Transition from upright stance to steady-state walking implies destabilisation of the current
antigravity postural set, whereas the opposite task, i.e. to stop walking, requires dissipation
of the kinetic energy attained by the body during forward progression, in order to allow
recovery of a static postural attitude. Both these tasks involve regulation of the external
forces by centrally initiated actions on the relative positions of COP and COM in the
anterior–posterior and mediolateral directions (Brenière et al. 1987, Crenna et al. 2001).
Initiation of gait, in particular, is triggered by a backward and lateral COP shift, with a
consequent fall of the body forward and toward the stance foot. The underlying motor
programme, which includes inhibition of postural activity in the soleus muscle followed by
activation of tibialis anterior and tensor fasciae latae, is initiated prior to any detectable
displacement of trunk and limb segments. This indicates the involvement of anticipatory
postural actions (APA; Crenna and Frigo 1991). In typically developing children, the
‘imbalance synergy’ associated with gait onset can be detected as early as within 1–4 months
of independent walking experience, and typically consists of a lateral tilt of the pelvis and
stance (trailing) leg, with consequent unloading of the contralateral swing (leading) limb
(Assaiante et al. 2000). In their earlier developmental stage these postural adjustments
display lower incidence, larger involvement of the upper parts of the body and more
consistent medio-lateral displacements, as compared to adults (Assaiante et al. 2000,
Malouin and Richards 2000). Follow-up of typically developing children from 21/2 to 8
years revealed progressively higher frequency of adult-like patterns, indicating that the
developmental tuning of such a feed-forward control takes a long time (Brenière and Bril
1998, Ledebt et al. 1998). Clinical experience in children with bilateral and unilateral spastic
CP does not reveal obvious difficulties in gait initiation. APA in the leg muscles associated
with arm rising while standing, which have been shown to rely on the same motor
programme adopted for gait onset (Crenna and Frigo 1991), have been detected in children
with bilateral spastic CP by recording the COP path under the support base during reaching
tasks (Jesinkey et al. 2005). This suggests that functional APAmay be substantially preserved
in CP (see also Stackhouse et al. 2007). Preliminary results from Malouin et al. (2003) in
children with unilateral spastic CP reported a reduced magnitude of the preparatory
adjustments on the paretic (swing) side, with simultaneous compensation by enhanced
anteroposterior actions on the non-paretic (stance) side. It should however be understood
that the standing attitude of children with CP differs from that of typically developing
children. Since the initial standing posture is known to affect the gait initiation process (e.g.
Couillandre and Brenière 2003), it is at present unclear to what extent the observed changes
should be ascribed to a primary dysfunction or should be related to the abnormal ‘postural
set’ of children with CP (and thus regarded as adequate adaptations). A further point worth
considering is that the above-reported findings refer to bilateral and unilateral forms of CP,
where deep (e.g. basal) brain structures are usually spared by neural lesion. In fact,
electrophysiological and imaging studies in adults suggest that control systems involved in
gait initiation include cortical, basal ganglia and brainstem structures and that APAs for gait
initiation can be severely affected in neural disorders of basal ganglia (Crenna et al. 1990;
for a review see discussion in Crenna et al. 2006). Further research, therefore, is necessary