firstly, if we put in a vertical position a person with a reduced skin sensibility of the foot, like in tabes, or even more easier to find, in diabetes with micro-angiopathies clinically detectable (retina or sole), we observe that the person is well balanced with eyes open, even while turning the head. However, if asked to close his eyes, the subject oscillates and loses his balance. This manoeuvre has been described by Romberg [1] who explains that the vision only cannot ensure balance;

secondly, the subject feels a certain discomfort accompanied by dizziness when standing in front of a monument occupying the whole visual field and having a general inclination of several degrees (as is the case with the Guadalupe Church in Mexico City). This can be explained by the discordance between the visual input having a false reference line and the cutaneous plantar sole input getting the true reference from contact with the ground. We repeated this experiment in our laboratory by placing some subjects in a room without any indication of linear geometric angles or sides, and some others in a completely dark room with a vertical or horizontal light line inclined until the dizziness appears. Physiologically it is equivalent to a typical optical vertigo.

Starting and Stopping

It is a decision that must be made in the premotor area, mobilising also the subcortical structures containing the motor library. Thus, subjects with Parkinson have a clear hesitation; they usually start with a series of short and quick steps. Stopping is also a technical complex procedure since we slow down and have to end up in a stable position. We studied these processes in the INSERM Unit and observed that the brutal stop on an obstacle requires the involvement of many muscles and crossing the obstacle shows the important role of visual guidance. Therefore vision records unconsciously topographical parameters, allowing the movement of the foot, which is not seen by the subject, at 1 cm from the edge of the obstacle in the case of a box 60 cm high. This is the same problem when walking on a rough terrain, where the foot placement must avoid obstacles (emerging stone, depression, …), which is done without too much effort thanks to the anticipation power of visual guidance. The blind replaces it with a guide touch, using a stick more or less sophisticated. In this regard, advanced automation walking is limited in a sequence of obstacles like the steps of a staircase because, as we have seen, having in the sequence a step 2 cm higher (not visible by the subject) leads necessarily to stumbling or falling.

Speed Control

It is a semi-automatic control in the sense that one can voluntarily change its walking speed. Two parameters are changed: firstly, cadence (number of steps) and secondly, the length of the step.

It should be remembered that a step is defined as being the same distance between two supports of the same foot (heel contact of the right and left foot). Some schools talk about a half-step to designate the distance between left and right support. This value can be asymmetric in some lameness. Trained walkers generally adopt a “comfort” rate level in flat ground, but it is interesting to note that there is an automatic regulation of the link cadence/step length, which is manifested in the rise of a slope usually with a decrease of both parameters or of one of them.

The military, normally prepared to obey without discussion, know how to walk in step, with or without music, but in this case, the perfect rhythmic answer demonstrates the power of the rhythm generation imposed by the hearing function. This is also the case for all forms of dance, from classical with strict traditional rules to modern with unbridled rhythmic fluttering and lonely on night clubs (like in a zoo).

Trajectory Control

Apparently simple, it is in fact quite complicated. We have seen that the quadruped performs its track changes by lateral bending of the spine (justifying the role of the epispinatus muscle). This also explains why their humerus and femur have no neck. Human beings in bipedal walking are changing direction with the hip, fully justifying the femoral neck lever operated by many muscle rotators. In reality, the changes are not always of large amplitude, but turning on the spot, which is apparently simple, requires the use of a differential length of left/right steps or vice versa, depending on the direction of rotation. We studied this problem when we produced, with the assistance of the Automation Division of Renault car factory, a “walking machine” operating in master-slave system for paraplegic patients. It was using a pneumatic splint with four hydraulic motors (hip/knee activation) driven by a physiotherapist coated with the same brace and with four potentiometers on hips and knees. In this model, the therapist allowed the patient to walk through appropriate control software. This allowed us to make two interesting observations:

firstly, during embryogenesis, the segmentation of the neural tube in metamers (from Greek meron, the segment) gives rise to a series of spinal nerves. They are formed, because of the separation of sensitive inputs and motor outputs, with the junction of a sensory root with spinal ganglion (the first relay of the sensory pathway) and a motor root containing motor fibres coming from the anterior horn of the spinal cord (alpha and gamma motor neurons);

secondly, the need to link the various segments by a functional interconnection system fully justifies the interneurons. This can be explained by the fact that most muscles have a multi-root innervation requiring to activate at the same time several metameric levels. The spinal cord has also an important reflex activity not only monosegmental but also plurisegmental like the flexor withdrawal reflex of the lower limbs after painful stimulation of the skin which also exists in paraplegic patient or like alternated flexion/extension reflex;

thirdly, this raises the issue of a possible gait pattern spinal generator, ensuring the automatic sequences of gait. This is still subject to controversy, and to understand, it is important to know the specific difference in the neurophysiology of an animal opposed to a man.

firstly, a beheaded duck, frequently experienced when living in the countryside, can do a few steps before collapsing completely;

secondly, in the hands of a good physiologist, a decerebrate cat can still walk on a treadmill;

thirdly, thanks to the French doctor Guillotin, inventor during the French Revolution of a cleaner and more efficient instrument than the axe, we have in France a great experience for the beheading of humans. A headless man standing up and making a few steps has never been observed (except in some exceptional miracle). Nowadays, public beheading is still practised in some countries. A man carrying a sword standing behind a condemned man executes the task, but nobody reported the victim running afterward;

fourthly, the newborn automatic gait described by André Thomas (1952) can be observed by putting a newborn immediately after delivery on his feet. The alternating movements of flexion of the lower limbs observed are in reality not comparable to real walking, but are, as the phenomenon disappears at about the age of 2 months, spinal flexor withdrawal reflexes in connection with an incomplete myelination of the major pathways of the nervous system.

starting out: bipodal support;