Bipedal Steps in the Development of Rhythmic Behavior in Humans

Possible Benefits of Synchronized Movements

Walking in pace represent one form of synchronized behavior. Our article will only explore a small selection of synchronized animal behavior. A review by Wilson and Cook (2016) explores, among others, possible benefits of synchronized movements across species in more detail. Investigating the evolution of (human) behavior is inherently difficult, because most behaviors leave little to no fossil evidence. All species are related to different degrees. Based on these phylogenies, one can compare human behaviors with those of other species, in particular our closest living relatives, the nonhuman primates.

Focusing on a branch of the phylogenetic tree further away from humans, Larsson (2012b) proposed that synchronized locomotion in fish could be adaptive. Fish swimming together produce overlapping and confusing acoustical signals likely resulting in confusion of the lateral line of predatory fish (Larsson, 2009). Schooling fish have the capacity to interrupt movements simultaneously providing quiet intervals, hence improving predator detection (Larsson, 2009). Animals of a similar size moving almost concurrently produce similar sounds, facilitating auditory grouping (Larsson, 2012a; Popper & Fay, 1993). Accordingly, the vertebrate brain may have become pre-programmed to develop synchronized behavior in appropriate ecological niches (Larsson & Abbott, 2018).

Wilson and Cook (2016) stated other possible advantages of moving in a synchronized manner: for example, animals swimming in choppy water may synchronize their limb strokes to the frequency of the waves, and a group of individuals coordinating movements may be more efficient in capturing preys and frightening a potential predator (Merker, 2000). Positive social interactions due to synchronization have been observed in macaques (Nagasaka, Chao, Hasegawa, Notoya, & Fujii, 2013). Synchronous movement has emotional impact on humans (McNeill, 1995): synchronization of movements in a group may be a potent way of creating and sustaining community and communication. A recent study (Cirelli, Einarson, & Trainor, 2014) has shown that 14-month-old children who were bounced to the beat of music played by an adult were more likely to retrieve an object dropped by the adult than were children bounced without keeping time to the beat (see also Cirelli, Wan, & Trainor, 2014; Cirelli, Wan, & Trainor, 2016; Trainor & Cirelli, 2015). The fact that synchronised movements may facilitate communication is, in principle, consistent both with the “bipedal experience in utero” and the “acoustical advantage” hypothesis.

Human Walking Sound

Human walking displays long-term constancies (Dingwell & Cusumano, 2010; Hausdorff et al., 1996). In unhindered, over-ground walking, regularities can be found in stride time, pace length, and speed (Terrier, Turner, & Schutz, 2005). Walking and running are periodic activities, with a single period known as the gait cycle. By definition, the gait cycle begins when one foot comes into contact with the ground and ends when the same foot again contacts the ground and is comprised of stance and swing phases (Novacheck, 1998). Human walking rates are generally in the range of 75 and 125 steps per minute (Sabatier & Ekimov, 2008). In laboratory studies as well as during long periods of unconstrained locomotor activity, the preferred cadence of walking is around 120 steps per minute (MacDougall & Moore, 2005).

In walking, the two initial portions of the stance phase, the initial contact and the loading response, normally produce more sound energy than other stance phase segments, although their combined duration is less than 10% of the gait cycle (Novacheck, 1998). In other words, the major amplitude component of the step is distinctly produced, and in that sense, it may resemble a beat. Footfall is likely to produce substantial self-perception of sound due to bone conduction (Moore, 2003), especially when running. During barefoot running at 4 m/s on a hard surface, the magnitude of the peak of the GRF is between 1.5 and 2.5 times the body weight. This sends a shock wave up the body that can be measured in the head within about 10 ms (Lieberman et al., 2010).

Walking sound of others conveys information to listeners about the sound source, and listeners learn to draw conclusions based on the features of the sound (Visell et al., 2009) about properties of the ground surface (Giordano et al., 2012), the gender (Li, Logan, & Pastore, 1991), and the posture of the walker (Pastore, Flint, Gaston, & Solomon, 2008).

Human locomotion may influence, and interact with, emotions, and locomotion sounds may be influenced by the emotion of the walker (Giordano & Bresin, 2006). Sounds produced on a firm surface lead to more aggressive walking patterns (Bresin, de Witt, Papetti, Civolani, & Fontana, 2010). Runners were shown to alter step length and, thereby, speed when presented with music of different emotional character (“relaxing” or “activating”), while retaining a pace of 130 beats per minute (BPM) (Leman et al., 2013).

