A combination of dual-tree discrete wavelet transform and minimum redundancy maximum relevance method for diagnosis of Alzheimer's disease

Assessment of Screening Approach in Early and Differential Alzheimer's Disease Diagnosis

Background: Alzheimer’s disease (AD) is the leading cause of dementia in the elderly population. Currently, diagnosis is based on invasive and expensive techniques, so there is a growing need to look for other possible tests, as well as carry out clinical validation. Studies from the literature showed potential diagnosis models, including some AD risk factors (age, gender, ApoE-ε4 genotype) and other variables (biomarkers levels, neuroimaging). Specifically, a recent model was performed from lipid peroxidation compounds in plasma samples to identify patients with early AD. However, there is a lack of studies about clinical validation of these preliminary diagnosis models. Methods: Plasma samples from participants classified into AD (n = 61), non-AD (n = 17), and healthy (n = 44) were analyzed. In fact, lipid peroxidation compounds were determined by liquid chromatography and mass spectrometry. Then, a previously developed diagnosis model was clinically validated, evaluating some diagnosis indexes. Results: The validation of the preliminary diagnosis model showed satisfactory diagnosis indexes (accuracy 77%, sensitivity 89%, specificity 61%, diagnostic odds ratio 12.5, positive predictive value 76%). Next, a useful screening tool, including the ApoE genotype, was developed, identifying patients with a higher risk of developing AD and improving the corresponding diagnosis indexes (accuracy 82%, sensitivity 81%, specificity 85%, diagnostic odds ratio 23.2, positive predictive value 90.5%). Conclusion: A new screening approach could improve the early, minimally invasive, and differential AD diagnosis in the general population.

Improved Measures of Redundancy and Relevance for mRMR Feature Selection

Many biological or medical data have numerous features. Feature selection is one of the data preprocessing steps that can remove the noise from data as well as save the computing time when the dataset has several hundred thousand or more features. Another goal of feature selection is improving the classification accuracy in machine learning tasks. Minimum Redundancy Maximum Relevance (mRMR) is a well-known feature selection algorithm that selects features by calculating redundancy between features and relevance between features and class vector. mRMR adopts mutual information theory to measure redundancy and relevance. In this research, we propose a method to improve the performance of mRMR feature selection. We apply Pearson’s correlation coefficient as a measure of redundancy and R-value as a measure of relevance. To compare original mRMR and the proposed method, features were selected using both of two methods from various datasets, and then we performed a classification test. The classification accuracy was used as a measure of performance comparison. In many cases, the proposed method showed higher accuracy than original mRMR.