s / Gait & Posture 42S (2015) S1–S90 S63 Table 1 Specificity and area under ROC curve (AuC) for distinguishing between normal and abnormal GM for variables derived from CIMA when sensitivity was set to 80%. Variables Specificity AuC Qmean 27 0.56 (0.45, 0.66) Qsd 28 0.56 (0.45, 0.67) Csd 52 0.70 (0.61, 0.79) CPP 57 0.75 (0.67, 0.83) FMD 61 0.75 (0.67, 0.84) CPP (cerebral palsy predictor) =Qmean+Qsd+Csd. Table 2 Specificity and area under ROC curve (AuC) for distinguishing between total presence (FM++) and total absence (FM−) of FMs for variables derived from CIMAwhen sensitivity was set to 80%. Variables Specificity AuC Qmean 54 0.68 (0.52, 0.84) Qsd 7 0.54 (0.36, 0.72) Csd 81 0.90 (0.80, 0.99) CPP 94 0.94 (0.86, 1.00) FMD 100 0.94 (0.87, 1.00) FMD (fidgety movements detector) =Qsd+Csd. in detecting the temporal organization of FMs in infants with intermittent (FM+) and sporadic (FM−/+) FMs. Furthermore, the distribution of the different categories of FMs may be different between a group of infants with high risk of later neurological impairment and a group of a mix of high risk and normal infants included in a former study using the same method . Nevertheless, the outcome variables evaluated in the present study is an important step in the development of more cost effective clinical detection of abnormal GMs.
N. Akalan, A. Apti, S. Kuchimov
Gait & Posture