Technology Research for Independent Living (TRIL) Centre and Intel Labs, Leixlip, Co. Kildare, Ireland.
IEEE Trans Biomed Eng. 2012 Apr;59(4):988-95. doi: 10.1109/TBME.2011.2181844. Epub 2011 Dec 26.
Cognitive decline and dementia have emerged as major challenges in modern healthcare with enormous associated societal and economic costs. Shifting demographics, owing to increasing numbers of people aged over 65 have greatly increased the potential scale of this problem in years to come. We report a novel quantitative method for assessment of cognitive decline (defined as a decline in mini mental state examination (MMSE) score of three or more) using quantitative parameters derived from body-worn inertial sensors. We sought to determine if baseline quantitative parameters and changes from baseline at follow-up, in those parameters could be used to automatically classify participants as cognitively declined or intact. Quantitative movement parameters were obtained at a baseline clinical assessment and in a follow-up assessment approximately 2 years later, using shank mounted triaxial gyroscopes. Data were obtained from 189 community dwelling older adults (aged over 60, 59 male, 130 female, mean age: 70.43 ± 6.57) while performing the timed up and go test. Nine participants who were deemed to be cognitively impaired at baseline (MMSE < 24) were excluded from analysis. Results suggest that quantitative parameters measured at baseline are 75.94% accurate in predicting cognitive decline in participants who were cognitively intact at baseline. A combination of baseline quantitative movement parameters and the change at follow-up (compared to baseline) in these parameters were 88.78% accurate in classifying final cognitive status in participants deemed cognitively intact at baseline. The reported method may be suitable for use as a portable cognitive screening tool, prompting further specialist clinical investigation and may also form part of a tool for longitudinal monitoring of cognitive function.
认知能力下降和痴呆已经成为现代医疗保健的主要挑战，给社会和经济带来了巨大的负担。由于 65 岁以上人口的增加，人口结构的变化极大地增加了未来几年这一问题的潜在规模。我们报告了一种使用来自佩戴式惯性传感器的定量参数评估认知能力下降（定义为简易精神状态检查（MMSE）评分下降 3 分或更多）的新定量方法。我们试图确定基线定量参数和随访时的基线变化是否可以用于自动将参与者分类为认知下降或完整。使用安装在腿上的三轴陀螺仪，在基线临床评估和大约 2 年后的随访评估中获得定量运动参数。数据来自 189 名居住在社区的老年人（年龄在 60 岁以上，59 名男性，130 名女性，平均年龄：70.43 ± 6.57），同时进行计时起坐测试。从分析中排除了 9 名基线时被认为认知受损（MMSE < 24）的参与者。结果表明，基线测量的定量参数在预测基线时认知正常的参与者认知下降的准确率为 75.94%。基线定量运动参数与随访时（与基线相比）这些参数的变化相结合，对基线时被认为认知正常的参与者的最终认知状态的分类准确率为 88.78%。该报告的方法可能适合用作便携式认知筛查工具，促使进一步进行专业临床调查，也可能成为认知功能纵向监测工具的一部分。
