Abstract:

Objective To explore the application value of visual paired comparison tasks based on eye-tracking technology in cognitive function assessment，and to provide an effective tool for large-scale early cognitive screening in community populations. Methods A cross-sectional study design was employed using convenience sampling. A total of 325 participants were recruited from the outpatient and inpatient departments of the Neurology Department at a tertiary A hospital in Guiyang City between December 2024 and October 2025. According to cognitive status, they were divided into the dementia group，the mild cognitive impairment group，and the cognitively normal group. After undergoing assessments using the Mini-Mental State Examination（MMSE） and the Montreal Cognitive Assessment（MoCA），participants simultaneously completed the visual paired comparison task using eye-tracking technology. The Wilcoxon rank-sum test was used to compare differences in eye movement parameters among the 3 groups. Receiver operating characteristic（ROC） curve analysis was conducted to evaluate the discriminatory performance of the visual paired comparison task based on eye-tracking technology assessment and its combined use with traditional scales for the 3 groups. Results The final study included 110 participants in the cognitively normal group，120 in the mild cognitive impairment group，and 95 in the dementia group. Comparisons of multiple eye-tracking indicators across the 3 groups revealed statistically significant differences（P<0.05）. The visual paired comparison task based on eye-tracking technology achieved an accuracy of 81.9%，a sensitivity of 74.2%，and a specificity of 77.3% in distinguishing between cognitively normal individuals and those with mild cognitive impairment. The accuracy increased to 97.1% when combined with traditional scales. When distinguishing between cognitively normal individuals and those with dementia，the accuracy of the standalone assessment was 91.4%，and 98.8% when combined with traditional scales. However，the accuracy for distinguishing between mild cognitive impairment and dementia was lower，at 64.3%，and increased to 71.7% after combined assessment. Conclusion The visual paired comparison task based on eye-tracking technology demonstrates high accuracy in distinguishing between cognitively normal and cognitively impaired individuals. However，for individuals already identified as cognitively impaired，their ability to distinguish between mild cognitive impairment and dementia is limited. It is recommended to combine this approach with traditional neuropsychological scales for a comprehensive assessment to improve the accuracy of clinical diagnosis.

Key words: Mild Cognitive Impairment, Dementia, Cognitive Function Assessment, Eye Tracking, Cognitive Screening, Nursing Care

摘要：

目的 探索基于眼动追踪技术的视觉配对比较任务在认知功能测评中的应用价值，为社区人群大范围早期认知障碍筛查提供有效工具。 方法 采用横断面研究设计，采用便利抽样法，选取2024年12月—2025年10月在贵阳市某三级甲等医院神经内科门诊就诊及住院的325例中老年人为研究对象，按认知情况分为痴呆组、轻度认知障碍组、认知正常组。研究对象在接受简易智能精神状态检查量表和蒙特利尔认知评估量表评估后，同步完成基于眼动追踪技术的视觉配对比较任务测评。采用秩和检验比较3组人群眼动指标的差异；通过受试者操作特征曲线下面积评估基于眼动追踪技术的视觉配对比较任务测评与传统量表联合使用对3组人群的鉴别效能。 结果 最终认知正常组纳入110例、轻度认知障碍组纳入120例、痴呆组纳入95例。3组眼动指标比较，差异有统计学意义（P<0.05）。基于眼动追踪技术的视觉配对比较任务区分认知正常与轻度认知障碍的准确率为81.9%，灵敏度为74.2%，特异度为77.3%；联合传统量表后准确率提升至97.1%；区分认知正常与痴呆时，单独测评准确率为91.4%，联合评估达98.8%。但区分轻度认知障碍与痴呆的准确率较低，为64.3%，联合评估后为71.7%。 结论 基于眼动追踪技术的视觉配对比较任务在区分认知正常与认知异常时具有较高准确性，但对于已识别出的认知异常个体，其在轻度认知障碍与痴呆的精细区分上效能有限，建议结合传统神经心理评估量表进行综合评估，以提高临床诊断的准确性。

关键词: 轻度认知障碍, 痴呆, 认知功能评估, 眼动追踪, 认知筛查, 护理