Listen to your feet

I run three-to-four times a week, sometimes five: just short 5 km runs in the morning, nothing too strenuous, sometimes 10 km. And although I love listening to music, at work and at home, I run with only the ambient sounds around me. This appears to be unusual, given the number of people I see running with earbuds firmly stuck in their ears. Although this is fine – yet potentially dangerous if you can't hear vehicles approaching as you run across roads – what does strike me is the sound these runners make: slap, slap, slap go their feet with every step. Even clomp, clomp, clomp. I'm not only talking about amateur (if not keen) runners like myself but also seasoned runners (you can tell by their T-shirts: ‘Finished Whatever Marathon or Triathlon Event in 2022, 2023, 2024’). Why is this an issue? Well, for a start, if you can hear their feet, then the footfall is rather heavy: one should be able to run silently, such as my own stealth mode in which I can feel my feet gently contacting the ground and making little sound. Why is this important? Well, in my opinion, silent running means that your stepping cycle is more efficient – making contact with the ball of your feet and pushing off by plantarflexing the ankle. I don't study the biomechanics or the sensorimotor control of locomotion, but it seems reasonable to me that a light footfall would be more efficient in terms of energy expenditure. So, is there anything in what we know about locomotion, and the roles of feedforward and feedback mechanisms in its neural control, that could support this idea? Running has developed in many quadrupeds, as well as in bipeds such as ourselves and flightless birds (the ostrich and emu), and the adopted running style of certain lizards. The ostrich is the fasted bipedal land animal and can run twice as fast as humans, even when both are running for their lives (or for that trophy!). Unlike humans the ostrich has only two long toes and the emu three; these are connected to long, compliant tendons that store elastic energy. The human foot is a more complicated structure, and although there certainly are biomechanical benefits to us running barefoot, given that it promotes making contact first with the ball of the foot and promotes use of the elastic energy stored in the structures of the arch – we are constrained by having very short toes. As such, we do not have the advantage of the forward propulsion from the long toes of the emu and ostrich. And although not as long as in the emu and ostrich our Achilles tendon does store elastic energy, making the propulsive plantarflexion generated by the triceps surae muscles more efficient. It is worth considering the possibility that, perhaps, the complicated composition of the human foot (with multiple bones and ligaments and four layers of intrinsic muscles) has necessitated a greater reliance on feedback mechanisms to make our running more efficient. Indeed, we have rich sources of proprioceptive feedback from the feet and the long dorsiflexor and plantarflexor muscles acting on the feet, and it is this feedback that we should be ‘listening’ to – in the figurative sense – when we run. And although it is known that running cadence can match the beat of the music, if the runner does engage with the rhythm (Van Dyck et al., 2021), I argue that the normal sensory feedback from the feet and legs is being overridden when people run to the sound of music, in which feedforward mechanisms are favoured. Sensory feedback from the feet is very important; we know that postural sway increases following occlusion of blood flow to the foot, which blocks sensory input from muscle and cutaneous afferents (Mauritz & Dietz, 1980), and we also know that proprioceptive feedback from the muscles acting on the ankle is more important than vestibular inputs in controlling upright stance (Fitzpatrick et al. 1994). Moreover we know that spindle afferents in the long muscles acting on the ankle can encode the small length changes in these muscles associated with postural sway (Day et al., 2017; Peters et al., 2017). We also know that human gait becomes very ataxic when muscle spindles are congenitally absent (Macefield et al., 2024), and that the brain is required to integrate proprioceptive feedback from the limbs for optimal locomotion (Santuz et al., 2022). My laboratory recently developed the approach of performing microelectrode recordings from the posterior tibial nerve at the ankle, which has allowed us to identify the firing properties of muscle spindle endings in the intrinsic muscles of the foot, and of cutaneous mechanoreceptors in the sole, during unsupported standing (Knellwolf et al., 2025). The location of the posterior tibial nerve as it runs around the medial malleolus means that one can obtain stable intraneural recordings while participants are seated with the leg pendant, during standing and even during imposed increases in postural sway. We showed that the responsiveness of muscle spindle afferents in the short muscles of the foot – which have very short or negligible tendons – to stretch and related joint movements was similar to that of muscle spindle afferents located in the intrinsic muscles of the hand, in that they responded often in more than one axis of rotation. Only 27% were spontaneously active in the unloaded condition, a similar proportion to those in the hand (30%; Burke et al., 1988), whereas 50% were active during unsupported free-standing (Knellwolf et al., 2025). Moreover, the firing rates of 67% of the endings were correlated with changes in the centre of pressure (CoP), primarily (88%) along the anteroposterior axis during standing (Knellwolf et al., 2025). Indeed some of these were in muscles that span the arch of the foot (abductor hallucis, flexor hallucis brevis and quadratus