Dual Semi-Supervised Learning for Classification of Alzheimer's Disease and Mild Cognitive Impairment Based on Neuropsychological Data

Deep learning has shown impressive diagnostic abilities in Alzheimer's disease (AD) research in recent years. However, although neuropsychological tests play a crucial role in screening AD and mild cognitive impairment (MCI), there is still a lack of deep learning algorithms using only such basic diagnostic methods. This paper proposes a novel semi-supervised method using neuropsychological test scores with scarce labeled data, introducing difference regularization and consistency regularization with pseudo-labeling. A total of 188 AD, 402 MCI, and 229 normal controls (NC) were enrolled from the ADNI database. We first selected the 15 features most associated with diagnostic outcomes via feature selection among seven neuropsychological tests. Next, we proposed a dual semi-supervised learning (DSSL) framework that uses two encoders to learn two different feature vectors. Sets of 60 and 120 diagnosed subjects were randomly selected as labeled training data. Experimental results show that DSSL achieves the best accuracy and stability in classifying AD, MCI, and NC (85.47% accuracy with 60 labels and 88.40% with 120 labels) compared to other semi-supervised methods, providing clinical insight for physicians to diagnose AD and MCI.

Francesco Di Lorenzo, Annibale Antonioni, Yan Wang, Xuming Gu, Wenju Hou, Meng Zhao, Li Sun, Chunjie Guo