Walking and running produce rhythms in the range of 75–190 BPM. Humans can synchronize walking movements with music over a broad spectrum of tempos, but synchronization is optimal in a narrow range around 120 BPM (Styns, van Noorden, Moelants, & Leman, 2007). Music is often played at a tempo similar to that of walking (Changizi, 2011). The tempo of popular dance music peaks at around 120–130 BPM (Leman et al., 2013). In healthy individuals attempting to walk in time to a metronome at 120 BPM, the average pace was 119.52 ± 3.12 steps per minute, suggesting a good potential to synchronize steps with rhythmic auditory sounds (Bilney, Morris, Churchyard, Chiu, & Georgiou-Karistianis, 2005). In brief, we have seen that humans skilfully extract information from footsteps, there is an emotional impact associated with footsteps, and the optimal synchronization to a metronome is around 120 BPM; all this seems consistent with both the “bipedal experience in utero” and the “acoustical advantage” hypothesis.

Walking Together

Walking side by side, people often unconsciously synchronize steps, suggesting that the perception of one’s partner directly influences gait in the absence of conscious effort or intent (Nessler & Gilliland, 2009; Zivotofsky & Hausdorff, 2007). When two individuals stroll on neighboring treadmills, their walking patterns tend to be substantially coordinated (Nessler & Gilliland, 2009). Each person makes fine adjustments to locomotion kinematics in order to adapt to their partner’s behavior (Nessler et al., 2013). In paired walking, participants can be phase-locked with a phase difference close to 0° (in phase), or they can be phase-locked with a phase difference close to 180° (antiphase or antisynchrony) with walkers contacting the ground simultaneously with opposite-side feet (Nessler & Gilliland, 2009; Nessler, Gonzales, Rhoden, Steinbrick, & De Leone, 2011). Walking in phase or antiphase is likely to produce a similar overall acoustical pattern and rhythm. These synchronous and antisynchronous patterns seem particularly common in humans, together with a few phylogenetically distant species (Ravignani, 2015). Similar leg length is significantly related to locking of step, and the level of frequency locking was not shown to significantly differ with variation in visual and auditory information, suggesting that minimal sensory information such as mechanical vibrations caused by the partner’s steps may induce unintentional synchronization (Nessler & Gilliland, 2009). Data on synchronization in runners are lacking.

Synchronization in the Evolution of Musical Rhythm?

Human dance and music are closely associated with moving in synchrony. This raises the question: did the human tendency and ability to move in synchrony evolve before or after the evolution of music? If the human preference to move in synchrony evolved first, what possible benefits may early humans have achieved by moving in synchrony? A related question is – why have apes, our closest relatives, not developed a similar tendency to move in synchrony (with synchrony intended in its more restrictive sense, e.g. (Ravignani, 2017) rather than loose coordination)? Some have hypothesized that groups of hominids walking in synchrony may have confused and frightened their enemies through mimicry of a large animal (Merker, 2000). The “acoustic advantages due to bipedal walking” hypothesis suggests that two or more humans walking in synchrony likely achieve a short period of silence in the middle of the step cycle, improving auditory awareness of their environment (Larsson, 2014). Larsson, Ekstrom, and Ranjbar (2015) showed that the masking potential of two individuals’ footsteps was reduced when walking in pace compared with not in pace although the decibel level was identical. While walking in such manner, the ability to perceive differences in pitch, rhythm, and harmonies could help the hominid brain to distinguish sound sources and facilitate synchronization of movements. Such attentive listening in nature in association with rhythmic group locomotion may have resulted in reinforcement possibly through dopamine release (Larsson, 2014). A primarily survival-based behavior may eventually have attained similarities to dance and music (Meehan, Abbott, & Larsson, 2017).

Associations Between Walking and Music

There may be an association between human locomotion and music/dance (Larsson, 2014). For example, passive listening to music, or imagining it, activates areas of the brain associated with motor behavior (Chen, Zatorre, & Penhune, 2006; Grahn & Brett, 2007). The connection between auditory and motor systems is important for the execution of rhythmic movements in humans, and music as structured auditory input has a remarkable ability to drive rhythmic, metrically organized, motor behavior such as dance (Zatorre, Chen, & Penhune, 2007). Meehan et al. (2017) proposed that music listening may mimic the sense of walking with other people, which might contribute to the positive effects of music and dance therapy in Parkinson’s disease (see also Pereira et al., 2019).

Both dance and music are extremely complex behaviors. Therefore, rather than thinking about one particular advantage from which they originated as a whole, it may be better to think about how adaptive their constituent components are. Particular traits of music and music-like behavior might have been advantageous and have had a more influential role under specific circumstances. For example, early mother-infant interaction (Dissanayake, 2000) motherese (Falk, 2004), and coalition signaling (Hagen & Bryant, 2003) have been discussed.

