In everyday living, humans are continuously challenged to maintain their body balance against the gravitational force. Particularly, when standing quietly in upright stance, we try to keep our body center of mass at a low energy consumption location over the support base, delimited by the positions of our feet, so that a reduced amount of torque at the ankles is enough to stabilize balance. On the contrary, sudden large-scale balance perturbations, such as tripping or slipping, require energy-consumption vigorous and fast muscular activation throughout the body to prevent a fall. At the first exposure to a perturbation, we are usually able to select a response good and strong enough to allow for balance recovery. Our primary response to an unusual postural perturbation, however, is characterized by poorly coordinated and energetically expensive movements. When suffering repeated balance perturbations, responses become more effective and economic, reducing the magnitude of muscular activation and amplitude of limb movements at the same time that the chance of falling is decreased. In my comments, I will explore a perspective distinct from the core idea brought by Shadmehr and Ahmed in Vigor: Neuroeconomics of Movement Control that movement vigor is determined by the amount and rapidity of reward acquisition in relation to the effort expended. Rather, I argue that in postural control our organism selects the vigor of reactive responses to sudden extrinsic perturbations guided by an optimization rule considering two contextual factors. First, the required magnitude of a postural response for balance recovery as indicated by afferent information from a myriad of sensory receptors, and second, the history of previous responses to similar perturbations. In this case, there exists a single relevant and immediately provided reward of achieving successful balance recovery without falling.

As seminal findings supporting my argument, Nashner (Reference Nashner1976) compared feet-in-place postural responses to a short set of repetitive displacements of the support base, either through toes-up rotations or backward translations. Results revealed task-specific adaptation over trials, with progressive reduced plantiflexor muscular activation for support base rotations, and progressive increased activation of the same muscles for support base translations. Both muscular response patterns were adaptive, as indicated by increased balance stability in the last as compared to the first perturbation trial. In this classic study, Nashner showed, thus, that response vigor of the plantiflexor muscles was modulated as a function of the particular task requirements for balance recovery and also by the history of consequences from previous responses. Further support for the notion that reactive postural responses are selected on the basis of functional task requirements and history of postural responses to previous perturbations comes from a recent study we conducted by applying stance perturbations leading to forward body sway with magnitudes of 6, 8, and 10% of body mass (Teixeira et al., Reference Teixeira, Maia Azzi, de Oliveira, Ribeiro de Souza, da Silva Rezende and Coelho2020). Responses were constrained to feet in place reactions to evaluate the effect of load on the magnitude of automatic postural responses at the ankles. One group was exposed to a series of perturbations of progressively increasing load magnitudes, whereas another group faced the opposite decreasing sequence. On the one hand, results showed instantaneous scaling at the very first reactive response to a given perturbation load, with vigor of postural responses corresponding to the magnitude of stance perturbation. On the other hand, the increasing in comparison with the decreasing load sequence led to reduced displacement and velocity of center of pressure under the feet, in parallel with lower activation rate of the agonist plantiflexor muscles. Namely, response magnitude to a given load was decreased or increased depending on whether the previous responses were generated to a lower or to a higher load, respectively. These results indicate that feedback from different sensory receptors signaling fast body sway (e.g., muscle spindles, mechanoreceptors under the feet soles, and vestibular apparatus) guides instantaneously the selection of vigor of postural responses in coherence with the perturbation magnitude, whereas the history of previous responses to a lower or higher load magnitude preset proactively the control system through feedforward processes for down- or up-sizing the postural response vigor.

