Ray Mary Katherine, Fleming Jorie, Aschenbrenner Andrew, Hassenstab Jason, Redwine Brooke, Burns Carissa, Arbelaez Ana Maria, Vajravelu Mary Ellen, Hershey Tamara
Department of Psychiatry, Washington University in St. Louis, 660 S. Euclid Avenue, St. Louis, MO, 63110, United States, 1 3142738322.
Department of Radiology, Washington University in St. Louis, St. Louis, MO, United States.
JMIR Form Res. 2025 Feb 11;9:e60275. doi: 10.2196/60275.
Studies have shown a relationship between worse glycemic control and lower cognitive scores in youths with type 1 diabetes (T1D). However, most studies assess long-term glucose control (eg, years-decades) and cognition at a single time point. Understanding this relationship at a higher temporal resolution (eg, minutes-hours) and in naturalistic settings has potential clinical implications. Newer technology (eg, continuous glucose monitoring [CGM] and ecological momentary assessment) provides a unique opportunity to explore the glucose dynamics that influence dynamic cognition; that is, cognitive functions that fluctuate short-term and are influenced by environmental factors.
Before we can assess this relationship, we need to determine the feasibility of measuring cognition in youths in daily life and determine the plausibility of obtaining glucose variation with CGM to be integrated with real-time cognition measures. This study's purpose was to assess the acceptability of measuring dynamic cognition using a smartphone app and adherence to cognitive testing in daily life in youths with and without T1D. Further, we assessed CGM-derived glucose measures at temporally related timeframes to cognitive testing in naturalistic settings.
Data were obtained from 3 studies including one in-laboratory study and 2 remote studies. For all studies, youths were asked to complete cognitive tests on the Ambulatory Research in Cognition (ARC) smartphone app that measured processing speed, associative memory, and working memory. For the in-laboratory study, youths completed testing 4 times during 1 session. For the remote studies, youths were asked to complete cognitive tests 5 times per day for either 10 or 14 consecutive days in daily life. Youths were asked to rate their impressions of the app. Youths with T1D wore a CGM.
74 youths (n=53 control; n=21 T1D) aged 4-16 years participated. Youths generally reported liking or understanding the ARC app tasks in a laboratory and remote setting. Youths had high testing adherence in daily life (2350/3080 to 721/900, 76.3%-80.2%) and none dropped out. The percentage of measurements within each glycemic range taken immediately before the app's cognitive testing was 3% (28/942) low glucose, 51% (484/942) euglycemia, 23% (221/942) high glucose, and 22% (210/942) very high glucose. In the 2-hour window before each cognitive task, mean glucose was 182.5 (SD 76.2) mg/dL, SD in glucose was 27.1 mg/dL (SD 18.7), and the mean maximum difference between the highest and lowest glucose was 85.5 (SD 53.7) mg/dL.
The results suggest that using the ARC smartphone app to assess dynamic cognitive functions in youths with and without T1D is feasible. Further, we showed CGM-derived glycemic variability at temporally associated timeframes of dynamic cognitive assessments. The next steps include using ecological momentary assessment in a fully powered study to determine the relationship between short-term glycemic control and cognition in youths with T1D.
研究表明，1型糖尿病（T1D）青少年的血糖控制较差与认知得分较低之间存在关联。然而，大多数研究评估的是长期血糖控制（如数年至数十年），且在单一时间点评估认知情况。在更高的时间分辨率（如分钟至小时）以及自然环境中了解这种关系具有潜在的临床意义。更新的技术（如连续血糖监测[CGM]和生态瞬时评估）为探索影响动态认知的血糖动态变化提供了独特机会；动态认知即短期波动且受环境因素影响的认知功能。
在评估这种关系之前，我们需要确定在青少年日常生活中测量认知的可行性，并确定通过CGM获得血糖变化数据以与实时认知测量相结合的合理性。本研究的目的是评估使用智能手机应用程序测量动态认知在有和没有T1D的青少年中的可接受性，以及他们在日常生活中对认知测试的依从性。此外，我们在自然环境中与认知测试时间相关的时间范围内评估了CGM得出的血糖测量值。
数据来自3项研究，包括1项实验室研究和2项远程研究。对于所有研究，要求青少年在用于测量处理速度、联想记忆和工作记忆的“认知动态研究”（ARC）智能手机应用程序上完成认知测试。对于实验室研究，青少年在1次实验过程中完成4次测试。对于远程研究，要求青少年在日常生活中连续10天或14天每天完成5次认知测试。要求青少年对应用程序发表他们的看法。患有T1D的青少年佩戴CGM。
74名年龄在4至16岁的青少年（n = 53名对照；n = 21名T1D）参与了研究。青少年普遍表示在实验室和远程环境中喜欢或理解ARC应用程序任务。青少年在日常生活中的测试依从性较高（2350/3080至721/900，76.3% - 80.2%），且无人退出。在应用程序进行认知测试前立即采集的每个血糖范围内的测量百分比为：低血糖3%（28/942），正常血糖51%（484/942），高血糖23%（221/942），极高血糖22%（210/942）。在每次认知任务前的2小时窗口内，平均血糖为182.5（标准差76.2）mg/dL，血糖标准差为27.1 mg/dL（标准差18.7），最高和最低血糖之间的平均最大差值为85.5（标准差53.7）mg/dL。
结果表明，使用ARC智能手机应用程序评估有和没有T1D的青少年的动态认知功能是可行的。此外，我们展示了在动态认知评估的时间相关时间范围内CGM得出的血糖变异性。接下来的步骤包括在一项充分有力的研究中使用生态瞬时评估来确定T1D青少年短期血糖控制与认知之间的关系。
