Cue-induced changes in the stability of finger force-production tasks revealed by the uncontrolled manifold analysis

Human movement strategies in uncertain environments: A synergy-based approach to the stability-agility tradeoff

Humans frequently prepare for agile movements by decreasing stability. This facilitates transitions between movements but increases vulnerability to external disruptions. Therefore, humans might weigh the risk of disruption against the gain in agility and scale their stability to the likelihood of having to perform an agility-demanding action. We used the theory of motor synergies to investigate how humans manage this stability-agility tradeoff under uncertainty. This theory has long quantified stability using the synergy index, and reduction in stability before movement transitions using anticipatory synergy adjustment (ASA). However, the impact of uncertainty - whether a quick action should be executed or inhibited - on ASA is unknown. Furthermore, the impact of ASA on execution and inhibition of the action is unclear. We combined multi-finger, isometric force production with the go/no-go paradigm. Thirty participants performed constant force (no-go task), rapid force pulse (go task), and randomized go and no-go trials (go/no-go task) in response to visual cues. We measured the pre-cue finger forces and computed ASA using the uncontrolled manifold method and quantified the spatio-temporal features of the force after the visual cue. We expected ASA in both go/no-go and go tasks, but larger ASA for the latter. Surprisingly, we observed ASA only for the go task. For the go/no-go task, 53% of participants increased stability before the cue. The high stability hindered performance, leading to increased errors in no-go trials and lower peak forces in go trials. These results align with the stability-agility tradeoff. It is puzzling why some participants increased stability even though 80% of the trials demanded agility. This study indicates that individual differences in the effect of task uncertainty and motor inhibition on ASA is unexplored in motor synergy theory and presents a method for further development.

Two aspects of feed-forward control of action stability: effects of action speed and unexpected events

We explored two types of anticipatory synergy adjustments (ASA) during accurate four-finger total force production task. The first type is a change in the index of force-stabilizing synergy during a steady state when a person is expecting a signal to produce a quick force change, which is seen even when the signal does not come (steady-state ASA). The other type is the drop in in the synergy index prior to a planned force change starting at a known time (transient ASA). The subjects performed a task of steady force production at 10% of maximal voluntary contraction (MVC) followed by a ramp to 20% MVC over 1 s, 3 s, and as a step function (0 s). In another task, in 50% of the trials during the steady-state phase, an unexpected signal could come requiring a quick force pulse to 20% MVC (0-surprise). Inter-trial variance in the finger force space was used to quantify the index of force-stabilizing synergy within the uncontrolled manifold hypothesis. We observed significantly lower synergy index values during the steady state in the 0-ramp trials compared to the 1-ramp and 3-ramp trials. There was also larger transient ASA during the 0-ramp trials. In the 0-surprise condition, the synergy index was significantly higher compared to the 0-ramp condition whereas the transient ASA was significantly larger. The finding of transient ASA scaling is of importance for clinical studies, which commonly involve populations with slower actions, which can by itself be associated with smaller ASAs. The participants varied the sharing pattern of total force across the fingers more in the task with "surprises". This was coupled to more attention to precision of performance, i.e., inter-trial deviations from the target as reflected in smaller variance affecting total force, possibly reflecting higher concentration on the task, which the participants perceived as more challenging compared to a similar task without surprise targets.

Integrating Motor Variability Evaluation Into Movement System Assessment

Common assessment tools for determining therapeutic success in rehabilitation typically focus on task-based outcomes. Task-based outcomes provide some understanding of the individual's functional ability and motor recovery; however, these clinical outcomes may have limited translation to a patient's functional ability in the real world. Limitations arise because (1) the focus on task-based outcome assessment often disregards the complexity of motor behavior, including motor variability, and (2) mobility in highly variable real-world environments requires movement adaptability that is made possible by motor variability. This Perspective argues that incorporating motor variability measures that reflect movement adaptability into routine clinical assessment would enable therapists to better evaluate progress toward optimal and safe real-world mobility. The challenges and opportunities associated with incorporating variability-based assessment of pathological movements are also discussed. This Perspective also indicates that the field of rehabilitation needs to leverage technology to advance the understanding of motor variability and its impact on an individual's ability to optimize movement.

IMPACT

This Perspective contends that traditional therapeutic assessments do not adequately evaluate the ability of individuals to adapt their movements to the challenges faced when negotiating the dynamic environments encountered during daily life. Assessment of motor variability derived during movement execution can address this issue and provide better insight into a patient's movement stability and maneuverability in the real world. Creating such a shift in motor system assessment would advance understanding of rehabilitative approaches to motor system recovery and intervention.

