The coordination between posture and movement is known to be a key factor in motor performance. Gait initiation, which is the transient phase between quiet standing posture and ongoing walking, is a classical model used in the literature to investigate this coordination (e.g., Mann et al., 1979; Brenière et al., 1987; Yiou et al., 2007; Honeine et al., 2016). It is composed of a postural phase preceding the swing foot-off, which corresponds to the “anticipatory postural adjustments (APA)”. The postural phase is followed by an execution phase ending when the peak of forward center of gravity (COG) velocity (or speed performance) is reached (Brenière et al., 1987; Lepers and Brenière, 1995). During these APA, the forward propulsive forces required to reach the peak COG velocity are generated by an anticipatory backward center of pressure (COP) shift. The larger this shift, the greater the speed performance (Brenière et al., 1987; Lepers and Brenière, 1995). The relationship between APA and speed performance during gait initiation illustrates the biomechanical concept of “Posturo-kinetic capacity (PKC)” (Bouisset and Zattara, 1987; Bouisset and Do, 2008), according to which the motor performance of any motor task (in terms of speed, force or precision) depends on the capacity of the central nervous system to generate appropriate APA. According to this concept, any factors that would impair (or conversely enhance) APA development may impair (or enhance) the motor performance. This PKC concept was substantiated by experimental studies which investigated the relationship between postural chain mobility, APA and motor performance during various motor tasks such as isometric ramp push (Le Bozec and Bouisset, 2004), pointing (Lino et al., 1992; Teyssèdre et al., 2000), and more recently, trunk flexion (Diakhaté et al., 2013) or sit-to-stand (Diakhaté et al., 2013; Alamini-Rodrigues and Hamaoui, 2016; Hamaoui and Alamini-Rodrigues, 2017a,b). In these studies, postural chain mobility was varied by changing the seat-thigh contact (Teyssèdre et al., 2000; Le Bozec and Bouisset, 2004; Diakhaté et al., 2013), by increasing the muscular tension along the torso (Hamaoui et al., 2004, 2011; Hamaoui and Le Bozec, 2014) or by restraining the spine mobility at different levels by means of splints (Alamini-Rodrigues and Hamaoui, 2016; Hamaoui and Alamini-Rodrigues, 2017a,b). These studies showed that the restriction of the postural chain mobility, and especially the spine mobility, has a negative influence on APA and motor performance. Conversely, according to the PKC concept, enhancing the postural chain mobility should have a positive influence on these parameters. Besides this purely mechanical influence, APA associated with stepping initiation are also known to be finely tuned to the continuous proprioceptive (Ruget et al., 2010) and cutaneous inflow (Do and Gilles, 1992; Ruget et al., 2008) arising from the postural body segments. Perturbations of this sensory inflow, e.g., by reducing the plantar support or by vibrating the ankle muscles, have been shown to alter APA and motor performance.

“Spinal Manipulative Therapy High-Velocity, Low-Amplitude” (SMT-HVLA) is a healing technique applied to the spine that has been used for centuries by healthcare practitioners including Osteopaths, Chiropractors and Physiotherapists to relieve symptomatic patients from acute and chronic low back/neck pain and/or to improve spine mobility (Wiese and Callender, 2005). As such, SMT-HVLA may have the potential to improve the PKC and thus motor performance. As stressed in the literature (e.g., the review of Pickar and Bolton, 2012), a number of sustained changes in the spinal biomechanics have been thought to occur as a result of SMT-HVLA. For example, the impulsive thrust delivered during the manipulation may alter the segmental biomechanics by releasing trapped meniscoïds, releasing adhesions, or by diminishing distortion in the intervertebral disc. In addition, recent studies reported relaxation of paraspinal muscles following SMT-HVLA as revealed with decreased electromyographic (EMG) activity (DeVocht et al., 2005; Lehman, 2012). Increased spine mobility might result from such changes in the spinal biomechanics and/or EMG activity. Interestingly, this technique is nowadays widely used by healthy athletes (runners, footballers, sprinters etc.) just before a competition in order to reach their “peak performance” (Leonardi, 1994). However, it must be noted that the effect of SMT-HVLA on the articular free play is still controversial (for review see Millan et al., 2012a), with mitigated results on sports performance (Miners, 2010). Shrier et al. (2006) compared jump height and running velocity with and without pre-event SMT-HVLA in elite healthy athletes. These authors found that there was no significant effect of SMT-HVLA on the countermovement jump height and sprint times. However, they also stressed that the direction and magnitude of the observed changes were consistent with a clinically relevant performance enhancement. A similar conclusion was stated by Humphries et al. (2013) with regard to the immediate effect of lower cervical spine manipulation on handgrip strength and free-throw accuracy of asymptomatic basketball players. These authors reported a slight increase in free-throw percentage, which according to them, deserved further investigation.

