Prediction of Clinical Scores for Subjective Cognitive Decline and Mild Cognitive Impairment

Abstract

Mild cognitive impairment (MCI) is a neurological disorder that occurs in older adults involving cognitive impairments. It may occur as a transitional stage between normal aging and dementia such as Alzheimer’s disease (AD). Recent studies found that subjective cognitive decline (SCD) may be the early clinical precursor of dementia that precedes MCI. SCD individuals with normal cognition may already have some medial temporal lobe atrophy. This paper proposes a machine learning framework by combination of sparse coding and random forest to identify the informative biomarkers for prediction of clinical scores in SCD and MCI using structural magnetic resonance imaging (MRI). The volumetric features are computed from brain regions and the subregions of hippocampus and amygdala in MRIs. Then, sparse coding is applied to identify the relevant features. Finally, the proximity-based random forest is used to combine three sets of volumetric features and establish a regression model for predicting clinical scores. Our method has double feature selections to better explore the relevant features for prediction. Our method is evaluated with the T1-weighted structural MR images from 36 MCI, 112 SCD, 78 Normal Control (NC) subjects. The results demonstrate the effectiveness of proposed method.