Influence of aging on the control of the whole-body angular momentum during volitional stepping: An UCM-based analysis

Evidence suggests that whole-body angular momentum (WBAM) is a highly controlled mechanical variable for performing our daily motor activities safely and efficiently. Recent findings have revealed that, compared to young adults, older adults exhibit larger range of WBAM during various motor tasks, such as walking and stepping. However, it remains unclear whether these age-related changes are ascribed to a poorer control of WBAM with age or not. The purpose of the present study was to examine the effect of normal aging on WBAM control during stepping. Twelve young adults and 14 healthy older adults performed a series of volitional stepping at their preferred selected speed. An Uncontrolled Manifold (UCM) analysis was conducted to explore the presence of synergies among the angular momenta of the body segments (elemental variables) to control WBAM (performance variable); i.e., to stabilize or destabilize it. Results revealed the existence of a stronger synergy destabilizing the WBAM in the sagittal-plane older adults compared to young adults during stepping, while there was no difference between the two groups in the frontal and transversal planes. Although older participants also had a larger range of WBAM in the sagittal plane compared to young adults, we found no significant correlation between synergy index and the range of WBAM in the sagittal plane. We concluded that the age-related changes in WBAM during stepping are not ascribed to alterations in the ability to control this variable with aging.

Humans prioritize walking efficiency or walking stability based on environmental risk

In human gait, the body's mechanical energy at the end of one step is reused to achieve forward progression during the subsequent step, thereby reducing the required muscle work. During the single stance phase, humans rely on the largely uncontrolled passive inverted pendular motion of the body to perpetuate forward motion. These passive body dynamics, while improving walking efficiency, also indicate lower passive dynamic stability in the anterior direction, since the individual will be less able to withstand a forward external perturbation. Here we test the novel hypothesis that humans manipulate passive anterior-posterior (AP) stability via active selection of step length to either achieve energy-efficient gait or to improve stability when it is threatened. We computed the AP margin of stability, which quantifies the passive dynamic stability of gait, for multiple steps as healthy young adults (N = 20) walked on a clear and on an obstructed walkway. Participants used passive dynamics to achieve energy-efficient gait for all but one step; when crossing the obstacle with the leading limb, AP margin of stability was increased. This increase indicated caution to offset the greater risk of falling after a potential trip. Furthermore, AP margin of stability increased while approaching the obstacle, indicating that humans proactively manipulate the passive dynamics to meet the demands of the locomotor task. Finally, the step length and the center of mass motion co-varied to maintain the AP margin of stability for all steps in both tasks at the specific values for each step. We conclude that humans actively regulate step length to maintain specific levels of passive dynamic stability for each step during unobstructed and obstructed gait.

Optimality, Stability, and Agility of Human Movement: New Optimality Criterion and Trade-Offs

This review of movement stability, optimality, and agility is based on the theory of motor control with changes in spatial referent coordinates for the effectors, the principle of abundance, and the uncontrolled manifold hypothesis. A new optimality principle is suggested based on the concept of optimal sharing corresponding to a vector in the space of elemental variables locally orthogonal to the uncontrolled manifold. Motion along this direction is associated with minimal components along the relatively unstable directions within the uncontrolled manifold leading to a minimal motor equivalent motion. For well-practiced actions, this task-specific criterion is followed in spaces of referent coordinates. Consequences of the suggested framework include trade-offs among stability, optimality, and agility, unintentional changes in performance, hand dominance, finger specialization, individual traits in performance, and movement disorders in neurological patients.

Locomotion control during curb descent: Bilateral ground reaction variables covary consistently during the double support phase regardless of future foot placement constraints

During community ambulation, anticipatory adaptations in gait are key for navigating built, populated and natural environments. It has been argued that some instability in gait can be functionally beneficial in situations demanding high maneuverability, and while the mechanisms utilized to maintain locomotor balance are well understood, relatively less is known about how the control of gait stability changes to facilitate upcoming maneuvers in challenging environments. The double support phase may be important in this regard; since both feet can push off the ground simultaneously, there is greater control authority over the body's movement during this phase. Our goal was to identify how this control authority is exploited to prepare for upcoming maneuvers in challenging environments. We used synergy indices to quantify the degree of coordination between the ground reaction forces and moments under the two feet for stabilizing the resultant force and moment on the body during the double support phase of curb descent. In contrast to our expectations, we observed that the kinetic synergy indices during curb descent were minimally influenced by expected foot targeting maneuvers for the subsequent step. Only the resultant moment in the frontal plane showed reduced stability when targeting was required, but the synergy index was still high, indicating that the resultant moment was stable. Furthermore, the synergy indices indicated that the main function of the ground reaction variables is to maintain stability of whole-body rotations during double support, and this prerogative was minimally influenced by the subsequent foot targeting tasks, likely because the cost of losing balance while descending a curb would be higher than the cost of mis-stepping on a visual target. Our work demonstrates the salience of stabilizing body rotations during curb negotiation and improves our understanding of locomotor control in challenging environments.