Leg Movements in Dance

Perception of music, particularly in regards of rhythms, can be studied, among others, through the theoretical framework of embodied cognition (Grahn & Brett, 2007; Penhune & Zatorre, 2019; Su, 2016a, 2016b). Even in non-embodied frameworks, rhythm perception and movement are tightly linked neurobiologically (Grahn & Brett, 2007; Penhune & Zatorre, 2019).

Given this auditory-motor coupling, rhythm information derived from music may be represented and retained in the brain as information about body movements (Konoike et al., 2012). Sensorimotor mechanisms similar to musical rhythm perception have been shown when human observers perceive dance movement (Su, 2016a, 2016b). Observers extract a visual beat from regular movement patterns of body parts, with leg movements most often chosen as the primary beat and vertical trunk movement occupying the space between the beats. When the four limbs move in tandem, visual beat perception in dance relies mainly on the pattern of leg movement (Su & Salazar-Lopez, 2016). In practice, across dance genres it is the footwork that is most often performed in time to the musical beat (Y. H. Su, personal communication in 2 April 2018). There are many dances, among them many African dances, German Schuhplattler, tap dance, Zapateado Peruano, or Flamenco, in which footwork is used to provide percussion.

Down-on-the-Beat Movements in Dance

Internal noise such as vibrations in the skeleton produced by walking, dance, and other movements is scarcely investigated. However, it is likely that down-on-the-beat movements appear more salient because they cause more internal sound to vibrations in the skeleton (Moore, 2003). A study of auditory-motor entrainment in street dancers (Miura, Kudo, & Nakazawa, 2013) showed that in rhythmic knee bending to the beat, up-on-the-beat (knee extension on the beat) was unintentionally replaced by down-on-the-beat (knee flexion on the beat) at high movement frequency. It may be that, in terms of motor control, knee flexion is biomechanically easier to perform than knee extension. Thus, synchronizing the former to an auditory target is a more economical and more natural form of movement. This action preference seems to be mapped in perception, as observers also perceive the downward trajectory of the knee-bending movement as more congruent with an auditory downbeat than upward movement (Su, 2014). Through training, dancers acquire the skill to perform the more demanding up-on-the-beat movement, which may require internal synchronizing to the beat in an antiphase manner. Footsteps in locomotion tend to be accompanied by moving the body downward, which largely generates the sound. Thus, we hypothesize that the downward movement is more coupled to the beat than is the upward movement, both in action and in perception.

Sensory Experience In Utero

Perception of sound and rhythm in utero has been suggested to influence the individual’s development of musical abilities (Parncutt, 1993; Parncutt & Chuckrow, 2017). Brain development is largely shaped by early sensory experience (Figure 1) (Webb, Heller, Benson, & Lahav, 2015). For instance, prenatal sensory experiences may influence taste preference of offspring in humans (Schaal, Marlier, & Soussignan, 2000) and other animals. Fetuses are exposed to scents of the mother’s diet, which influences taste preferences after birth (Schaal et al., 2000). Neonates seem to remember the smell of amniotic fluid, which attracts them more than other cues (Tyzio et al., 2006). Full-term infants indicate orientation of sound by turning the head towards the source. If they are shown an object at the same time, they will move their gaze to the sound, implying that hearing is more mature than vision at birth (Lagercrantz, 2014; Wilkinson & Jiang, 2006). Visual perception is unlikely to be experienced in utero, while sound is a primary source of varied and consistent stimulus to the developing brain (Teie, 2016).

OPEN IN VIEWER

Figure 1.

A pregnant woman.

The Fetal Acoustic Environment and Associations With Musical Elements

It has been proposed that features of music correspond to sounds that are present in the womb, and that the fetal acoustic environment may provide the basis for the fundamental musical elements found in the music of all cultures (Parncutt, 1987, 1993; Parncutt & Chuckrow, 2017; Teie, 2016). Although the role of footfalls in rhythm development has been previously discussed (Parncutt, 1987, 2009), heartbeat and pulse have been more often considered in this context (Teie, 2016; Ullal-Gupta, Vanden Bosch der Nederlanden, Tichko, Lahav, & Hannon, 2013).

Fetuses of all mammals perceive maternal heartbeat to some extent, thus perception of heartbeat in utero has little explanatory value with respect to musical ability as a strictly human phenomenon. In other words, chimpanzee and human fetuses will hear a similar sound from the mother’s heart. Thus, that fetal experience relating to heartbeat will not contribute to explain the difference between human and nonhuman primates in regards of musical and rhythmical skill. Instead, we argue for a relatively more influential role of maternal footfall.

