• cognition
• development
• locomotion
• movement
• space
• trajectories

This doctoral thesis has a wide scope, resulting from a research effort made in three directions: to understand innate motor behaviour, to observe the shaping of voluntary motor actions through development, and to trace back the entanglement between movement and cognition in locomotor navigation. As it usually occurs in scientific investigation, the result was by no means comparable with the effort, but nonetheless all worth it.

The observation of spontaneous motor activity as a diagnostic tool

In the first place, I have been involved as a child neurologist in the early diagnosis and prognosis of motor disorders. A deep change of paradigm has taken place in neonatal and infant neurology in the last thirty years, from adult-derived, analytical examination protocols to a more age-specific, global and qualitative approach: Prechtl's method on the qualitative assessment of general movements (GMs) (Cioni, Ferrari & Prechtl 1989; Einspieler et al. 2005a). This has become a clinical gold standard, in particular for the early diagnosis of cerebral palsy, thanks to its high reliability and non-invasiveness (Prechtl et al. 1997). Chapter 2 in this thesis is an account of the state of the art of the methodology, its clinical application and functional meaning (Cioni et al., in press).

Motor development from spontaneous variability to goal-oriented optimisation

The speculative interest of GM assessment is of no less momentum than its clinical power, for at least two reasons.
Firstly, on the examiner's side, it provides a crystal-clear instance of how the global and qualitative judgement can outperform the analytical and quantitative one, when properly trained. This is an intensely pursued and beloved result of GM history, whose expectation dates back to when Heinz Prechtl was taught the power of Gestalt perception by his mentor, Konrad Lorenz (1959).
Secondly, on the infant's side, it definitely points out the richness of innate motor repertoire, its strict correlation with brain well-being and with the quality of later development. It must be stressed, and it will be never enough, that the power of GM assessment, and in general of Gestalt perception, relies on the global, phenomenological appreciation of movement quality, not on the analysis of particular items and much the less on movement quantity. What contributes the most to a positive judgement is the variability of motor patterns, i.e. their neverending variation in shape, speed, force, anatomical distribution and sequential organisation. In other words, normal GMs can be regarded as a sort of motor babbling, featuring the widest possible range of variation in almost all kinesiological features. On the opposite side, severely abnormal GMs feature a complete lack of variation.
In this perspective, GM assessment is in perfect agreement with the theory of neuronal group selection (Edelman 1977; Sporns & Edelman 1993): voluntary, goal-oriented motor actions would emerge from the selection of successful spontaneous motor patterns, by means of an on-going trial-and-error sensorimotor loop. Variability in innate motor behaviour (or, maybe better, "variation", as proposed by Hadders-Algra in a recent review, 2010) should be then considered the foundation of the later flexibility and optimisation of motor solutions. As explained in the next paragraph and more in detail in Chapter 3, one of the key signs of optimal motor coordination in adults is kinematic stereotypy, which, in a sense, is the opposite of variation. Stereotypy is however only one side of the coin, whereas the other is flexibility: only their integration allows to select the best solution to attain a goal (optimisation).
In conclusion, the best imaginable motor repertoire in a mature individual is made up of accurately selected, and therefore stereotyped, motor patterns, integrated in a vast repertoire of coping solutions. On the other hand, an "optimality concept" was coined by Prechtl in 1980, to indicate the best imaginable neurological condition at birth, which can be now identified with the best possible quality of GMs, i.e. the maximal amount of variation. The "optimal" motor development is then a journey from "optimal" variation to "optimal" selection.

The formation of motor trajectories: a bridge between different spaces

Edelman's theory of neuronal group selection is a possible answer to the question raised by Nikolaj Bernstein in 1935 (Bernstein 1967, first English translation): how does the brain cope with the huge number of degrees of freedom offered by a multi-articulated body (Sporns & Edelman, 1993)? The degrees of freedom problem is still central in modern movement science. Two simple observations are at its basis: 1) more than one motor signal can lead to the same spatial trajectory of a given end-effector, e.g. a hand (redundancy); 2) different end-effectors, e.g. a hand, a foot or the whole body, can easily produce topologically identical trajectories, even those body parts that were not associated with that action during training (motor equivalence). In other words, it is not the biomechanical chain to be encoded at the highest level of control, but the Gestalt of the motor action (Bernstein, 1967).
This raises a yet unsettled question: is there a role for spatial cognition in motor coordination, or are movements produced by a direct stimulus-response coupling? The former hypothesis has been adopted by Alain Berthoz and his co-workers at the Collège de France, with whom I have collaborated in the second and third parts of my thesis. Their description of the close analogies between hand- and locomotor trajectories, from both a geometric and a kinematic point of view, unveils a common computational module for movements of such different biomechanical implementations (Vieilledent et al. 2005; Hicheur et al. 2005b). Two fundamental principles govern the formation of locomotor trajectories in adults: 1) the walking direction is always anticipated by head- and eye-orientation (Grasso et al. 1998b; Bernardin et al. 2012), and 2) trajectory geometry and kinematics are highly stereotyped in a world-centred spatial reference frame (Hicheur et al. 2007). Both findings point to the existence of a top-down, feed-forward control loop for locomotor optimisation, strongly related to the cognitive representation of the external space (Pham et al. 2007; Pham & Hicheur 2011). In Chapter 3, the first systematic investigation into the typical development of head anticipation and trajectory formation in goal-oriented locomotion is presented (Belmonti et al., submitted). Our major finding is that both behaviours are consolidated as late as in early adolescence, i.e. well after the maturation of gait biomecanichal patterns and in a period when spatial cognition definitely shifts to a preferential allocentric strategy (Bullens et al. 2010).

From locomotion to navigation, from action to cognition

Finally, having realized the strong entanglement between action, perception and cognition, we have investigated the development of locomotor navigation in healthy children and in children with cerebral palsy. Navigation is the process or activity of accurately ascertaining one's position and planning and following a route (OED). It is a complex function, requiring several basic motor, perceptual and cognitive skills, and relying on different, sometimes concurrent cognitive strategies (Berthoz et al. 1995; Berthoz 1997; Maguire et al. 1998; O'Keefe et al. 1998; Chapter 4 in this thesis, for a review). To this respect, the comparison between hand movements and locomotion has taught us another lesson: encoding locations in the locomotor space is by no means the same as encoding locations in the reaching space. Inspired by previous work by Laura Piccardi, Cecilia Guariglia and their co-workers at the University La Sapienza in Rome (Piccardi et al. 2008; 2011; 2012), a new computerized tool for the presentation and recording of spatial sequences in the locomotor space has been devised: the Magic Carpet. The Magic Carpet employs the same spatial layout of the classical Corsi Block-tapping Test for visual-spatial memory, but enlarged to room size and with tiles to walk on instead of blocks to tap. It is equipped with pressure sensors and LEDs under tiles, to record and cue the subject's displacement. In Chapter 4, the results of its first application in typical development and in cerebral palsy are presented (Belmonti et al., in preparation). On the basis of previous reports and of our findings on trajectory formation, we hypothesized that navigational strategies would show major changes during school age. The comparison of memory performance on the table Corsi and on the Magic Carpet, however, did not reveal a major transition between 6 and 12 years, but only a slow trend towards a relative advantage in navigation with increasing age and intellectual level. In cerebral palsy, navigational skills were more often preserved than visual-spatial memory in the reaching space. Only children with temporal brain lesions were less favoured in navigation. In summary, navigation seems more an opportunity than a complication for most children with cerebral palsy, which is probably due to its multi-sensory and multi-strategy nature, at least when all sources of spatial information are available.