Besides the potential increase in spine mobility, movement kinematics may also be potentially influenced by neurophysiological changes induced by SMT-HVLA. For example, studies on the anesthetized cat have shown that spinal manipulation induced changes in the discharge of somatosensory afferents from the paraspinal region (Pickar, 2002; Pickar and Bolton, 2012; Reed et al., 2015), including those afferents innervating muscle spindles, Golgi Tendon Organs and high threshold mechanoreceptors. There are currently no unequivocal data regarding whether SMT-HVLA activates nociceptors. In humans, changes in the sensori-motor pathways following SMT-HVLA have been reported in the literature, but with sometimes contradictory results. For example, studies using the Hoffman reflex (H-reflex) technique indicated that spinal manipulation induced a decreased motoneuronal excitability in asymptomatic subjects (Murphy et al., 1995; Dishman and Burke, 2003) and in low back pain patients (Suter et al., 2005), while Niazi et al. (2015) indicated, on the contrary, an increased excitability. Data collection and data analysis methodology of the H-reflex have been evoked by these latter authors to explain this discrepancy with the literature. At the cortical level, it seems that there exists a consensus concerning the alteration of the sensorimotor processing and sensorimotor integration following spinal manipulation, as evidenced with the somatosensory-evoked potential technique (e.g., Haavik-Taylor and Murphy, 2007; Taylor and Murphy, 2008; Haavik Taylor and Murphy, 2010; see “Discussion” Section on this aspect).

As stressed in the literature (e.g., Pickar and Bolton, 2012), the extent to which these mechanical and neurophysiological responses to spinal manipulation reflect beneficial outcomes (e.g., pain relief or enhanced spine mobility) remains unclear. However, each of these responses has the potential to induce changes in the coordination between posture and movement, which strongly relies on both sensory inputs from the postural limbs and postural joint mobility as stressed above. The present study, therefore, aimed to investigate the short-term effect of SMT-HVLA on spine mobility, APA and speed performance during gait initiation in young healthy adults. We first hypothesized that a SMT-HVLA applied to the ninth thoracic vertebra (T9) will increase the spine range of motion and facilitate APA development in the gait initiation paradigm, which is known to involve spine mobility (e.g., Ceccato et al., 2009). Second, we also assumed that the various short-term neurological effects of this manipulation may either improve or reduce the PKC and task performance.

This study investigated the effect of SMT-HVLA manipulation applied to T9 on spine ROM and on the biomechanical organisation of gait initiation. Participants purposely performed series of gait initiation trials before and after sham or SMT-HVLA manipulations. Spine ROM and classical biomechanical gait initiation parameters were compared in the pre- and post-manipulation conditions.

The results showed that the spine ROM, and especially the maximal thoracic flexion, was larger post-manipulation than pre-manipulation in the HVLA group, which was not the case in the sham group. The mean increase was 20%. The HVLA manipulation applied to T9 had, therefore, a beneficial effect on spine mobility, even in a group composed of young adults with no known spine pathologies. A similar result was found in the sole study to date that tested the effect of thoracic HVLA manipulation on spine ROM (Schiller, 2001). However, this study only examined right and left thoracic lateral flexion using a goniometer (BROM 2), and the population tested included patients with thoracic back pain. To our knowledge, the other studies focusing on thoracic HVLA measured the cervical ROM, and reported small or no beneficial effects (González-Iglesias et al., 2009; Suvarnnato et al., 2013). These negligible variations have led us to exclude the cervical section from spine ROM measurement.

SMT-HVLA has been shown to have a beneficial effect on spine-related pain, both clinically and in experimentally-induced pain (for reviews see Millan et al., 2012a,b). However, it remains unclear from the literature if it has an immediate noticeable biomechanical effect on spinal motion that can be measured in terms of an increased ROM (Millan et al., 2012a). One of the main goals of healthcare practitioners applying SMT-HVLA manipulation is to increase spine ROM, e.g., in athletes before competition or in patients with spine stiffness. A recent review on this aspect emphasized that some studies found spinal manipulation to have limited effect on the ROM, while others found it had none (Millan et al., 2012a). These mitigated effects could probably stem from many factors such as the different tools used in research and in clinical practice to measure ROM (single/double/triple inclinometers, goniometers, a rangiometer, tape measures, visual estimation, spine motion analyzer, etc.), the direction, duration and force applied to the participant’s spine, the expertise of the practitioner etc. The present study shows that analogical inclinometers may be used to detect an increased spine ROM following thoracic SMT-HVLA.

