UCRISE, University of Canberra, Canberra, Australian Capital Territory, Australia.
Statistical Consulting Unit, Australian National University, Canberra, Australian Capital Territory, Australia.
Percept Mot Skills. 2021 Feb;128(1):283-303. doi: 10.1177/0031512520977683. Epub 2020 Dec 3.
The Active Movement Extent Discrimination Apparatus (AMEDA) has been used for measuring proprioception at various joints in the body for more than two decades. The utility of this instrument for discriminating groups has been reported in terms of an area under the curve (AUC) derived from an absolute identification test. This metric has supported statistically significant group differences, but it is not clear whether the AMEDA's testing protocol is suitable for measuring individual proprioception acuity changes. This study aimed to test the reliability, variance and absolute AUC scores obtained with the AMEDA with reference to other studies that have tested absolute identification acuity in other domains and the theoretical underpinnings of the testing protocol. We re-analyzed raw data from a 2013 study involving 65 people, most of whom were tested three times over two separate sessions on the ankle AMEDA by now assessing the accuracy of individual responses and calculating the sensitivity index, d', in addition to the AUC. To assess reliability, we calculated the intraclass correlation coefficient (ICC) and standard error of measurement (SEM) for those who completed all three tests. Fewer than 50% of presented stimuli were accurately identified and relatively poor discrimination was achieved between adjacent stimuli (in only one case was the median d' value greater than 1). The ICC of AUC scores across the three tests was poor (0.47). The SEM was 0.04, while 90% of participants' AUC scores fell between 0.59 and 0.76. The variation in performance at the individual level was substantial, producing a large SEM relative to the population spread of scores. We considered potential theoretical factors that may be affecting these results and concluded that an alternative approach will be needed in order for the apparatus to be used to explore individual proprioceptive performance.
主动运动幅度辨别仪（AMEDA）已被用于测量身体各关节的本体感受超过二十多年。该仪器在区分组群方面的效用已通过基于绝对识别测试的曲线下面积（AUC）来报告。该指标支持了具有统计学意义的组间差异，但尚不清楚 AMEDA 的测试方案是否适合测量个体本体感受锐度的变化。本研究旨在测试 AMEDA 的可靠性、变异性和绝对 AUC 评分，参考其他研究在其他领域测试绝对识别锐度以及测试方案的理论基础。我们重新分析了 2013 年一项涉及 65 人的研究的原始数据，其中大多数人在两次单独的踝关节 AMEDA 测试中进行了三次测试，现在评估了个体反应的准确性，并计算了灵敏度指数（d'），除了 AUC。为了评估可靠性，我们计算了所有三项测试完成者的组内相关系数（ICC）和测量标准误差（SEM）。只有不到 50%的呈现刺激被准确识别，相邻刺激之间的辨别力相对较差（只有一种情况下中位数 d'值大于 1）。三次测试的 AUC 评分的 ICC 较差（0.47）。SEM 为 0.04，而 90%的参与者的 AUC 评分在 0.59 到 0.76 之间。个体水平的表现差异很大，相对于评分的人群分布，SEM 相对较大。我们考虑了可能影响这些结果的潜在理论因素，并得出结论，为了使用该仪器探索个体本体感受性能，需要采用替代方法。