Expectation of volitional arm movement has prolonged effects on the grip force exerted on a pinched object

Humans closely coordinate the grip force exerted on a hand-held object with changes in the load arising from the object's dynamics. Recent work suggests the grip force is responsive to the predictability of the load forces as well. The well-known grip-force-load-force coupling is intermittent when the load arising from volitional movements fluctuates predictably, whereas grip force increases when loads are unpredictable. Here, we studied the influence of expected but uncertain volitional movements on the digit forces during a static grasp. Young, healthy participants used a pinch grasp to hold an instrumented object and track visual targets by moving the object. We quantified the mean grip force, the temporal decline in grip force (slacking), and the coupling between the pressing digit forces that yield the grip force during static prehension with no expectation of movement, and during the static phase of a choice reaction time task, when the participant expected to move the object after a variable duration. Simply expecting to move the object led to sustained (for at least 5 s) higher magnitude and lower slacking in the grip force, and weaker coupling between the pressing digit forces. These effects were modulated by the direction of the expected movement and the object's mass. The changes helped to maintain the safety margin for the current grasp and likely facilitated the transition from static to dynamic object manipulation. Influence of expected actions on the current grasp may have implications for manual dexterity and its well-known loss with age.

Intramuscle Synergies: Their Place in the Neural Control Hierarchy

We accept a definition of synergy introduced by Nikolai Bernstein and develop it for various actions, from those involving the whole body to those involving a single muscle. Furthermore, we use two major theoretical developments in the field of motor control—the idea of hierarchical control with spatial referent coordinates and the uncontrolled manifold hypothesis—to discuss recent studies of synergies within spaces of individual motor units (MUs) recorded within a single muscle. During the accurate finger force production tasks, MUs within hand extrinsic muscles form robust groups, with parallel scaling of the firing frequencies. The loading factors at individual MUs within each of the two main groups link them to the reciprocal and coactivation commands. Furthermore, groups are recruited in a task-specific way with gains that covary to stabilize muscle force. Such force-stabilizing synergies are seen in MUs recorded in the agonist and antagonist muscles but not in the spaces of MUs combined over the two muscles. These observations reflect inherent trade-offs between synergies at different levels of a control hierarchy. MU-based synergies do not show effects of hand dominance, whereas such effects are seen in multifinger synergies. Involuntary, reflex-based, force changes are stabilized by intramuscle synergies but not by multifinger synergies. These observations suggest that multifinger (multimuscle synergies) are based primarily on supraspinal circuitry, whereas intramuscle synergies reflect spinal circuitry. Studies of intra- and multimuscle synergies promise a powerful tool for exploring changes in spinal and supraspinal circuitry across patient populations.

Recent Advances in the Neural Control of Movements: Lessons for Functional Recovery

We review the current views on the control and coordination of movements following the traditions set by Nikolai Bernstein. In particular, we focus on the theory of neural control of effectors - from motor units to individual muscles, to joints, limbs, and to the whole body - with spatial referent coordinates organized into a hierarchy with multiple few-to-many mappings. Further, we discuss synergies ensuring stability of natural human movements within the uncontrolled manifold hypothesis. Synergies are organized within the neural control hierarchy based on the principle of motor abundance. Movement disorders are discussed as consequences of an inability to use the whole range of changes in referent coordinates (as in spasticity) and an inability to ensure controlled stability of salient variables as reflected in indices of multi-element synergies and their adjustments in preparation to actions (as in brain disorders, including Parkinson's disease, multiple-system atrophy, and stroke). At the end of the review, we discuss possible implications of this theoretical approach to peripheral disorders and their rehabilitations using, as an example, osteoarthritis. In particular, "joint stiffening" is viewed as a maladaptive strategy, which can compromise stability of salient variables during walking.