plantae). These are known to increase their activity during standing, thereby facilitating elastic energy storage in the tendon, and to actively control the shape of the arch – acting in a manner similar to that of active suspension in a car (Kelly et al., 2014, 2019). And although Golgi tendon organ afferents (GTOs) were not encountered in our study (all afferents responded to muscle stretch), there is certainly evidence for a role of GTOs in encoding changes in load forces during locomotion (Dietz & Duysens, 2000). By inserting the microelectrode into cutaneous fascicles of the posterior tibial nerve we could also assess how mechanoreceptors in the sole of the foot behaved in unsupported free standing. The firing of these afferents depended on receptor type and location of receptive field. Cutaneous mechanoreceptors with small receptive fields (type I) – the fast-adapting (FA I) and slowly adapting SA I afferents – which have a high density on the toes, responded when contact was made with the substrate on standing, but in the loaded condition these two types of tactile afferent also responded to spontaneous adjustments in posture (Knellwolf et al., 2025). Conversely the SA II endings – slowly adapting receptors with large receptive fields that are sensitive to skin stretch – could, like muscle spindles, encode fluctuations in CoP through changes in their ongoing discharge in the loaded condition (Knellwolf et al., 2025). These data emphasise the important contributions of sensory sources in the foot to maintaining and responding to perturbations in upright posture, and it is likely that they also provide useful information during running. It is well known that each limb goes through four phases during locomotion – (i) the support phase, (ii) the lift-off phase, (iii) the swing phase and (iv) the touch-down phase – and it is also known that these elements are controlled by central pattern generators (CPGs) within the spinal cord, as evidenced by maintenance of this basic pattern following complete dorsal rhizotomy and, therefore, loss of somatosensory input to the spinal cord (Grillner & Kozlov, 2021). How one runs is determined by top-down control of spinal CPGs, with specific interneurons in the spinal cord (V1 neurones) determining the speed (Gosgnach et al., 2006). These CPGs have been preserved through evolution and receive sensory information from the periphery on every phase of the step cycle. However the voluntary initiation of locomotion, and the cadence and speed, is very much cortically driven, with the motor cortex also contributing to the ongoing activity of human walking (Petersen et al., 2012). Moreover specific neurones in the brainstem control the direction of locomotion towards the left or right (Cregg et al., 2020). And although crabs avoid the complexities of co-ordinating four legs on each side by walking sideways, quadrupeds co-ordinate the forelimbs and hindlimbs to suit the pattern of locomotion. Moreover, during human locomotion co-ordination of the arms and legs is organized in a way similar to that in the cat: corticospinal excitation of upper-limb motoneurones is mediated indirectly via propriospinal neurones in the cervical spinal cord, which allows a task-dependent neuronal linkage of cervical and thoracolumbar propriospinal circuits that control leg and arm movements (Balter & Zehr, 2007; Dietz, 2002). Nevertheless, as I have tried to highlight above, sensory feedback to the CPGs and to the brain plays important roles in controlling the complex patterns of flexor and extensor muscle activity, as well as of other muscles engaged during locomotion. Some of this feedback comes from muscle spindles and Golgi tendon organs, which detect changes in muscle length and tension, respectively, as well as from mechanoreceptors in the skin that detect the changes in skin strain around the joints and, to some extent, from receptors within the joint capsules themselves (Grillner & Kozlov, 2021; Nielsen, 2016; Nielsen & Sinkjaer, 2002). It is the sensory feedback during the lift-off and, especially, the touch-down phase that is critical: tripping over an object during lift-off or contact with an irregular substrate during the touch-down phase needs rapid correction to avoid falls or otherwise interfere with locomotion. We are currently starting to record from muscle and cutaneous afferents from the foot during human treadmill walking, and although it will be impossible to do this during running, there is no doubt that both muscle spindle afferents from the intrinsic muscles of the foot and cutaneous afferents from the sole of the foot provide important contact and phase-related information during running. In my opinion running while listening to music engages feedforward mechanisms in which you are trying to align your stride to the tempo of the music, perhaps leading to some degree of desynchronization of sensory feedback from your legs and feet with the motor commands. So, rather than listening to the beat, listen to your feet. It seems to me that if you can't hear your feet as you run then everything is working as it should, and that your feet are engaging the substrate in a manner that optimises sensory feedback. In other words, for those Lord of the Rings fans, run like Legolas, not like Gimli. Or, for the sci-fans out, there adopt ‘Silent Running.’ Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. The author has no competing interests. V.M. conceived or designed the work; drafted the work or revised it critically for important intellectual content; final approval of the version to be published; agreement to be accountable for all aspects of the work. All authors have approved the final version of the manuscript and agreed to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed. No funding was obtained.