BACKGROUND: External support has been viewed as an important biomechanical constraint for children with deficits in postural control. Nonlinear analysis of head stability may be helpful to confirm benefits of interaction between external trunk support and level of trunk control. OBJECTIVE: The purpose of this study was to compare the effect of biomechanical constraints (trunk support) on neural control of head stability during development of trunk control. DESIGN: This was a quasi-experimental repeated-measures study. METHODS: Data from 15 children (4-16 years of age) with moderate (Gross Motor Function Classification System [GMFCS] IV; n=8 [4 boys, 4 girls]) or severe (GMFCS V; n=7 [4 boys, 3 girls]) cerebral palsy (CP) were compared with previous longitudinal data from infants with typical development (TD) (3-9 months of age). Kinematic data were used to document head sway with external support at 4 levels (axillae, midrib, waist, and hip). Complexity, predictability, and active degrees of freedom for both anterior-posterior and medial-lateral directions were assessed. RESULTS: Irrespective of level of support, CP groups had lower complexity, increased predictability, and greater degrees of freedom. The effect of support differed based on the child's segmental level of control. The GMFCS V and youngest TD groups demonstrated better head control, with increased complexity and decreased predictability, with higher levels of support. The GMFCS IV group had the opposite effect, showing decreased predictability and increased complexity and degrees of freedom with lower levels of support. LIMITATIONS: Infants with typical development and children with CP were compared based on similar segmental levels of trunk control; however, it is acknowledged that the groups differed for age, cognitive level, and motor experience. CONCLUSIONS: The effect of external support varied depending on the child's level of control and diagnostic status. Children with GMFCS V and young infants with TD had better outcomes with external support, but external support was not enough to completely correct for the influence of CP. Children with GMFCS IV performed worse, with increased predictability and decreased complexity, when support was at the axillae or midribs, suggesting that too much support can interfere with postural sway quality.