PhD Course in Basic and Developmental Neuroscience Head: Prof. Giovanni Cioni
The construction of motor spaces through
typical development and in children with
cerebral palsy
Tutor Prof. Giovanni Cioni Candidate Vittorio Belmonti CYCLE XXIV (2009-2011) SSD MED39Contents
Acknowledgements V Abbreviations VII 1 Introduction 3 The observation of spontaneous motor activity as a diagnostic tool 3 Motor development from spontaneous variability to goaloriented optimisation 3 The formation of motor trajectories: a bridge between different spaces 4 From locomotion to navigation, from action to cognition 5 2 Early diagnosis and prognosis of cerebral palsy 9 from a book chapter by Giovanni Cioni, Vittorio Belmonti and Christa Einspieler published in Cerebral Palsy, ed. by Roberta Shepherd (in press) Introduction 10 Techniques for the clinical assessment of the neonatal nervous system 11 Prechtl's method on the qualitative assessment of General Movements 13 Absent Fidgety Movements at 35 months is the most sensitive sign for CP prediction 16 CrampedSynchronized General Movements at preterm and term age are a very specific predictive sign for spastic CP 17 The asymmetry of selective distal movements at 3 months predicts unilateral CP 17 GM features that can predict dyskinetic CP 19 Motor optimality scores and the prediction of CP severity 19 Integrating GM assessment with traditional neurological examination and neuroimaging techniques 20 Conclusions 21 3 The typical development of anticipatory orienting strategies and trajectory formation in goaloriented locomotion 23 by Vittorio Belmonti, Giovanni Cioni and Alain Berthoz (submitted) Introduction 25 Materials and methods 28 Results 38 Discussion 464 Shortterm spatial memory for reaching and for navigation: typical development and differential impairment in cerebral palsy 51 by Vittorio Belmonti, Pierre Leboucher, Paola Brovedani, Cristiana Susino, Ylenia Capuzzo, Giovanni Cioni and Alain Berthoz (in preparation) Introduction 53 Materials and methods 56 Results 67 Discussion 75 5 Conclusions and future directions 83 Personal publications and scientific communications 85 References 87
Acknowledgements This doctoral thesis has been mainly carried out at the Department of Developmental Neuroscience of the Fondazione Stella Maris, Calambrone (Pisa), Italy, directed by Prof. Giovanni Cioni, in close collaboration with the Laboratoire de Physiologie de la Perception et de l'Action (LPPA) of the Collège de France, Paris, directed by Prof. Alain Berthoz. My work as a doctoral student has been thoroughly funded by La Fondation Motrice, Paris. La Fondation Motrice (LFM) is a nonprofit institution founded in Paris in July 2006 from the joining of three previous associations and entirely devoted to promote research in every field connected with cerebral palsy. My first two years were funded by the 2008 call for projects, my last year by a new project, the PerceptionAction CognitionEnvironment pilot project 2011, specifically aimed at understanding the interconnection between these four aspects of behaviour, for the improvement of assessment and treatment protocols. In addition, both the Fondazione Stella Maris and the LPPA have provided the necessary equipments, technical support, scientific databases, secretarial services, rooms and, most importantly, human relationships. My first thanks are for all the children and families who have participated in the studies I had the courage to propose to them. Sometimes I had to ask for their collaboration. More often, I was convinced to hold on by their enthusiasm. To Prof. Giovanni Cioni I give my longest due thanks. I have always worked with him since I first entered the specialty course for Child and Adolescent Neuropsychiatry as a medical student, and he has never stopped believing in me since then. This PhD thesis has been entirely supervised and inspired by Prof. Alain Berthoz, whom I thank for his neverending cognitive and moral support (that was often needed), his scientific rigour, and his demanding qualitative standard, without which I would have had much more rest and much less satisfaction. A number of people have helped me in several ways. I hope not to forget anyone. Michele Coluccini, kinesiologist at the Fondazione Stella Maris, was my first mentor and taught me part of what I know on motor development. We collaborated for some years and started together the study on locomotor trajectories. Sergio Scapellato, engineer, helped us create the experimental setup. Giovanna Macrì and Vito Monaco, engineers of the Scuola Superiore Sant'Anna (SSSUP), Pisa, contributed for a while to data analysis. Chiara Martelloni, also engineer from the SSSUP, was involved in the first steps of the project. QuangCuong Pham, Hideki Kadone, Delphine Bernardin and Colas Authié, postdoctoral students at the LPPA, and Laura Marazzato, PhD student at the LPPA and SSSUP, spent some of their time sharing with me their mehods, results and interpretations. I cannot forget my feeling so confident at each and every new explanation, and then getting lost again after some lonely hours in front of my laptop. Many hours, unfortunately. Pierre Leboucher, formerly engineer and researcher at the LPPA created the Magic Carpet, installed it in Pisa, inspired my early work with it and payed me two visits I remember with great pleasure. Mohamed Zaoui, also engineer and researcher at the LPPA, has created the virtual version of the Magic Carpet and the new digital version,
which promises to be a big improvement for assessment and rehabilitation protocols. Laura Piccardi and Prof. Cecilia Guariglia, neuropsychologists at the University La Sapienza, Rome (now Laura teaches at the University of L'Aquila) had the brilliant idea of the WalkingCorsi test, from which the Magic Carpet was born. They also had the kindness of sharing with me their data and thoughts, greatly favouring the Magic Carpet project. Prof. Gilles Kemoun and Anaick Perrochon, neurologists and researchers, are part of the Magic Carpet network, are studying the Magic Carpet task in elderlies and have shared with me fundamental ideas, among which their analysis of errors. Prof. Jacques Droulez, researcher and now director of the LPPA, has enlightened my thought about sequential analysis and inspired my method of error analysis on the Magic Carpet. Paola Brovedani and Chiara Pecini, neuropsychologists at the Fondazione Stella Maris, and Maria Chiara Di Lieto, PhD student, have helped me with the assessment of children with CP and with the unraveling of neuropsychological data.
