JMIR Publications

ABSTRACT

Background:

Dementia has become a major public health concern due to its heavy disease burden. Mild cognitive impairment (MCI) is a transitional stage between healthy aging and dementia. Early identification of MCI is an essential step of dementia prevention.

Objective:

Based on machine-learning methods, this study aimed to develop and validate a stable and scalable panel of cognitive tests for the early detection of MCI and dementia based on the Chinese Neuropsychological Consensus Battery (CNCB) in the Chinese Neuropsychological Normative Project (CN-NORM) cohort.

Methods:

The CN-NORM was a nationwide, multicenter study conducted in China with 876 participants, including an MCI group (n = 358), a dementia group (n = 189) and a cognitively normal group (CN, n = 358). Based on each test z score, we used the following four algorithms to select candidate variables: the F score according to the SelectKBest method, area under the curve (AUC) from logistic regression (LR), p values from the logit method, and backward stepwise elimination. Based on the results, different models were constructed after considering the administration duration and complexity of combinations of various tests. The performance of each model was assessed in terms of discrimination and calibration. Receiver-operating characteristic (ROC) curve and AUC metrics were used to evaluate the discriminative ability of the models via stratified sampling cross-validation and LR and support vector classification (SVC) algorithms to reduce the variance and bias of models. This stable and scalable model was further validated in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort, which included 420 CN subjects, 237 patients with MCI and 90 patients with dementia.

Results:

We developed a stable and scalable composite neurocognitive test based on machine learning to discriminate patients with MCI from CN subjects and dementia patients. This composite test consisted of the Hopkins verbal learning test-5 minutes recall and Trail making test-B, it was time efficient and easily administered. This composite test achieved similar discrimination and better calibration in distinguishing between MCI and CN individuals as well as between MCI and dementia individuals after independent verification by LR and SVC algorithms. These results were further validated in the ADNI cohort. Therefore, this test has both high scalability and high stability for the early discrimination of dementia.

Conclusions:

We developed a stable and scalable composite neurocognitive test based on machine learning that could not only differentiate between MCI patients and controls but can between patients with different stages of cognitive impairment. This composite neurocognitive test is a feasible and practical digital biomarker that can potentially be used in large-scale cognitive screening and intervention studies.