Abstract:Alzheimer's disease (AD) is a common form of dementia. Early and differential diagnosis of AD has always been an arduous task for the medical expert due to the unapparent early symptoms and the currently imperfect imaging examination methods. Therefore, obtaining reliable markers with clinical diagnostic value in easily assembled samples is worthy and significant. Our previous work with laser Raman spectroscopy (LRS), in which we detected platelet samples of different ages of AD transgenic mice and non-transgenic controls, showed great effect in the diagnosis of AD. In addition, a multilayer perception network (MLP) classification method was adopted to discriminate the spectral data. However, there were disturbances, which were induced by noise from the machines and so on, in the data set; thus the MLP method had to be trained with large-scale data. In this paper, we aim to re-establish the classification models of early and advanced AD and the control group with fewer features, and apply some mechanism of noise reduction to improve the accuracy of models. An adaptive classification method based on the Gaussian process (GP) featured, with predictive probabilities, is proposed, which could tell when a data set is related to some kind of disease. Compared with MLP on the same feature set, GP showed much better performance in the experimental results. What is more, since the spectra of platelets are isolated from AD, GP has good expansibility and can be applied in diagnosis of many other similar diseases, such as Parkinson's disease (PD). Spectral data of 4 month and 12 month AD platelets, as well as control data, were collected. With predictive probabilities, the proposed GP classification method improved the diagnostic sensitivity to nearly 100%. Samples were also collected from PD platelets as classification and comparison to the 12 month AD. The presented approach and our experiments indicate that utilization of GP with predictive probabilities in platelet LRS detection analysis turns out to be more accurate for early and differential diagnosis of AD and has a wide application prospect.

Laser Raman Detection of Platelets for Early and Differential Diagnosis of Alzheimer’s Disease Based on an Adaptive Gaussian Process Classification Algorithm

Laser Raman Detection of Platelet As a Non-Invasive Approach for Early and Differential Diagnosis of Alzheimer's Disease

Raman Spectroscopy Detection of Platelet for Alzheimer’s Disease with Predictive Probabilities

Laser Tweezers Raman Spectroscopy Combined with Machine Learning for Diagnosis of Alzheimer's Disease.

Ultrasensitive and point-of-care detection of plasma phosphorylated tau in Alzheimer's disease using colorimetric and surface-enhanced Raman scattering dual-readout lateral flow assay

Early Detection of Alzheimer’s Disease Using Non-invasive Near-Infrared Spectroscopy

Laser Raman Detection for Oral Cancer Based on an Adaptive Gaussian Process Classification Method with Posterior Probabilities

Platelet-Derived Amyloid- Protein Precursor as a Biomarker of Alzheimer's Disease

Detection of Alzheimer's Disease by Raman Spectra of Rat's Platelet with a Simple Feature Selection

Platelet biomarkers for a descending cognitive function: A proteomic approach

Blood-based CNS Regionally and Neuronally Enriched Extracellular Vesicles Carrying Ptau217 for Alzheimer’s Disease Diagnosis and Differential Diagnosis

Acoustofluidic multimodal diagnostic system for Alzheimer's disease

Performance of Plasma Amyloid β, Total Tau, and Neurofilament Light Chain in the Identification of Probable Alzheimer's Disease in South China

Blood-based CNS regionally and neuronally enriched extracellular vesicles carrying pTau217 for Alzheimer's disease diagnosis and differential diagnosis

Early Rapid Diagnosis of Alzheimer’s Disease Based on Fusion of Near- and Mid-Infrared Spectral Features Combined with PLS-DA

High-Performance Plasma Biomarker Panel for Alzheimer's Disease Screening Using a Femtomolar-Level Label-Free Biosensing System

Precise Prediction of Cerebrospinal Fluid Amyloid Beta Protein for Early Alzheimer's Disease Detection Using Multimodal Data

Integrated Ultrasound‐Enrichment and Machine Learning in Colorimetric Lateral Flow Assay for Accurate and Sensitive Clinical Alzheimer's Biomarker Diagnosis

An explainable machine learning based prediction model for Alzheimer's disease in China longitudinal aging study

Blood-based Alzheimer’s Disease Diagnosis Using Fluorescent Peptide Nanoparticle Arrays

Optical Sensor Array for the Early Diagnosis of Alzheimer’s Disease