I want to express my gratitude and friendship to Prof. Heinz Prechtl and all the tutors of the GM Trust (in addition to Giovanni Cioni): Prof. Christa Einspieler, Prof. Fabrizio Ferrari, Prof. Arend Bos, Maria Federica Roversi, Paola Bruna Paolicelli, Andrea Guzzetta and Natascia Bertoncelli. We collaborated in many courses and discussions, they taught me GM assessment (and much more) and made me the honour of accepting me as official tutor of the Trust. I am deeply thankful to all the students of Child and Adolescent Neuropsychomotricity (there is no English word to express it correctly, Physical Therapy and Occupational Therapy being far from its meaning) who worked at their theses under my supervision: Ilaria Felici, Elisa Pucciarelli, Ylenia Capuzzo, Simone Giannotti and Cristiana Susino. Ilaria also provided part of her family for enrollment to the study on trajectories, and Ylenia spent three months with me administering tests and the Magic Carpet at school. Many thanks to Michela La Marca, the teachers and pupils of the Scuola Elementare G. Viviani, Marina di Pisa, who enthusiastically participated in the Magic Carpet study. Many of my colleagues and coworkers at the Fondazione Stella Maris deserve a particular gratitude. I expressly refuse to cite all of them here, firstly because the list would fill several pages, and secondly because their help was only loosely linked to the present thesis. It had more to do with friendship and sentiments I cannot express in a doctoral thesis (I have been told that tears are not accepted here). However, I want to mention Giuseppina Sgandurra, who shared with me her own PhD work, Andrea Guzzetta, Roberta Battini, Paola Bruna Paolicelli and Silvia Perazza, who tutored me for a long time and accorded me access to their patients. I would like to mention the three secretaries with whom I have worked the most: Monica Casacci and Patrizia Neri at the Fondazione Stella Maris, and Emilie Gaillard, at La Fondation Motrice. Samuele Lupi and Maria for the coffees and the beautiful rests they provided. I have always thought that my family should not enter this list, because I tend to separate work from love and personal affairs. However, Silvia, my wife, has played a joke on me by creating the beautiful cover of this thesis. I take this opportunity to thank her for supporting my research efforts, and not loving me for the sake of their results.
Abbreviations AKD Average Kinematic Deviation ATD Average Trajectory Deviation CBT Corsi blocktapping test CI Curvature Index CP Cerebral Palsy CS CrampedSynchronized FM Fidgety Movement GM General Movement GMFCS Gross Motor Function Classification System HHD HeadHeading Deviation LPPA Laboratoire de Physiologie de la Perception et de l'Action MC Magic Carpet MRI Magnetic Resonance Imaging MTD Maximal Trajectory Deviation PR Poor Repertoire SD Standard Deviation WM Writhing Movement
«Reason, by whose aspiring influence, We take a flight beyond material sense, Dive into mysteries, then soaring pierce, The flaming limits of the universe.» [...] And 'tis this very reason I despise. This supernatural gift, that makes a mite, Think he is the image of the infinite: Comparing his short life, void of all rest, To the eternal, and the ever blessed. This busy, puzzling, stirrer up of doubt, That frames deep mysteries, then finds 'em out. John Wilmot, Count of Rochester, A Satyr against Mankind You must not find symbols in everything you see. It makes life impossible. Oscar Wilde, Salomé
1
Introduction
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 agespecific, 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 noninvasiveness (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 goaloriented 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 crystalclear 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 wellbeing 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, goaloriented motor actions would emerge from the selection of successful spontaneous motor patterns, by means of an ongoing trialanderror sensorimotor loop. Variability in innate motor behaviour (or, maybe better, "variation", as proposed by HaddersAlgra 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 multiarticulated 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 endeffector, e.g. a hand (redundancy); 2)
different endeffectors, 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 stimulusresponse 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 eyeorientation (Grasso et al. 1998b; Bernardin et al. 2012), and 2) trajectory geometry and kinematics are highly stereotyped in a worldcentred spatial reference frame (Hicheur et al. 2007). Both findings point to the existence of a topdown, feedforward 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 coworkers 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 Blocktapping Test for visualspatial 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 visualspatial 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 multisensory and multistrategy nature, at least when all sources of spatial information are available. To gain more insight into the cognitive strategies employed on the Magic Carpet task, an automated routine for the analysis of incorrect responses has been lately devised. Three global error measures are computed (topographic, topokinetic and length errors), yielding a classification of five strategies: 1) simple omissions or insertions, 2) topokinetic encoding, 3) strict topographic encoding, 4) approximate topographic encoding, and 5) undetermined. Preliminary results from healthy controls indicate that topographic encoding is greatly favoured at all ages, except for short sequences of 3 or 4 tiles. We propose two possible explanations: 1) topokinetic encoding allows for a more synthetic representation, embedding place and order information in a single representation, but becomes impracticable when too many locations are to be stored; 2) topokinetic encoding is closer to motor programming, being accomplished in a completely egocentric reference frame, whereas topographic encoding, not necessarily but preferably accomplished in an allocentric reference frame, is closer to the structure of the external space: as stimulus complexity increases, a topographic strategy becomes more efficient.