One more time about motor (and non-motor) synergies

We revisit the concept of synergy based on the recently translated classical book by Nikolai Bernstein (On the construction of movements, Medgiz, Moscow 1947; Latash, Bernstein's Construction of Movements, Routledge, Abingdon 2020b) and progress in understanding the physics and neurophysiology of biological action. Two aspects of synergies are described: organizing elements into stable groups (modes) and ensuring dynamical stability of salient performance variables. The ability of the central nervous system to attenuate synergies in preparation for a quick action-anticipatory synergy adjustments-is emphasized. Recent studies have demonstrated synergies at the level of hypothetical control variables associated with spatial referent coordinates for effectors. Overall, the concept of synergies fits naturally the hierarchical scheme of control with referent coordinates with an important role played by back-coupling loops within the central nervous system and from peripheral sensory endings. Further, we review studies showing non-trivial changes in synergies with development, aging, fatigue, practice, and a variety of neurological disorders. Two aspects of impaired synergic control-impaired stability and impaired agility-are introduced. The recent generalization of the concept of synergies for non-motor domains, including perception, is discussed. We end the review with a list of unresolved and troubling issues.

Sensitivity of the Toe Height to Multijoint Angular Changes in the Lower Limbs During Unobstructed and Obstructed Gait

Tripping while walking is a main contributor to falls across the adult lifespan. Trip risk is proportional to variability in toe clearance. To determine the sources of this variability, the authors computed for 10 young adults the sensitivity of toe clearance to 10 bilateral lower limb joint angles during unobstructed and obstructed walking when the lead and the trail limb crossed the obstacle. The authors computed a novel measure-singular value of the appropriate Jacobian-as the combined toe clearance sensitivity to 4 groups of angles: all sagittal and all frontal plane angles and all swing and all stance limb angles. Toe clearance was most sensitive to the stance hip ab/adduction for unobstructed gait. For obstructed gait, sensitivity to other joints increased and matched the sensitivity to stance hip ab/adduction. Combined sensitivities revealed critical information that was not evident in the sensitivities to individual angles. The combined sensitivity to stance limb angles was 84% higher than swing limb angles. The combined sensitivity to the sagittal plane angles was lower than the sensitivity to the frontal plane angles during unobstructed gait, and this relation was reversed during obstacle crossing. The results highlight the importance of the stance limb joints and indicate that frontal plane angles should not be ignored.

Stability of Action and Kinesthetic Perception in Parkinson's Disease

We present a review of action and perception stability within the theoretical framework based on the idea of control with spatial referent coordinates for the effectors at a number of hierarchical levels. Stability of salient variables is ensured by synergies, neurophysiological structures that act in multi-dimensional spaces of elemental variables and limit variance to the uncontrolled manifold during action and iso-perceptual manifold during perception. Patients with Parkinson’s disease show impaired synergic control reflected in poor stability (low synergy indices) and poor agility (low indices of anticipatory synergy adjustments prior to planned quick actions). They also show impaired perception across modalities, including kinesthetic perception. We suggest that poor stability at the level of referent coordinates can be the dominant factor leading to poor stability of percepts.

The Influence of Recent Actions and Anticipated Actions on the Stability of Finger Forces During a Tracking Task

The authors examined how the stability of the current total isometric force (FT) produced by four fingers is influenced by previous and expected voluntary changes in FT. The authors employed the synergy index obtained from the across-trial uncontrolled manifold analysis to quantify the stability of FT. The authors compared two tasks with similar histories of FT changes; one in which participants expected changes in FT in the future, and one in which they expected no changes in FT. The stability of FT was lower in the former task, indicating the existence of a novel type of anticipatory synergy adjustment. Disparate histories of FT changes yield inconsistent changes in stability, driven by individual differences in the covariation in the finger forces that leave FT invariant. Future research should focus on exploring these individual differences to better understand how previous and expected behavior changes influence the stability of the current motor behavior.