Postural mobility, and especially spine mobility, is known to be a key factor of the PKC (Bouisset and Zattara, 1987; Bouisset and Do, 2008). According to the PKC concept, any factor that may enhance (or conversely, alter) this capacity would favor (or hinder) the motor performance and postural stability. This concept has been substantiated by many recent experimental studies which manipulated spine mobility using various means, e.g., by the application of splints at different levels of the spine (Alamini-Rodrigues and Hamaoui, 2016), by experimentally-induced trunk muscular tension (Hamaoui et al., 2004, 2011; Hamaoui and Le Bozec, 2014), or by changing the contact surface between the thighs and seat in the sitting posture (Lino et al., 1992; Teyssèdre et al., 2000; Le Bozec and Bouisset, 2004; Diakhaté et al., 2013). It has been shown that constraining spine mobility may disturb postural equilibrium when maintaining erect posture as revealed by measuring the COP. In dynamical tasks, such as sit-to-stand (Diakhaté et al., 2013; Alamini-Rodrigues and Hamaoui, 2016), maximal isometric ramp push (Le Bozec and Bouisset, 2004), arm pointing (Lino et al., 1992; Teyssèdre et al., 2000) or trunk flexion (Diakhaté et al., 2013) from the sitting posture, facilitating spine mobility has been shown to favor APA development and thus motor performance. Based on the results of these studies—and given that spine mobility is known to be highly solicited during locomotion and gait initiation (e.g., Thorstensson et al., 1984; Ceccato et al., 2009; Cusin et al., 2017), APA development and motor performance could have been expected to be facilitated following SMT-HVLA. It is also noteworthy that T9 is described as the “walking vertebra”, a concept arising from the classical article of Wernham (1985). This concept is based on the fact that the T9 vertebra is the inflexion point of the curvature change of thoracic cyphosis in lumbar lordosis. This vertebra ensures the junction of the thoracic and lumbar segments, mainly in their counter-rotation movement, especially while walking. As such, the T9 HVLA manipulation is commonly used by healthcare practitioners in patients with locomotor deficiencies. However, its impact on the locomotor function has to date never been evaluated in systematic studies. In contrast to our expectations, APA amplitude and duration decreased following manipulation in the HVLA group, by 24% and 6%, respectively. This was not the case in the sham group, which shows that this result could not be ascribed to a placebo effect. Not only were the APA parameters reduced, but their efficiency (computed as the ratio [COG velocity at foot-off/APA duration]) was reduced (by 23% as compared to the pre-manipulation condition). In other words, the capacity of the postural system to generate forward propulsive forces during the limited duration of APA was less efficient post-manipulation. As a consequence of the lower initial (foot-off) COG velocity, the peak COG velocity (speed performance) and step length both reached lower values post-manipulation in the HVLA group (compared to the pre-manipulation condition, the decrease was 14%, 12% and 11%, respectively). This finding was expected since it is well-known that the two latter step parameters are positively correlated with the amplitude of APA, i.e., the higher the peak anticipatory backward COP shift is, the higher the speed performance and step lengths are (Brenière et al., 1987). Because there was no change in APA parameters post-manipulation in the sham group, step length and speed performance remained the same as in the pre-manipulation condition. Because in the present study, spine mobility was increased following the HVLA manipulation, which is known to be a factor of improved motor performance, the question arises as to why APA development and speed performance were impaired instead of being improved.