In summary, spatial cognition can be regarded as the top level of sensorimotor control. The ongoing exploration of the physical world leads the nervous system to try and select the best available motor patterns. Initially, they are too
variable (unoptimised) and, at the same time, too rigidly tied to the context in which they have been learned. Some innate patterns are so rigid that they are called "reflexes". Gradually, these patterns are optimised (by selection) and integrated in an increasingly complex hierarchy of control levels. At the bottom levels, neuronal groups still encode strictly bodyrelated and highly automated sensorimotor signals, but at the top levels, neuronal firing patterns become more and more independent from body structure, and more and more coherent with the world's structure (or, better, with the bodyinworld structure). This leads to increasing optimisation and flexibility in motor control, as well as in thought (Kelly 1955; Guidano 1991). As soon as a new problem is faced, available solutions are first tried, decomposed, and then further integrated in a novel, more optimised and worldcoherent coping solution. Finally, the most advanced strategies stand out as examples of utmost elegance and embodied simplicity, what Berthoz calls "simplexity" (Berthoz, 2009)
from a book chapter by Giovanni Cioni, Vittorio Belmonti and Christa Einspieler, published in Cerebral Palsy, ed. by Roberta Shepherd (in press)
2
Early diagnosis and prognosis of cerebral palsy
Introduction
The quality of care for the infants born with an extremely low birth weight, or with other conditions of high neurological risk, has achieved extraordinary progress in the last decades. Survival and quality of life have considerably increased for these children, but they still remain at risk for neurological damage, caused by perinatal infections, hypoxicischaemic damage, haemorrhagic insult, or by a combination of these factors. Negative outcome includes cerebral palsy (CP), intellectual disability, perceptual and sensory disorders, behavioural disorders and other.
The incidence of CP in particular has been considered the marker of the quality of neonatal care and utilized as outcome measure in many studies, although cognitive and behavioural problems are by far more common. Those more subtle disorders, but important for participation and quality of life as well, are difficult to diagnose in the first weeks or months of life. Conversely, it is largely agreed that an early diagnosis of CP should be possible, or better should be necessary even in the neonatal period, for a number of reasons.
Although all infants discharged by a NICU and potentially at risk for a neurodevelopmental disorder should enter a followup programme, children diagnosed as carrying a CP require medical and social resources that are limited in several countries. Furthermore, an early diagnosis of CP is necessary in order to start early intervention, to tailor it according to the infant’s needs, to modify positively the natural history of her/his condition, and to evaluate the results of the intervention. Randomization of the subjects for early interventional trials, based on neuroprotective drugs, environmental changes etc., requires detailed prognostic data. Finally, another major reason for all the efforts made on early diagnosis of CP deals with the information due to the parents, who obviously ask about the neurodevelopmental outcome of their infants. Their questions are: “will he/she be able to walk?” or “to take and use an object?”
Although prognostic hints for a CP are needed in the neonatal period, many textbooks and manuals still postpone the possibility of a diagnosis of CP to the end of a so called “silent period”, lasting some weeks or even, for mild forms of CP, some months after term age.
The possibility that neonatal neuroimaging can enable recognition of brain pathology that may lead to CP is today largely accepted in terms of type and timing of the brain lesions typical of the different forms of CP (KrägelohMann 2004), although it is questioned by some authors (O’Shea et al. 1998). A recent paper by De Vries et al. (2011) has unraveled the “myth” that CP cannot be predicted by neonatal neuroimaging, when this technique is correctly applied. The authors show evidence from a number of papers that a severe, non ambulatory CP can be predicted in the majority of cases by a combination of sequential brain US and by MRI carried out at term age. Additional useful indices can be obtained by assessing the myelination of the PLIC (Posterior Limb of the Internal Capsule) and DWI (DiffusionWeighted Imaging) techniques. Moreover, there are interesting research directions that may further improve, by sophisticated MRI techniques, the prognostic value of neonatal neuroimaging in the less severe forms of CP. However, in most centres brain MRI requires the transfer of the infant from the NICU. In addition, the equipment is extremely expensive and therefore not affordable for all centres and countries. It has also to be remarked that even the most advanced neuroimaging techniques can only show structural changes of the brain and do not provide information on the functional status of the nervous system. For this and other reasons, a clinical assessment is always necessary. Techniques for the clinical assessment of the neonatal nervous system Neonatal neurological examination is generally carried out through traditional neurological methods, largely based on muscle tone and neonatal reflexes. However, it is now widely recognized that the human nervous system can express many complex and rapidly changing functions from as early as the first weeks of gestation. Foetuses and newborns are complex organisms producing a great deal of endogenously generated behaviours. Some of the standard neurological examination methods, though standing as milestones of modern infant neurology, are still influenced by considerations drawn from adult neurology and experiments on animal models. For instance, SaintAnne Dargassies (1977) developed a pioneering examination protocol based on the evaluation of active and passive tone. Other methods were then proposed in the following decades, including items for muscle tone, postural motor milestones and, in some cases, behavioural aspects.
The neurological examination protocol proposed in the same period by Prechtl (Prechtl 1977) (not to be confused with the method the same author later proposed), has been standardized and validated only for the examination of
infants at term. It includes the extremely important concept of behavioural states, but many of its items are still based on muscle tone and responses integrated at a lowlevel in the CNS. Moreover, it is rather timeconsuming and cannot be applied to preterm infants.
