Chen Yu, Gerling Kathrin, Verbert Katrien, Vanden Abeele Vero
e-Media Research Lab, Faculty of Engineering Technology, KU Leuven, Leuven, Belgium.
Augment Group, Department of Computer Sciences, KU Leuven, Leuven, Belgium.
JMIR Ment Health. 2025 Aug 5;12:e71304. doi: 10.2196/71304.
Early assessment of mild cognitive impairment (MCI) in older adults is crucial, as it enables timely interventions and decision-making. In recent years, researchers have been exploring the potential of gamified interactive systems (GISs) to assess pathological cognitive decline. However, effective methods for integrating these systems and designing GISs that are both engaging and accurate in assessing cognitive decline are still under investigation.
We aimed to comprehensively investigate GISs used to assess MCI. Specifically, we reviewed the existing systems to understand the different game types (including genres and interaction paradigms) used for assessment. In addition, we examined the cognitive functions targeted. Finally, we investigated the evidence for the performance of assessing MCI through GISs by looking at the quality of validation for these systems in assessing MCI and the diagnostic performance reported.
We conducted a scoping search in IEEE Xplore, ACM Digital Library, and Scopus databases to identify interactive gamified systems developed for assessing MCI. Game types were categorized according to genres and interaction paradigms. The cognitive functions targeted by the systems were compared with those assessed in the Montreal Cognitive Assessment (MoCA). Finally, we examined the quality of validation against the reference standard (ground truth), relevance of controls, and sample size. Where provided, the diagnostic performance on sensitivity, specificity, and area under the curve was reported.
A total of 81 articles covering 49 GISs were included in this review. The primary game types used for MCI assessment were classified as casual games (30/49, 61%), simulation games (17/49, 35%), full-body movement games (4/49, 8%), and dedicated interactive games (3/49, 6%). Of the 49 systems, 6 (12%) assessed cognitive functions comprehensively, compared to those functions assessed via the MoCA. Of the 49 systems, 14 (29%) had validation studies, with sensitivities ranging from 70.7% to 100% and specificities ranging from 56.5% to 100%. The reported diagnostic performances of GISs were comparable to those of common screening instruments, such as Mini-Mental State Examination and MoCA, with some systems reporting near-perfect performance (area under the curve>0.98). However, these findings often stemmed from small samples and retrospective designs. Moreover, some of these systems' model training and validation exhibited substantial deficiencies.
This review provides a comprehensive summary of GISs for assessing MCI, exploring the cognitive functions assessed by these systems and evaluating their diagnostic performance. The results indicate that current GISs hold promise for the assessment of MCI, with several systems demonstrating diagnostic performance comparable to established screening tools. Nevertheless, despite some systems reporting impressive performance, there is a need for improvement in validation, particularly concerning sample size and methodological rigor. Future work should prioritize prospective validation and present greater methodological consistency.
对老年人的轻度认知障碍（MCI）进行早期评估至关重要，因为这有助于及时进行干预和决策。近年来，研究人员一直在探索游戏化交互系统（GIS）评估病理性认知衰退的潜力。然而，将这些系统进行整合并设计出既引人入胜又能准确评估认知衰退的GIS的有效方法仍在研究中。
我们旨在全面研究用于评估MCI的GIS。具体而言，我们回顾了现有系统，以了解用于评估的不同游戏类型（包括类型和交互范式）。此外，我们研究了所针对的认知功能。最后，通过查看这些系统在评估MCI时的验证质量和报告的诊断性能，我们研究了通过GIS评估MCI的性能证据。
我们在IEEE Xplore、ACM数字图书馆和Scopus数据库中进行了范围搜索，以识别为评估MCI而开发的交互式游戏化系统。游戏类型根据类型和交互范式进行分类。将系统所针对的认知功能与蒙特利尔认知评估（MoCA）中评估的功能进行比较。最后，我们检查了针对参考标准（基本事实）的验证质量、对照的相关性和样本量。在有数据的情况下，报告了敏感性、特异性和曲线下面积的诊断性能。
本综述共纳入81篇涵盖49个GIS的文章。用于MCI评估的主要游戏类型分为休闲游戏（30/49，61%）、模拟游戏（17/49，35%）、全身运动游戏（4/49，8%）和专用交互式游戏（3/49，6%）。在49个系统中，有6个（12%）全面评估了认知功能，与通过MoCA评估的功能相比。在49个系统中，有14个（29%）进行了验证研究，敏感性范围为70.7%至100%，特异性范围为56.5%至100%。报告的GIS诊断性能与常见筛查工具（如简易精神状态检查表和MoCA）相当，一些系统报告了近乎完美的性能（曲线下面积>0.98）。然而，这些发现往往源于小样本和回顾性设计。此外，这些系统中的一些模型训练和验证存在严重缺陷。
本综述全面总结了用于评估MCI的GIS，探讨了这些系统评估的认知功能并评估了它们的诊断性能。结果表明，当前的GIS在评估MCI方面具有前景，一些系统的诊断性能与既定的筛查工具相当。然而，尽管一些系统报告了令人印象深刻的性能，但在验证方面仍需改进，特别是在样本量和方法严谨性方面。未来的工作应优先进行前瞻性验证并提高方法的一致性。