For repeated similar perturbations vigor of postural responses is diminished, leading to more effective and economic movements. The first perturbation of a sequence has been shown to be featured by excessively strong muscular activation, resulting in exaggerated amplitude of limb and trunk movements (Oude Nijhuis et al., Reference Oude Nijhuis, Allum, Borm, Honegger, Overeem and Bloem2009; Tang, Honegger, & Allum, Reference Tang, Honegger and Allum2012). Interestingly, these strong responses rather than leading to fast recovery of stance stability provokes indeed further balance perturbation. Over repeated perturbations of the same kind, one will see more economic movements associated with increased balance stability. This effect has been explored for improvement of reactive responses in perturbation-based balance training. Results have shown that training reactive balance responses by means of serial perturbations leads to decrement of the following response parameters: (a) hip angular velocity (Krause et al., Reference Krause, Freyler, Gollhofer, Stocker, Brüderlin, Colin and Ritzmann2018), (b) number and/or length of compensatory steps (Mansfield, Peters, Liu, & Maki, Reference Mansfield, Peters, Liu and Maki2010; McIlroy & Maki, Reference McIlroy and Maki1995), and (c) amplitude of arms and trunk displacement (Akinlosotu, Alissa, Sorkin, Wittenberg, & Westlake, Reference Akinlosotu, Alissa, Sorkin, Wittenberg and Westlake2020; Hurt, Rosenblatt, & Grabiner, Reference Hurt, Rosenblatt and Grabiner2011; Takazono, de Souza, de Oliveira, Coelho, & Teixeira, Reference Takazono, de Souza, de Oliveira, Coelho and Teixeira2020). Reinforcing the adaptive value of previous exposure to perturbations for selecting response vigor in future events, research has also shown retention of stability gains over time (König et al., Reference König, Epro, Seeley, Catalá-Lehnen, Potthast and Karamanidis2019; McCrum, Karamanidis, Willems, Zijlstra, & Meijer, Reference McCrum, Karamanidis, Willems, Zijlstra and Meijer2018), and generalizability of gains to contexts different from that specifically experienced during the perturbation-based balance training (Lee, Bhatt, & Pai, Reference Lee, Bhatt and Pai2016; Takazono et al., Reference Takazono, de Souza, de Oliveira, Coelho and Teixeira2020). Through different measurements, then, these findings show that the history of previous experiences with perturbatory events to balance stability prospectively attenuates the vigor of ensuing reactive postural responses through feedforward processes, making them at the same time more economic in energy consumption and more effective for balance recovery. Neurophysiologically, response vigor can be thought to be modulated at two control levels. At the lower level, it has been shown that balance training through self-induced perturbations by standing on an unstable support surface leads to attenuation of fast peripheral reactions in the lower limb, as revealed by decreased excitability of the H-reflex following training (Keller, Pfusterschmied, Buchecker, Müller, & Taube, Reference Keller, Pfusterschmied, Buchecker, Müller and Taube2012; Taube et al., Reference Taube, Kullmann, Leukel, Kurz, Amtage and Gollhofer2007). At higher control levels, perturbation-based balance training has been shown to lead to increased activation of the prefrontal and parietal cortices (Patel, Bhatt, DelDonno, Langenecker, & Dusane, Reference Patel, Bhatt, DelDonno, Langenecker and Dusane2019), which can be thought to underlie adaptive selection and scaling of compensatory responses to unanticipated balance perturbations.

An additional instance supporting the argument that contextual factors guide modulation of vigor of muscular responses can be seen under organismic constraints. In bipedal creatures such as humans, the two legs are coordinated to share the duty of producing muscular forces for maintaining stability of upright balance control. When keeping quiet stance, for example, the two legs share equivalent control responsibilities. In situations that one individual's leg is disabled like in unilateral stroke, the unimpaired leg compensates for the weak responses of the impaired leg in automatic postural reactions to extrinsic stance perturbations (Coelho, Fernandes, Martinelli, & Teixeira, Reference Coelho, Fernandes, Martinelli and Teixeira2019). To study compensatory control between the legs, we recently performed an experiment evaluating reactive lower leg muscular responses to unanticipated forward stance perturbations in the condition that the plantiflexor muscles of one leg only were fatigued (Rinaldin et al., Reference Rinaldin, de Oliveira, de Souza, Scheeren, Coelho and Teixeira2021). Results revealed that a low muscular activation of the fatigued leg when responding to stance perturbations was compensated for by stronger muscular activation of the non-fatigued leg in comparison with the pre-fatigue state. As further findings of interest, we observed progressive decrement of muscular activation in the non-fatigued leg over a series of perturbation trials, and an after–effect featured by conservation of greater muscular activation of the non-fatigued leg following fatigue dissipation. In both instances, our findings revealed feedforward processes as previous fatigue-related responses affected ensuing muscular activation for balance recovery. Additionally, the between-leg compensatory control was observed in the medial and lateral gastrocnemii but not in the soleus muscle, suggesting that vigor of muscular activation was set presumably on the basis of the potential contribution of each individual muscle (because of their structural and functional properties) to the aim of reestablishing upright balance following the specific perturbation employed. These results support the perspective that the vigor of reactive postural responses can be predictively up- or down-sized taking into consideration physiological constraints and memory of previous responses.

As concluding remarks, in my comments to Vigor: Neuroeconomics of Movement Control, I discussed evidence that vigor of reactive postural responses to sudden extrinsic perturbations to stance stability is modulated on the basis of an optimization rule taking into consideration two contextual factors: The required postural response for balance recovery as signaled by sensory information, and the history of previous responses to similar perturbations. From this perspective, I argue that feedback and feedforward processes interact to determine the vigor with which we respond to extrinsic perturbations to body balance stability.