The Neonatal Behavioural Assessment Scale (NBAS) is a technique developed by Brazelton (Brazelton and Nugent 1973) for examining the behaviour of term infants during the first couple of months of age. Its conceptual basis is founded on the assumption that the newborn has active and specific responses to environmental stimulations, rather than passive behaviour. On the basis of the NBAS, Als et al. (1982) standardized a behavioural scale for preterm infants, the Assessment of Preterm Infant Behaviour (APIB), thought to be employed in neonatal intensive care units also for providing and monitoring individualized intervention programmes. These techniques are timeconsuming and not easily applicable in clinical settings. Moreover, for the same test a high intra individual daytoday variability in the responses has been shown (Sameroff, 1978). Their main applications are in research and early intervention protocols. Currently, the most recently updated and extensively validated method for the neurological examination of preterm and fullterm newborn infants is the Hammersmith Neonatal Neurological Examination (HNNE), firstly published by Dubowitz and Dubowitz (1981) and updated by Dubowitz et al. (1999). These authors adapted tests drawn from the previous works of Prechtl, Saint Anne Dargassies, and Brazelton, in a simplified and userfriendly proforma, also including items based on the concepts of Prechtl and coworkers on spontaneous motor activity (see below). The items are organized in six sections: posture and tone, tone patterns, reflexes, movements, abnormal signs, behaviour. Typical normal and abnormal patterns are extensively described in the manual (Dubowitz et al. 1999) and have proven easily recognizable and clinically useful for diagnosis and prognosis. For research purposes, an optimality score for fullterm and preterm newborn infants has also been calculated (Mercuri et al. 2003). On the basis of the neonatal examination, the same authors also developed a protocol for use after the neonatal period in infants up to 24 months of age: the Hammersmith Infant Neurological Examination (HINE) (Dubowitz et al. 1999), divided into three sections, one nonagedependent neurological items, a second providing a summary of motor milestones and a third made of three simple behavioural items. Its prognostic value as to motor outcome has been found high both in preterm infants born before 31 weeks of gestation (Frisone et al. 2002) and in term infants with hypoxicischaemic encephalopathy (Haataja et al. 2001).
Although both HNNE and HINE have been tested in several clinical and research settings, they also present some limitations. Most items are still
correlated with muscle tone and reflexes and the distinction between normal and abnormal patterns may turn out, to some extent, rigid and schematic, hardly comprising the whole complexity of the infant's repertoire. Moreover, while several studies have reported statistically significant correlations between clinical findings, and mid and longterm outcome, some others have reported a relevant number of false positive and false negative results, especially among preterm infants (see Volpe 2008, for a review of followup studies).
Prechtl's method on the qualitative assessment of General Movements The basic requirements for an ideal method of neurological assessment for newborn infants were outlined by Heinz Prechtl as follows: it had to be non invasive, nontimeconsuming, and highly sensitive to variations of the age specific functional repertoire (Prechtl, 1990). None of the traditional protocols of neonatal neurological assessment matched completely these criteria. The observation and categorization of all spontaneous movements in the first months of life led Prechtl and his coworkers to the identification of several normal and abnormal motor patterns. Among others, the socalled "general movements" (GMs) were identified as ongoing global movements involving all body parts and appeared particularly suitable for assessment (Prechtl, 1990; 2001). This gave birth to Prechtl’s Method on the Qualitative Assessment of General Movements in Preterm, Term and Young Infants (see the published handbook: Einspieler et al. 2005a). Though based on global and qualitative judgment, GM assessment has proven a very sound method. Its validity and reliability have been extensively evaluated in a number of studies, mainly dealing with the prediction of CP (Prechtl et al. 1997; Ferrari et al. 2002; Cioni et al. 2000), but also with that of minor neurological disorders (Bruggink et al. 2008; Einspieler et al. 2007; Groen et al. 2005), Rett Syndrome (Einspieler et al. 2005b and c), cognitive development (Bruggink et al. 2010) and autistic spectrum disorders (Phagava et al. 2008). Normal GMs involve the whole body in a complex sequence of arm, leg, neck and trunk movements. They wax and wane in intensity, force and speed, have a gradual beginning and end. Rotations along the axis of the limbs and slight changes in the direction of movement make them fluent and elegant and create the impression of complexity and variability. GMs appear as early as 9 to 12 weeks postmenstrual age (PMA) and continue after birth without substantially changing their form until 4649 weeks PMA, irrespective of when birth occurs. At 4649 weeks PMA a major change in form sets in, being fulfilled around 3 months postterm, when the socalled "fidgety movements" (FMs) appear. From
5 to 6 months postterm, GMs fade out, while new, voluntary motor patterns emerge.
Though complex and variable by definition, GMs can be classified into a limited number of recognizable patterns, related to PMA (postmenstrual age) and either normal or abnormal. Normal GMs are described in Table 2.1.
Table 2.1 Typical time course of the normal patterns of General Movements
Pattern Age (wks. PMA) Description
Writhing Movements
(WMs) Range: 9 to 49Peak: 40 (term) Variable amplitude, slow to moderate speed, typically ellipsoid limb trajectories lying close to the sagittal plane with superimposed rotations.
Fidgety Movements
(FMs) Range: 46 to 64Peak: 52 Smaller than WMs, moderate average speed with variable acceleration in all directions, migrating through all body parts as an ongoing flow of movement. Continual in the awake infant, except during fussing, crying and focused attention. Legend: PMA=Postmenstrual age. As already stated, in the FM period various other motor patterns gradually emerge and mingle with GMs, thus building up the socalled "associated motor repertoire", whose richness and ageadequacy have been related to the optimality of later motor coordination (Bruggink et al. 2008) and cognitive functioning (Bruggink et al. 2010).
GMs of infants with cerebral impairment lack complexity, fluency and/or variability. Abnormal GM patterns can be sorted into two groups depending on whether they are observed before or after the onset of FMs, i.e. in the preterm/Writhing Movement period or in the FM period. The description of these patterns is reported in Table 2.2.