Besides its mechanical effect on spine mobility (for reviews see Pickar, 2002; Millan et al., 2012a), SMT-HVLA is known to induce transient changes in the sensorimotor pathways and structures involved in the coordination between posture and movement. As stressed in the “Introduction” Section, studies using the H-reflex technique to investigate the effect of SMT-HVLA on motoneuronal excitability reported controversial findings, i.e., both an increased (Niazi et al., 2015) and a decreased excitability (Murphy et al., 1995; Dishman and Burke, 2003; Suter et al., 2005) have been found. It seems however that there exists a consensus concerning the effects of spinal manipulation on the sensorimotor processing and integration at the cortical level, as evidenced with the somatosensory evoked potential technique (SEP). Specifically, recent studies reported an alteration of the amplitude of the cortical SEP peaks N20 and N30 following SMT-HVLA (Haavik and Murphy, 2012; Lelic et al., 2016). The N20 peak is known to represent the arrival of the afferent volley at the primary somatosensory cortex (Desmedt and Cheron, 1980; Nuwer et al., 1994; Mauguière, 1999), while later peaks such as the N30 SEP peak are thought to reflect early sensory-motor integration (Rossi et al., 2003; regarding the possible generators of this peak, see Haavik and Murphy, 2012; Lelic et al., 2016). In addition, studies in the anesthetized cat have shown that spinal manipulation induced changes in the discharge of mechanoreceptors from the paraspinal region, especially group Ia spindle afferents (Pickar, 2002; Reed et al., 2015). The extent to which the cortical and afferent responses to spinal manipulation reflect beneficial outcomes (e.g., pain relief), remains largely unclear in the literature; however, what is clear, is that these responses have the potential to induce changes in the coordination between posture and movement, which is known to strongly rely on proprioceptive inputs from the postural limbs, and even more importantly, on how the CNS processes, interprets and transforms this afferent information into motor commands (Paulus and Brumagne, 2008; Haavik and Murphy, 2012). Lelic et al. (2016) recently speculated that since spinal manipulation is known to reduce pain and improve function in clinical trials (Botelho and Andrade, 2012; Mieritz et al., 2014; Schneider et al., 2015), the observed reduction of the N30 amplitude following SMT-HVLA may reflect a beneficial change. However, it should be noted (as the authors did) that reduced N30 SEP peak amplitudes have also been found in the literature in pathological populations such as Parkinson’s disease (Cheron et al., 1994; Kang and Ma, 2016), known to have deficits in APA production during both voluntary lower (e.g., Delval et al., 2014) and upper limb tasks (e.g., Bazalgette et al., 1987). Kang and Ma (2016) even reported that frontal N30 status indicated the motor severity of Parkinson’s disease. During gait initiation, disturbances in Parkinson’s disease include reduced APA and abnormal APA timing (Delval et al., 2014). During arm elevation, postural movements are known to be less anticipatory in Parkinson’s patients than in controls (Bazalgette et al., 1987). In the present study, APA were also less anticipatory, had a smaller amplitude and were less efficient in the HVLA group following manipulation than prior to manipulation. Globally taken, the results from the literature may thus suggest that a reduction of the N30 amplitude after HVLA manipulation may reflect a transitory alteration in the cortical integration of sensory-motor information, and may thus reflect a negative change. If so, such alteration has the potential to affect motor coordination during locomotor tasks such as gait initiation. In other words, we propose that a neural effect, possibly mediated by a transient alteration in the early sensory-motor integration following SMT-HVLA could be one of the mechanisms responsible for the present results. This neural effect may have masked the potential mechanical benefits associated with increased spine mobility.

There are some limitations to the present study that should be pointed out. First, this study only focused on a biomechanical investigation. It is clear that studies linking the changes in motor behavior observed in the present study, to the changes of activity in the neural structures and pathways reported in the literature should be carried out to further substantiate the data interpretation. This is why we used the term “pilot” in the title of this article. Second, it should be emphasized that only short-term effects were investigated. It is not excluded that thoracic HVLA manipulation may have a long-term beneficial effect on APA and speed performance. Third, the biomechanical responses described in this study were obtained from young healthy participants and may not be generalizable to other populations, including patients with spinal pain. Finally, it is known that a manipulation is rarely specific to only the adjustment site (Ross et al., 2004). This non-specificity is amplified by the technique used in this article as it is an indirect technique. There is no direct contact of the practitioner with the chosen vertebra since the compressive force is indirectly transmitted by the hand of the patient between his own vertebrae and the thorax of the practitioner. We point out that the role of T9 vertebra and the interest of its manipulation is based solely on empirical knowledge although these notions are still taught in physiotherapy and osteopathy schools. Currently, some studies suggest that the center of rotation of the thoracic zone in the horizontal plane corresponds to a very wide area (T7–L3; Konz et al., 2006).

The present results showed that thoracic HVLA manipulation in young healthy participants has an immediate beneficial effect on spine mobility but a detrimental effect on APA development and speed performance during gait initiation. It thus seems that HVLA manipulation should be considered with caution by participants who seek an immediate increase of speed performance during locomotor tasks.

SD and EY designed the study; collected, analyzed and interpreted the data; drafted and revised the manuscript; gave final approval. AH and AD interpreted the data; drafted and revised the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling Editor currently co-hosts a Research Topic with one of the authors EY, and confirms the absence of any other collaborations. He states that the process met the standards of a fair and objective review.