, we need to evaluate the state of his bones. For a person with incomplete SCI, they could be good because there is margin for the improvement of muscles strength since he is still able to walk. For the second problem related to understanding how CNS adapts, we have to say that exoskeletons can give bilateral or unilateral assistance. The assistance can be fixed, adaptive, or no assistance. The exoskeleton walking is tested in all these modalities for 10 minutes each one, then they go back to walking without exoskeletons. They found that by assisting with exoskeletons, in these 10 minutes tested there was no adaptation of muscles, so no less effort spent as we would have expected by thinking at upper arm reaching tasks. The question became if we don’t adapt in a short time, or we don’t adapt at all. Also, by comparing maximum assistance and no assistance with exoskeletons, and normal walking, they found that there was a huge difference between exoskeleton and without, because the exoskeleton for its mechanical structure imposes a pattern. This is expla