Table 2.2. Main abnormal patterns of GMs A) WM period Pattern Description Poor Repertoire (PR) Monotonous sequences, few movement components, repetitive and not so complex as in normal WMs. Fluency may be reduced too (but usually more spared than complexity and variability). CrampedSynchronized (CS) No complexity, fluency and variability: all limb and trunk muscles contract and relax almost simultaneously. Chaotic (Ch) Large amplitude, high jerk and chaotic order without any fluency or smoothness. Rare, often evolving into CS. B) FM period Pattern Description Absent FMs (F) FMs are never observed in the whole period. Abnormal FMs (AF) Fidgetylike movements, but amplitude, average speed and jerkiness are exaggerated.
The global visual perception of movement quality (Gestalt perception) has proven a powerful and reliable instrument to recognize normal and abnormal GMs, but only if scorers are properly trained and the technique is carefully applied. A thorough description of the standardized assessment procedure can be found in the GMs handbook (Einspieler et al. 2005a). Notably, the standard GM assessment is performed offline on selected video recordings, but it has also proven reliable (especially in the FM period) when performed live, as part of the routine neurological examination.
The methodology of Prechtl's GM assessment has evolved to today's standardized and highly reliable form through several years. Concerns about its possible biases (e.g. because of poor validation in nonEuropean countries) may still arise (Darsaklis et al. 2011), but they are easily overestimated if earlier studies are considered together with newer ones, or if results obtained in different clinical populations (e.g. highrisk and lowrisk babies) are mixedup. Recently, the most predictive features of GMs have been definitely pointed out
and their clinical role extensively reviewed, as summarized in the following paragraphs, especially in relation to the prediction of CP. Absent Fidgety Movements at 35 months is the most sensitive sign for CP prediction In 1997, Heinz Prechtl and associates carried out the todate most important study on the predictive value of GM assessment, indicating it to be a reliable and valid tool for distinguishing between infants who are at significant risk of developing CP and infants who are not (Precthl et al. 1997a). The findings were based on a longitudinal study on 130 infants who represented the whole spectrum of perinatal brain ultrasound findings. Central to the study were the agespecific FMs, i.e. normal FMs observed at least once between 3 and 5 months postterm age. Ninetysix percent of the infants with normal FMs (N = 70) had a normal neurological outcome. Abnormal quality or total absence of FMs were followed by neurological abnormalities (most of them CP) in 95% of the 60 infants. Specificity and sensitivity of the assessment of FMs (96% and 95%, respectively) were higher than those of cranial ultrasound (83% and 80%, respectively). Since then, various groups have emphasized the significance of FMs for the early prediction of CP. Burger and Louw (2009) reviewed 15 studies on the predictive value of FMs and reported a sensitivity > 91% and a specificity > 81%. So far, the largest sample recruited in a longitudinal study has been of 903 children, which yielded a sensitivity of 98% and a specificity of 94% (Romeo et al. 2008).
As already mentioned, GM assessment is based on global and qualitative judgment, but has proven highly reliable and consistent, both inter and intra subjectively, especially in the FMs period. Its high level of objectivity has been documented by an interscorer agreement ranging from 89% to 93%, and by an average Kappa of 0.88, both obtained in a total of 15 studies (Einspieler et al. 2005a; Fjørtoft et al. 2009). Such high values can be achieved after a some days of extensive training (Valentin et al. 2005). A high intraindividual consistency of GM quality was demonstrated by Kappa values from 0.90 to 0.96 (Mutlu et al. 2008). Notwithstanding that, because of the utmost importance of FMs for prognosis, recent research has struggled to provide automated and objective methods for their identification. A computerbased video analysis technique has been recently developed by Adde and coworkers (Adde et al. 2009; 2010), yielding a sensitivity of 85% and a specificity of 88% for CP prediction.
The mere absence of FMs, however, has never been found specific for a particular CP subtype, nor it can predict CP severity. This fact indicates that several neural structures, at least the corticospinal fibres, the basal ganglia and
the cerebellum, need be intact to generate normal FMs. The latter are thought to be a necessary step for an optimal calibration of the sensorymotor system (Prechtl, 1997b). Interestingly enough, normal FMs are also absent in infants with genetic disorders. FMs are thus very sensitive for prognosis, while other motor features, combined with the absence of FMs, have proven useful to predict CP type and severity (see next paragraphs).
CrampedSynchronized General Movements at preterm and term age are a very specific predictive sign for spastic CP
As mentioned, fetal, preterm and writhing GMs display three patterns of abnormality, among them the socalled ‘crampedsynchronized GMs’ (CS). CS GMs appear very rigid and abrupt, all limb and trunk muscles appear to contract almost simultaneously and relax almost simultaneously (Ferrari et al. 1990). If normal GMs are characterized by fluency, complexity and variability, CS GMs lack all these three main features. Observing this pattern consistently over several weeks is highly predictive (98%) for the eventual development of spastic CP (Table 2.3) (Prechtl et al. 1997a). The sooner CS GMs evolve and the longer they last, the more severe the future motor impairment will be (Ferrari et al. 2002). Conversely, transient CS feature (i.e. observed on one of several longitudinal observation of the same infant), usually does not predict CP. Moreover, CS GMs are not very sensitive (may lack or be inconsistent in CP) and do not usually occur in children with nonspastic CP.
The asymmetry of selective distal movements at 3 months predicts unilateral CP As already stated, children with spastic CP show abnormal GMs during their first weeks of life and have no FMs at 3 to 5 months postterm age, which is also true for unilateral forms. By the way, this finding denies the presence of a 'silent period' in children with unilateral CP (Cioni et al. 2000; Guzzetta et al. 2001). At about 2 to 4 months postterm, the first asymmetries can also be observed (Table 2): the socalled ‘segmental movements’ (i.e. isolated finger and toe movements which are part of the spontaneous motor repertoire) are reduced or even absent in the side contralateral to the lesion, regardless of head position (Cioni et al. 2000; Guzzetta et al. 2001; 2009). Noteworthy, at this age neurological examination may still yield normal results (see below).