The coordination between digit forces is altered by anticipated changes in prehensile movement patterns

Stability is the ability of a system to maintain a desired static or dynamic motor pattern. Maneuverability, on the other hand, is the ability to transition between motor patterns, and it is antagonistic to stability. Animals frequently reduce the stability of an ongoing task to facilitate anticipated movement transitions. Such stability-maneuverability tradeoffs are observed in human locomotion. However, the notion applies to other behaviors and this paper reports the first study on the stability-maneuverability tradeoff in human prehension. We tested if the coordination between the digit forces during the manipulation of a hand-held object is altered in response to an expected change in the manipulation pattern. We focused on the coupling between the grip and the load force and between the opposing forces exerted by the thumb and the four fingers, and on the transition from rhythmic vertical oscillation to non-vertical oscillation of the object. The nature of these couplings depends on the oscillation direction. Therefore, the stability-maneuverability tradeoff predicts that an expected volitional change to the object's movement will diminish the strength of these couplings so that the force patterns generating the current movement can efficiently transition into new ones that generate the new movement. The strength of the grip-load coupling did not alter in tasks that required a change in movement compared to tasks that did not. We speculate that participants preferred safety over maneuverability and maintained the grip-load coupling strength to counter high inertial loads and avoid object slip. In contrast, the strength of the coupling between the thumb and the four fingers' opposing forces reduced in tasks that required a change in movement compared to tasks that did not. Thus, the stability-reduction aspect of the stability-maneuverability tradeoff occurs in prehensile behavior. Future work should focus on associating the reduction in stability with gains in maneuverability, and on developing a comprehensive account of this tradeoff in prehensile tasks.

Speed impacts frontal-plane maneuver stability of individuals with incomplete spinal cord injury

BACKGROUND

Following incomplete spinal cord injury, people often move slowly in an effort to maintain stability during walking maneuvers. Here we examine how maneuver speed impacts frontal-plane stability in people with incomplete spinal cord injury. We hypothesized that the challenge to control frontal-plane stability would increase with maneuver speed; specifically, the minimum lateral margin of stability would be smaller and the required coefficient of friction to avoid a slip would be greater during fast vs. preferred speed maneuvers.

METHODS

We measured kinematics and ground reaction forces as 12 individuals with incomplete spinal cord injury performed side-step, lateral maneuvers at preferred and fast speeds. We examined four sequential steps: the Setup and Pushoff steps initiated the maneuver, and the Landing and Recovery steps arrested the maneuver.

FINDINGS

Our hypotheses were partially supported. Maneuver time was shorter during fast vs. preferred speed maneuvers (p = 0.003). Minimum lateral margin of stability was smaller during the Setup step of fast vs. preferred speed maneuvers (p = 0.026). We found no differences in minimum lateral margin of stability between speeds for the Landing and Recovery steps (p > 0.05). The required coefficient of friction was not different between fast and preferred speed maneuvers (p = 0.087).

INTERPRETATION

The greatest effect of increasing maneuver speed occurred during the Setup step; as speed increased, participants reduced their minimum lateral margin of stability ipsilateral to the maneuver direction. This action allowed maneuvers to be performed more quickly without requiring a greater lateral impulse during the Pushoff step. However, this strategy reduced passive stability.

Preparation to a quick whole-body action: control with referent body orientation and multi-muscle synergies

We examined the control of postural stability in preparation to a discrete, quick whole-body sway toward a target and back to the initial position. Several predictions were tested based on the theory of control with referent body orientation and the notion of multi-muscle synergies stabilizing center of pressure (COP) coordinate. Healthy, young adults performed fast, discrete whole-body motion forward-and-back and backward-and-back under visual feedback on the COP. We used two methods to assess COP stability, analysis of inter-trial variance and analysis of motor equivalence in the muscle activation space. Actions were always preceded by COP counter-movements. Backward COP shifts were faster, and the indices of multi-muscle synergies stabilizing COP were higher prior to those actions. Patterns of muscle activation at the motion onset supported the idea of a gradual shift in the referent body orientation. Prior to the backward movements, there was a trend toward higher muscle co-activation, compared to reciprocal activation. We found strong correlations between the sets of indices of motor equivalence and those of inter-trial variance. Overall, the results support the theory of control with referent coordinates and the idea of multi-muscle synergies stabilizing posture by confirming a number of non-trivial predictions based on these concepts. The findings favor using indices of motor equivalence in clinical studies to minimize the number of trials performed by each subject.