Sound In Utero

Maternal vocalization—empirical work has shown that the intrauterine sound levels of the mother’s voice are enhanced by an average of 5.2 dB, while external voices are attenuated by roughly 2 to 3 dB (Richards, Frentzen, Gerhardt, McCann, & Abrams, 1992). From week 32 to birth, the fetus is likely to be exposed to around 6.7 × 10⁵ words (Mehl, Vazire, Ramirez-Esparza, Slatcher, & Pennebaker, 2007). Full-term fetuses detect and respond to the mother’s voice (Voegtline, Costigan, Pater, & DiPietro, 2013). Fetuses’ hearts may respond to the mother’s voice from week 32–34 (Kisilevsky & Hains, 2011). Maternal tissues filter out many aspects of speech; for instance, the fetus probably cannot distinguish fricatives. Nonetheless, features like prosody and rhythm are transmitted, and biases are present for acoustic patterns in the prosody of languages spoken by the mother (Abboub, Nazzi, & Gervain, 2016).

Maternal heartbeat is the most common rhythmic auditory stimulus, occurring 24 hours a day resulting in a stream of more than 5 × 10⁶ maternal heartbeats between week 32 and birth (Teie, 2016). The average frequency of healthy females is 81.7 BPM (Patrick, Campbell, Carmichael, & Probert, 1982), quite close to the pace that is common in music. The sound is 25 dB above basal noise from the external environment (Querleu, Renard, Versyp, Paris-Delrue, & Crepin, 1988). Magnetoencephalographic recordings show that the fetus responds to the sound of the mother’s heartbeat with a latency of around 200 ms (which is a generally accepted marker of fetal response to auditory stimuli) (Porcaro et al., 2006).

Maternal breathing is audible in utero (Parncutt, 2016). The resting mother’s breathing creates a rhythm of 12–20 BPM (Barbosa Pereira et al., 2017). In a walking mother, breathing may be coupled to footfall, since locomotion and respiration are frequently coupled (Bramble & Carrier, 1983; Funk, Milsom, & Steeves, 1992).

Maternal bowel sounds are not particularly rhythmic but are likely to be perceived and contribute as background sound (Teie, 2016).

Maternal footfalls may represent a major contribution to in utero sound. Human uninhibited walking cadence is approximately 120 steps per minute (MacDougall & Moore, 2005), a common musical tempo (Changizi, 2011). The fetus will move rhythmically to the mother’s audible footsteps (Parncutt, 2016). Footfalls will be heard during limited periods. Notice that a fetus moving rhythmically to the mother’s footsteps is consistent with whole-body involvement in the dance of many cultures (e.g., Fitch, 2016; Miura, Kudo, Ohtsuki, & Kanehisa, 2011). However, alternative theoretical frameworks have also explained the role of whole-body movement in dance (e.g., in terms of action simulation, imagery, embodiment, etc., see Bartenieff & Lewis, 2013; Cross, Hamilton, & Grafton, 2006; Hanrahan, Tetreau, & Sarrazin, 1995; Miura et al., 2013).

Testable Hypotheses

Useful hypotheses need to be testable. Although the hypotheses presented here are preliminary, we believe it is important to suggest empirical avenues to support or refute speculations. One possibility would be to investigate whether children of non-walking mothers show differences in musical and general rhythmic abilities (H1). Do offspring of mothers who experienced severe incapacitating girdle pain show selectively reduced rhythmic abilities but intact musical skills in melody, harmony, and timbre (H1)? Does this even extend to grammatical linguistic (dis)advantages (Gordon et al., 2015)? Across geographical areas, does average human leg length correlate with preferred tempo, as potentially predicted by our hypothesis? How do the recently-found genetic correlations between rhythm and walking or breath connect with the literature we review here (Niarchou et al., 2019)? Although it is an everyday experience that dyads of similar leg-lengths tend to walk in pace, the scientific study of human paced walking is scarce. For example, we were unable to find even a single empirical study about perceptual mechanisms of militaries walking in pace. A future empirical question is how militaries (vs. non-militaries) use senses to achieve common pace. Is it hearing or vision or combination of both that is used (H2)? Future zoological research may map the number of limbs to locomotion patterns, and hence sounds produced, across species. Systematically mapping the bio-physics of animal gaits and the rhythmic patterns they produce will allow testing our hypothesis in a comparative framework (H1 and H2). Applied to other species, our hypothesis (H2) predicts that pronking (a gait of quadrupeds, involving jumping high into the air by lifting all four feet off the ground simultaneously) animals should be more sensitive to, or prefer, isochronous patterns, while horses should be more sensitive to quaternary patterns (H1 and H2). It would also be of interest to study the role of incidental sound of locomotion in synchronized animal groups, such as the flapping of wings in bird and bat flocks, or the sound of body movements and breathing in moving cetaceans (H2).