Table 2.3 Developmental trajectories with a high predictive power for normal development and the development of CP.
GMs during
preterm age Writhing GMs(at term) Fidgety GMs(3-5 months) Neurological outcome Reference
Poor repertoire
or normal GMs Poor repertoire or normal GMs Normal fidgety movements Normal Prechtl (1990; 1997; 2001), Ferrari et al. (1990), Prechtl et al. (1997), Cioni et al. (1997a; b), Hadders-Algra (2004), Einspieler et al. (2004; 2005a) Poor repertoire or cramped-synchronized GMs
Cramped-synchronized GMs Absence of fidgety movements; abnormal findings in neurological examination
Bilateral spastic CP Prechtl (1990), Prechtl et al. (1997; 2001), Ferrari et al. (1990; 2002; 2011), Cioni et al. (1997a; b), Einspieler et al. (2004; 2005), Hadders-Algra (2004), Snider et al. (2008), Romeo et al. (2008), Bruggink et al. (2009), Spittle et al. (2009), Spittle et al. (2010), Adde et al. (2010), Hamer et al. (2011), Burger & Louw (2011) Poor repertoire or cramped-synchronized GMs Poor repertoire or cramped-synchronized GMs Absence of fidgety movements and asymmetrical segmental movements; normal or abnormal findings in neurological examination Unilateral spastic CP Cioni et al. (2000), Guzzetta et al. (2003; 2010), Einspieler et al. (2004), Romeo et al. (2008), Einspieler (2008) Poor repertoire GMs Poor repertoire GMs, circular arm movements and finger spreading Absence of fidgety movements, absence of foot-to-foot contact, circular arm movements and finger spreading
Dyskinetic CP Einspieler et al. (2002; 2004)
GM features that can predict dyskinetic CP
Until the second month postterm, infants who will later become dyskinetic show a socalled ‘poor repertoire' (PR) pattern of GMs (Einspieler et al. 2002). A monotonous sequence of movement components and a lack of complexity characterize PR GMs (Ferrari et al. 1990). Apart from the PR pattern, which is by no means specific, these infants have been found to move their arms repeatedly and stereotypically in circle and to spread their fingers in an exaggerated manner (Einspieler et al. 2002). Characteristically, these abnormal circular arm movements are present at least until the age of 5 months postterm. They are uni or bilateral, monotonous, slow forward rotations originating in the shoulder. From 3 to 5 months, FMs and limb movements towards the midline (particularly foottofoot contact) are absent (Table 2.3). Motor optimality scores and the prediction of CP severity A semiquantitative assessment of GM quality can be achieved by applying Prechtl's optimality concept (Prechtl 1980). A score for optimal or nonoptimal performance is given to every movement criterion, such as amplitude, speed, movement character, sequence, range in space, onset and offset of GMs. Two different optimality scoring lists have been reported: the first one for preterm and term age (Ferrari et al. 1990); and the second one covering the whole motor behaviour, not only GMs, of 3 to 5monthold infants (Bos et al. 2003). In fact, as previously mentioned, the motor repertoire of infants aged 3 to 5 months consists not only of FMs but also of other motor patterns, e.g. kicking, swiping and wigglingoscillating arm movements, movements towards the midline, leg lifts, arching and axial rolling (Einspieler et al. 2008). The motor optimality score at 35 months is thus the sum of five components: 1) the presence and quality of FMs, 2) the presence and normality of other movement patterns, 3) the presence and normality of postural patterns, 4) the ageadequacy of the concurrent motor repertoire, 5) the quality of the concurrent motor repertoire. The presence and quality of FMs is always the most important feature and is weighted more than the other components. GM optimality scores can be used for statistical calculations and comparisons with other measures, especially in research settings. Optimality scoring should never be carried out prior to or together with the global assessment of GMs by means of Gestalt observation, as the latter is easily disrupted by detailed analysis.
A recent study by Bruggink et al. (2009) investigated into the predictive value of the motor optimality score at 35 months, in relation to the level of selfmobility
of CP children at school age. In this study, the abnormal quality of the concurrent motor repertoire was separately classified as monotonous, jerky and/or cramped. Of the 347 prospectively enrolled participants, 37 developed CP and were classified according to the GMFCS (Palisano et al. 1997). The higher the motor optimality score, the better was the GMFCS level. The positive and the negative predictive values of the optimality score (cutoff = 9) for the outcome (either levels III or levels IIIV) were both around 70%. Among the various single features of the motor repertoire, (a) a cramped movement character; (b) a reduced ageadequate motor repertoire; (c) monotonous kicking; and, obviously (d) the absence of FMs, were all associated with lower levels of selfmobility. Integrating GM assessment with traditional neurological examination and neuroimaging techniques GM assessment may be carried out as an isolated means of observation (for research purposes, serial checks, or when the infant cannot be otherwise evaluated), but is more often part of a comprehensive assessment protocol. As indicated in the introduction of this chapter, irrespectively of their clinimetric properties, the traditional neurological examination and the assessment of GMs have different conceptual backgrounds. The previous one invariably consists of a list of individual items, each of which is separately scored and aims to evaluate a particular 'function' or structure of the nervous system. Most of the items are usually based on muscle tone, reflex reactions, postural patterns, single 'bad signs', and motor milestones. On the contrary, GM assessment is a global, simultaneous appreciation of movement quality, which cannot be further distinguished into any individual 'components', at least not during the same session (detailed analysis and the assessment of isolated distal movements must take place after the GM pattern has been determined). GMs are more related to the general wellbeing of the central nervous system than any other single behavioural feature, and have been recognized as the most prominent and consistent agespecific motor behaviour shown by young infants (Prechtl, 1990; 1997b). On the other hand, they still provide little information on the anatomical structures involved in pathological processes. In summary, while traditional neurological items can be seen in tighter relationship with brain impairment (what the child cannot do), GMs are more sensitive to the agespecific brain function (Palmer et al. 2004). The integration of GM assessment with traditional neurological examination is particularly useful for distinguishing unilateral from bilateral CP. In an Italy wide multicentre study (Romeo et al. 2009), 13 children out of 903 preterm
infants were eventually diagnosed with unilateral CP. Eleven of them had shown no FMs, which is remarkable, since nine of them had had only a persistent flare on the brain ultrasound with no signs of unilateral damage. Surprisingly, the HINE scores of all but one infant were within the normal range. On the contrary, most of the infants who later developed bilateral CP had abnormal HINE scores. These results also lead to the important conclusion that a 3 to 4 monthold infant with a normal neurological score but an absence of fidgety movements and asymmetric segmental movements is at a high risk of developing unilateral CP (Einspieler et al. 2008).
