“…The first option, that internal models can be expected to become less accurate when the unimodal input itself is not well defined, is supported by a recent study in 7–13 year old children, testing the accuracy of the unimodal estimates of vision and proprioception. Using a localization task, the study showed that proprioceptive-based estimates become increasingly more reliable in older children (King, Pangelinan, Kagerer, & Clark, 2010; Pickett & Konczak, 2009); as a result, younger children up-weight visual information, whereas older children up-weight proprioceptive input when task demands require this. An earlier study in 5–11 year old children using a localization task in connection with a tendon vibration perturbation (Hay, Bard, Ferrel, Olivier, & Fleury, 2005) showed an interesting pattern of movement amplitude accuracy: constant amplitude errors showed a U-shaped function of age, with the highest accuracy at 5 and 11 years of age, and lower accuracy at 7 and 9 years.…”