Predicting Locomotor Skill Acquisition

Training characteristics will be collected during therapy sessions as potential predictor variables for acquisition and retention of locomotor skill, including measures of error, movement index, postural control and movement variability. These will be measured using video coding of recorded therapy sessions using Datavyu video coding software (45 ). Error during prone locomotor training will be defined as the number of corrective turns generated by the child. Corrective turns are a change in direction less than 2 s following another turning movement. Error during upright locomotor training will be defined as the number of losses of balance. Analyses will be performed with total amount of error, and frequency of error normalized to the distance travelled (error rate). The number and rates of error will be measured using the crawling and walking training robots. Movement index is the percent of time moving during each therapy session. This variable will be calculated using IMU data. Postural control is the percent of time with the head lifted during crawling therapy sessions, and the percent of time spent in an upright posture during walking therapy sessions. This variable will be determined by video coding. Movement variability during crawling is the activities that engage the SIPPC assist mechanism, measured by the IMUs. Movement variability during walking is the number of different motor activities in which the child engages, determined by video coding.