Another important tool for the early diagnosis of brain impairment is neonatal Magnetic Resonance Imaging (MRI). So far, only few studies have addressed the relationship between specific MRI findings and GM patterns. In very preterm infants, both white matter lesions (Spittle et al. 2009) and a reduced cerebellar diameter, but not grey matter abnormalities, were associated with absent FMs (Spittle et al. 2010). In infants born at term, however, the severity of the injury to the central grey matter and basal ganglia correlated with a lack of FMs (Ferrari et al. 2011). In the latter study, CS GMs were highly specific for CP (100%), while MRI was by far more sensitive (100%). This high sensitivity of MRI abnormalities may sometimes be misleading for diagnosis, for instance in the case of tiny focal lesions: about half of neonatal cerebral infarctions are known not to develop into CP (Wu et al. 2005). Definite GM abnormalities (especially absent FMs and CS GMs), have on the contrary a very high positive predictive value, as seen, which means that CP will very likely follow. Again, the integration of different approaches, born with different purposes and from distant professional areas, is the best way of obtaining comprehensive information on the child's actual wellbeing and future development.
Conclusions
The methodological breakthrough of the GM assessment, which is non intrusive, easy to acquire and costeffective, lies in its predictive value for the development of neurological deficits, in particular CP, at a much earlier age than before. Recognition of abnormal GMs can help to improve earlier detection of CP if this technique can be successfully incorporated into followup programs and developmental surveillance (Palmer, 2004), combining it with neuroimaging (especially MRI) and neurological assessment. The great advantage of detecting an increased risk of CP at such an early stage consists in the possibility of intervention long before the emergence of pathological and adaptive features. The consistent presence of crampedsynchronized GMs, and even more so the absence of fidgety movements, puts an infant at such a high
risk of CP that early physiotherapeutic intervention is justified. It is no less important to identify infants who, despite an increased risk based on their clinical history, have normal GMs and can thus be expected to have a normal neurological outcome. In addition, the design of randomized control studies to test the effectiveness of early physical or other interventions for those infants who had neonatal adverse events and are at risk for neurodevelopmental disorders, require precise criteria for case selection and group allocation. The evaluation of GMs can be an important tool for these studies.
and trajectory formation in goaloriented locomotion
by Vittorio Belmonti, Giovanni Cioni and Alain Berthoz (submitted)Abstract Background In goal oriented locomotion, healthy adults generate highly stereotyped‐ trajectories (Hicheur et al. 2007) and a consistent anticipatory head orienting‐ behaviour (Grasso et al. 1998), both evidence of a topdown, openloop control (Pham & Hicheur, 2011). Aim To describe the typical development of anticipatory orienting strategies and trajectory formation. Our hypothesis is that fullblown anticipatory control requires advanced navigational skills. Methods 26 healthy subjects (14 children: 411 yrs; 6 adolescents: 1317 yrs; 6 adults). Task: to freely walk towards one out of three visual targets, in a randomised order. Optoelectronic motion capture (SMART®, BTS) with 15 body markers. Signals extracted (yaw plane): whole body displacement, orientation of head,‐ trunk and pelvis, heading direction, foot placements. Parameters: maximal head heading deviation, headheading anticipation time, trajectory curvature, indexes of variability of the trajectories, foot placements and kinematic profiles. Results Mean headheading anticipation and trajectory curvature did not significantly differ between agegroups. In children, however, head anticipation was more often lacking ( ²=9.08, p<0.01χ ) and there were huge intra and intersubject variations. Trajectory curvature could be very high in children, while it settled on consistent lower values in adolescence (χ²=77.17, p<0.0001). The indexes of spatial and kinematic variability all followed a decreasing developmental trend (R²>0.5, p<0.0001). Conclusions Children under 11 do not plan locomotor trajectories as adolescents and adults do. Anticipatory head orientation and trajectory formation develop in late childhood, well after gait maturation (Vaughan et al. 2003). Navigational skills, such as path planning and shifting from ego to allocentric spatial reference frames, are thought necessary for mature locomotor control.
