Pii: S0304-3940(01)01925-5

We analysed the co-ordination of the elevation angles of the thigh (a t), shank (a s) and foot (a f) during walking in 19 adults and 21 children (aged 11±144 months), including the very ®rst unsupported steps in four. Cross-correlation functions (CCF) maturation of pairs of elevation angles was quanti®ed by a global error parameter (Et(CCF)) re ̄ecting the difference between particular CCF value of toddlers and mean adult value (Ea(CCF)). During the very ®rst step, Et(CCF) could be ®ve times higher than Ea(CCF). With walking experience, Et(CCF) for both a t-a s and a s-a f pairs evolved following a biexponential pro®le, with a fast time constant below 6 months. Adult-like CCF parameters were reached earlier for a s-a f than a t-a s, indicating disto-proximal maturation of the temporal co-ordination of the lower limb segments in human locomotion. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Locomotion; Development; Temporal co-ordination; Lower limb Studies in motor control have increasingly concentrated on the emergence of co-ordinative rules and integration of sensorimotor experiences into prewired neural networks [11]. In this context, the role of maturation in such a fundamental behaviour as locomotion appears critical but has received sparse attention until recently. Several parameters of postural and dynamical control have been studied with respect to early gait development, including head, trunk and hip stabilisation, intralimb co-ordination and centre of mass parameters [1,2,4,5,7,8,16]. In adults, a series of experimental studies of locomotion have demonstrated that the elevation angles of the lower limb segments provide a kinematic template revealing a motor organisation rule [3], which is related to minimisation of mechanical energy expenditure [14]. When the elevation angles of the thigh, shank and foot are plotted one versus the others, they describe a regular loop which lies close to a plane. Recently, we found that this planar covariation rule emerges early in toddler locomotion [7], suggesting a rapid dynamical integration of forward propulsion and postural control into the motor command. Two important aspects of the co-ordination of angular evolution of the lower limb segments, namely excursion amplitudes and timing, are expected to in ̄uence the planar covariation. The temporal relationships between these angles evolution appear particularly crucial for locomotor co-ordination. Statistical approaches such as cross-correlation function (CCF) have proved to be adequate for deciphering consistent temporal relationships between kinematics signals during free-form movements [1,6]. In this study we introduce a global parameter based on CCF of lower limb segments elevation to quantify the gait kinematic maturation of children from the very ®rst step through adolescence. Twenty-one healthy children (13 girls and 8 boys, aged 11±144 months) and 19 healthy adults (9 females and 10 males, aged 25 ^ 4 (mean ^ SD) years) participated in this study. The procedures were approved by the local Ethics Committee of the University Children Hospital Queen Fabiola and conformed with the Declaration of Helsinki. Special attention was given to recording the very ®rst steps of some toddlers. Daily recording sessions were programmed around the parents' expectation of the very Neuroscience Letters 308 (2001) 123±127 0304-3940/01/$ see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S0304-3940(01)01925-5 www.elsevier.com/locate/neulet * Corresponding author. Tel.: 132-2-6502187; fax: 132-26503745. E-mail address: [email protected] (G. Cheron). ®rst day of walking until unsupported locomotion was recorded. When we succeeded in recording this event, the same child was recorded again in order to follow the early maturation of walking. This was done for four children, who started to walk at 11, 11, 13 and 14 months, respectively. The other 17 children spanned the range of 3 weeks to 131 months of unsupported walking experience. The age of independent walking onset was reported by the parents. For the recording of the very ®rst steps, the toddlers were initially supported by a parent's hand. Then, when the parent moved forward, letting go of the child's hand, the child was encouraged to walk on the laboratory ̄oor coated with a red band of linoleum (0.6 m wide and 8 m long). Toddlers were never supported during the actual recording of locomotion trials. For each subject, 10 to 15 trials were recorded in the same condition. Subjects were instructed to look straight forward and to walk as naturally as possible from one end of the ground band to the other end. Kinematics of the locomotor movements was recorded and analysed using the opto-electronic ELITE system. This system consists of two CCD-cameras detecting retrore ̄ective markers using a sampling rate of 100 Hz. The cameras were placed on a line parallel to and 4-m away from the progression line of the subjects, 1 m above the ̄oor, 3 m apart. After calibration, two-dimensional data were corrected for optical distortion and converted to 3D co-ordinates [10]. The position in space of 10 passive re ̄ective markers (1.5 cm in diameter), de®ning 9 links, was recorded (Fig. 1A). The markers were fastened over the nose at the horizontal extent of the lower border of the orbit, ear meatus, acromial process, lateral condyle of the elbow, styloid process of the wrist, tubercle of the anterosuperior iliac crest, greater trochanter, lateral condyle of the knee, lateral malleolus and 5th metatarsal. After reconstruction of the stick diagrams representing successful locomotion, we focused our analysis on the orientation of the trunk and the lower limb segments with respect to the vertical. The following segments were analysed: trunk (de®ned by the line connecting the acromion and the iliac spine markers), thigh (trochanter-knee), shank (knee-lateral malleolus) and foot (lateral malleolous ± 5th metatarsal). The elevation angles of the thigh, shank and foot in the sagittal plane are noted a t, a s and a f, respectively (Fig. 1A). In order to test the temporal co-ordination between a t, a s and a f, CCF between each set of two variables were calculated. The span of time lags or leads was analysed for a time window (T ) corresponding to one gait cycle de®ned as the time interval between two successive maxima in the time series of a t. The CCF between two functions, e.g. a t and a s, was de®ned as: CCFat ;as …t† ˆ 1 T Z