Computerised cognitive testing and multi-domain structural Magnetic Resonance Imaging in patients with idiopathic Normal Pressure Hydrocephalus and Alzheimer's disease
Computerised cognitive testing and multi-domain structural Magnetic Resonance Imaging in patients with idiopathic Normal Pressure Hydrocephalus and Alzheimer's disease
复制标题
特发性正常压力脑积水和阿尔茨海默病患者的计算机认知测试和多域结构磁共振成像
DOI：
10.1002/alz.074654
复制
发表时间：
2023
期刊：
影响因子：
--
通讯作者：
Giovane M
中科院分区：
文献类型：
--
作者：
Giovane M
背景特发性正常压力脑积水（iNPH）的特点是步态障碍、尿失禁和认知障碍。通过脑脊液 (CSF) 改道，症状可能会逆转；因此，早期、准确的诊断至关重要。 iNPH 的症状和影像学特征通常与阿尔茨海默病 (AD) 和其他神经退行性疾病重叠，这是一个重大的诊断挑战。我们的目的是展示计算机化神经心理学测试在帮助 iNPH 诊断方面的实用性，以及探索 iNPH 认知功能障碍的神经解剖学相关性。MethodiNPH (N = 11) 和阿尔茨海默病 (N = 14) 患者接受了在线计算机化认知测试（图 1）和 MRI 扫描。使用一种新颖的规范方法将原始认知分数转换为“与预期的偏差”(DfE) 分数，将线性模型训练到 N = ∼200,000 个规范，同时考虑人口统计数据。对 DfE 分数进行针对 0 的 T 检验，以识别对 iNPH 和/或 AD 敏感的任务。方差分析用于识别 AD 和 iNPH 相对于 10 名年龄匹配的健康对照 (HC) 的区域脑容量和束特异性分数各向异性 (FA) 值的差异。使用一般线性模型来研究对 iNPH 和 AD 敏感的认知任务是否可以预测已识别的 MRI 异常。结果 AD 患者在所有任务中均表现出显着损伤，而 iNPH 患者在运动控制、处理速度、视觉空间和执行功能任务方面的得分显着低于对照组（图 2）。单因素多元方差分析确定了各组之间的认知差异（F(22,554)= 14.47，p<0.001），与 AD 相比，iNPH 在记忆和语言测试中的表现明显更好。与HC相比，iNPH患者胼胝体和骨骼的FA较低，扣带皮层体积减少，心室总体积较高。相反，与 HC 患者相比，AD 患者的下纵束和骨骼中 FA 减少，颞叶和顶叶体积较低（图 3）。对于 iNPH，目标检测和单词延迟回忆预测胼胝体 FA，而视觉空间任务预测心室和扣带回体积。在 AD 组中，多重比较校正后，任务表现与结构变化之间没有检测到统计学上显着的联系。结论远程计算机认知测试可以区分 AD 和 iNPH，并有可能预测与疾病相关的 MRI 异常。
BackgroundIdiopathic normal pressure hydrocephalus (iNPH) is characterised by gait disturbance, urinary incontinence and cognitive impairment. Symptoms are potentially reversible via cerebrospinal fluid (CSF) diversion; thus early and accurate diagnosis is crucial. Symptoms and imaging features of iNPH often overlap with Alzheimer’s disease (AD) and other neurodegenerative conditions, which represent a major diagnostic challenge. We aimed to demonstrate the utility of computerised neuropsychology testing to aid iNPH diagnosis, as well as exploring the neuroanatomical correlates of cognitive dysfunction in iNPH.MethodiNPH (N = 11) and Alzheimer’s disease (N = 14) patients underwent online computerised cognitive testing (Figure 1) and MRI scanning. Raw cognitive scores were converted into ‘deviation from expected’ (DfE) scores using a novel normative approach, training linear models to N = ∼200,000 norms while accounting for demographics. T‐tests against 0 were conducted on the DfE scores to identify tasks sensitive to iNPH and/or AD. ANOVAs were used to identify differences in regional brain volumes and tract‐specific fractional anisotropy (FA) values in AD and iNPH relative to 10 age‐matched healthy controls (HC). General linear models were used to investigate whether cognitive tasks sensitive to iNPH and AD could predict the identified MRI abnormalities.ResultAD patients demonstrated significant impairments in all tasks, while iNPH patients scored significantly lower than controls in motor control, processing speed, visuospatial, and executive function tasks (Figure 2). One‐way multivariate ANOVA identified cognitive differences between groups (F(22,554)= 14.47, p<0.001), with iNPH performing significantly better on memory and language tests relative to AD. Compared to HC, iNPH patients had lower FA in the corpus callosum and skeleton, reduced cingulate cortex volume and higher total ventricular volume. AD patients, instead, had reduced FA in the inferior longitudinal fasciculus and skeleton, with lower temporal and parietal volumes relative to HC (Figure 3). For iNPH, target detection and words delayed recall predicted corpus callosum FA, while visuospatial tasks predicted ventricular and cingulate volumes. In the AD group, no statistically significant link was detected between task performance and structural changes following multiple comparisons correction.ConclusionRemote computerised cognitive testing can discriminate between AD and iNPH and has the potential to predict disease‐related MRI abnormalities.