3
The typical development of anticipatory orienting
strategies and trajectory formation in goaloriented
locomotion
Introduction Locomotion is a complex action in space, of which gait (here meant in a strict sense as the cyclic motor pattern allowing locomotion) is but a means to achieve wholebody displacement. To attain a goal, the latter also requires control on walking direction and the ability to orient it in space, i.e. steering (Grillner et al. 2008), by which locomotor trajectories on the horizontal plane are generated (Vieilledent et al. 2005). To our knowledge, this is the first study to systematically describe bodyorienting strategies and trajectory formation in goaloriented locomotion in typically developing subjects from a wide age range, from preschoolers to adulthood.Several investigations have addressed the typical development of gait in humans. Early works (Sutherland et al. 1980; Beck et al. 1981) set the maturation of gait kinematic and kinetic patterns in the first three and five years of life, respectively. Later studies, normalising temporal and distance parameters by anthropometric measures, moved the acquisition of mature gait to 78 years (Vaughan et al. 2003). From a neural point of view, gait control seems to develop through locomotor experience from a very stereotyped, innate and phylogenetically preserved stepping template, as revealed by EMG and kinematic patterns (Forssberg 1985; Ivanenko et al. 2005; Dominici et al. 2011). In particular, the emergence of planar covariation of lower limb elevation angles (a sign of central neural control on gait) takes place in the first year of independent walking (Ivanenko et al. 2005; 2008), but it takes much more locomotor experience to select adultlike coordination patterns (Dominici et al. 2011).
In addition to limb movements, gait includes axial postural synergies aimed at maintaining balance and orienting the body in space (Massion 1998). The typical development of dynamic postural control has been shown to consist of two basic achievements, both fulfilled by the age of 7: moving upward the postural reference frame from the pelvis to the head, and shifting from a
simplified, en bloc postural organisation to a flexible, articulated one (Assaiante 1998; Assaiante et al. 2005).
So far, however, only few investigators have addressed the development of locomotor control in steering. Grasso et al. (1998a) were the first to describe anticipatory head orientation in children as young as 3.5 years performing a cornerturning task. Vallis & McFadyen (2005) studied obstacle cirvumvention and found that 10yearsolds change direction differently than adults, i.e. partitioning the task into a first head and trunkreorienting phase and a second CoMdeviation phase. While so little is known about steering development, some basic principles underlying curvilinear locomotion emerge from adult studies and are reviewed in the next paragraphs. Head and gaze anticipation The direction of walking is consistently anticipated by the yaw orientations of head (Grasso et al. 1998a,b; Prévost et al. 2002; Patla et al. 1999; Courtine & Schieppati 2003; Sreenivasa et al. 2008) and gaze (Grasso et al. 1998b; Prévost et al. 1998; Hollands et al. 2002; Bernardin et al. 2012). Head and gaze anticipation is a universal locomotor behaviour, though not indispensable (Cinelli & Warren 2012). It does not strictly depend on the availability of visual cues, being reported even in darkness (Grasso et al. 1998b), when walking blindfolded (Pham et al. 2011; Prévost et al. 2002; Courtine & Schieppati 2003), and in locomotor imagery (Wagner et al. 2008). If head rotation is prevented, the whole turning behaviour is affected (Hollands et al. 2001). In more complex tasks, as for instance in challenging ground conditions, head and gaze behaviours are more varied, but still aiming at anticipating the future walking trajectory (Marigold & Patla 2007; 2008). In children, head anticipation has been reported as early as at 3.5 years of age (Grasso et al. 1998a), but no extensive study has ever been performed throughout development.
Head stabilisation
Head orientation is stabilised on the pitch and roll planes with respect to gravity in a wide range of grossmotor activities, including locomotion (Pozzo et al. 1998b). While in straight walking the head compensates for steprelated oscillations of the gravitoinertial acceleration vector (GIA), in steering, when the GIA is tilted by centripetal acceleration, there is a substantial alignment
between head and GIA (Imai et al. 2001). On the yaw plane, this results in stabilising the head with respect to the walking direction (Hicheur et al. 2005a). The formation of locomotor trajectories Locomotor and hand trajectories share important analogies, in agreement with the principle of 'motor equivalence' (Bernstein 1967). A prominent kinematic invariant of hand movements is for instance the power law relating tangential velocity and curvature (Viviani & Terzuolo 1982; Richardson & Flash 2002) and a similar relationship has been found for locomotion (Hicheur et al. 2005; Vieilledent et al. 2005). The latter studies employed predetermined geometrical trajectories. More recently, the spontaneous formation of goaloriented trajectories has been investigated. As a major finding, the variability of trajectory geometry and kinematics (temporal profiles of tangential velocity, head, trunk and pelvis orientation) proved strikingly low in adults (Hicheur et al. 2007). This 'stereotypy' in trajectory formation has been confirmed across trials of different speed and in the absence of vision (Pham et al. 2011). According to the Authors, this is consistent with an openloop (i.e. anticipatory) control model for the optimisation of locomotor trajectories (Pham et al. 2007; Pham & Hicheur 2011; Arechavaleta et al. 2008). In addition, the adult brain is capable of using several geometries, including nonEuclidean ones, to generate trajectories in different walking conditions (Bennequin et al. 2009), which opens further questions on how internal models for motor control are built through ontogenesis.
Aims of the study
Our first aim is to describe the typical development of anticipatory orienting strategies, global trajectory geometry and kinematics in goaloriented locomotion, across a wide agerange. More specifically, we are interested in measuring head anticipation and the variability of locomotor trajectories in children, adolescents and adults. Our hypothesis is that fullblown anticipatory orienting strategies and stereotyped trajectories, as hallmarks of a topdown, openloop control on locomotion (Pham & Hicheur 2011), can only be consolidated through locomotor experience in late childhood, i.e. after the acquisition of mature navigational skills (Bullens et al. 2010) and in any case well after the fulfillment of gait maturation at 78 years.
In addition, we are collecting reference data for future studies on children with locomotor disorders. In particular, children with cerebral palsy have never been