Stability and predictability in human control of complex objects

Previous research on movement control suggested that humans exploit stability to reduce vulnerability to internal noise and external perturbations. For interactions with complex objects, predictive control based on an internal model of body and environment is needed to preempt perturbations and instabilities due to delays. We hypothesize that stability can serve as means to render the complex dynamics of the body and the task more predictable and thereby simplify control. However, the assessment of stability in complex interactions with nonlinear and underactuated objects is challenging, as for existent stability analyses the system needs to be close to a (known) attractor. After reviewing existing methods for stability analysis of human movement, we argue that contraction theory provides a suitable approach to quantify stability or convergence in complex transient behaviors. To test its usefulness, we examined the task of carrying a cup of coffee, an object with internal degrees of freedom. A simplified model of the task, a cart with a suspended pendulum, was implemented in a virtual environment to study human control strategies. The experimental task was to transport this cart-and-pendulum on a horizontal line from rest to a target position as fast as possible. Each block of trials presented a visible perturbation, which either could be in the direction of motion or opposite to it. To test the hypothesis that humans exploit stability to overcome perturbations, the dynamic model of the free, unforced system was analyzed using contraction theory. A contraction metric was obtained by numerically solving a partial differential equation, and the contraction regions with respect to that metric were computed. Experimental results showed that subjects indeed moved through the contraction regions of the free, unforced system. This strategy attenuated the perturbations, obviated error corrections, and made the dynamics more predictable. The advantages and shortcomings of contraction analysis are discussed in the context of other stability analyses.

Expectation of movement generates contrasting changes in multifinger synergies in young and older adults

Anticipatory synergy adjustment (ASA) is a feed-forward control mechanism that describes a continuous decrease in the stability of the current motor state beginning about 150 ms prior to a state transition. Recently, we described an associated phenomenon in which the system stability was reduced solely in response to a cue that generates an expectation of a state change, independent of whether the state change actually occurs. Both phenomena are of the same kind (stability reduction), but evoked by distinct antecedent conditions. Since, logically, cuing for movement must occur before the initiation of that movement, we named this new phenomenon 'Stage-1 ASA' and rechristened the well-established version 'Stage-2 ASA'. Here, we used a four-finger, isometric force production task to explore (1) the effect of healthy aging on Stage-1 ASA, and (2) if Stage-1 ASA resulted in a more rapid state change. Young and older adult participants produced 10% of their maximal force when they did not expect to produce any change in the force, and when they expected to change their force in an unknown direction and at an unknown time. In the latter condition, the 10% constant-force phase was followed by a choice reaction time task, in which the participants rapidly changed their force to follow a moving target presented on a computer monitor. Both young and older adults displayed equivalent amount of Stage-1 ASA. This was driven by a 42% reduction in finger-force variability in young adults. In contrast, it was driven by a 38% increase in finger-force variability in older adults. We speculate that the reduction in finger force variability assists the young adults in rapid state changes via two mechanisms: (1) the finger forces occupy a restricted set of states that are optimal for quick state transitions, and (2) lower variability during steady state translates into lower self-motion during state transition. Self-motion is the covariation between finger forces that fails to change the total force. The older adults are unable to adopt this strategy, and the increase in finger-force variability arises from (1) the adoption of an alternative strategy of destabilizing the attractor associated with the current state to facilitate state transitions and (2) the inability to coordinate multiple finger forces. Finally, older adults displayed longer reaction times than young adults, but a clear relation between Stage-1 ASA and consequent behavioral benefit in terms of reduced reaction time remained elusive.

Stability of vertical posture explored with unexpected mechanical perturbations: synergy indices and motor equivalence

We explored the relations between indices of mechanical stability of vertical posture and synergy indices under unexpected perturbations. The main hypotheses predicted higher posture-stabilizing synergy indices and higher mechanical indices of center of pressure stability during perturbations perceived by subjects as less challenging. Healthy subjects stood on a force platform and held in fully extended arms a bar attached to two loads acting downward and upward. One of the loads was unexpectedly released by the experimenter causing a postural perturbations. In different series, subjects either knew or did not know which of the two loads would be released. Forward perturbations were perceived as more challenging and accompanied by co-activation patterns among the main agonist-antagonist pairs. Backward perturbation led to reciprocal muscle activation patterns and was accompanied by indices of mechanical stability and of posture-stabilizing synergy which indicated higher stability. Changes in synergy indices were observed as early as 50-100 ms following the perturbation reflecting involuntary mechanisms. In contrast, predictability of perturbation direction had weak or no effect on mechanical and synergy indices of stability. These observations are interpreted within a hierarchical scheme of synergic control of motor tasks and a hypothesis on the control of movements with shifts of referent coordinates. The findings show direct correspondence between stability indices based on mechanics and on the analysis of multi-muscle synergies. They suggest that involuntary posture-stabilizing mechanisms show synergic organization. They also show that predictability of perturbation direction has strong effects on anticipatory postural adjustment but not corrective adjustments. We offer an interpretation of co-activation patterns that questions their contribution to postural stability.