Multimodal classification of Dementia using functional data, anatomical features and 3D invariant shape descriptors

Abstract

Multimodality classification of Alzheimer’s disease (AD) and its prodromal stage, Mild Cognitive Impairment (MCI), is of interest to the medical community. We improve on prior classification frameworks by incorporating multiple features from MRI and PET data obtained with multiple radioligands, fluorodeoxyglucose (FDG) and Pittsburg compound B (PIB). We also introduce a new MRI feature, invariant shape descriptors based on 3D Zernike moments applied to the hippocampus region. Classification performance is evaluated on data from 17 healthy controls (CTR), 22 MCI, and 17 AD subjects. Zernike significantly outperforms volume, accuracy (Zernike to volume): CTR/AD (90.7% to 71.6%), CTR/MCI (76.2% to 60.0%), MCI/AD (84.3% to 65.5%). Zernike also provides comparable and complementary performance to PET. Optimal accuracy is achieved when Zernike and PET features are combined (accuracy, specificity, sensitivity), CTR/AD (98.8%, 99.5%, 98.1%), CTR/MCI (84.3%, 82.9%, 85.9%) and MCI/AD (93.3%, 93.6%, 93.3%).

Identifying regions of interest for discriminating Alzheimer's disease from mild cognitive impairment

“…To overcome this problem, dimensionality reduction and feature selection techniques are very important, and have been applied to the diagnosis of AD [5,6]. Some techniques segment the brain in regions of interest (ROIs) associated with atrophy caused by the disease [7,8].…”

Amyloid‐β PET in Alzheimer's disease: A systematic review and Bayesian meta‐analysis

“…Two hundred eighty‐seven articles were excluded through the filter at the official PubMed website, and 1420 articles were further excluded by reading the titles and abstracts. After carefully reading the full text and supplementary materials and judging the accuracy and reliability of the extracted data, the final number of articles included in our study was 48 (Beach et al., 2014; Brück et al., 2013; Camus et al., 2012; Clark et al., 2012; Fleisher, 2011; Hatashita & Yamasaki, 2013; Hatashita et al., 2014; Hosokawa et al., 2015; Jack et al., 2012; Kaneko et al., 2014; Kerbage et al., 2015; Li et al., 2015; Mattsson et al., 2014; Mikhno et al., 2012; Newberg et al., 2012; Ng et al., 2007; Rabinovici et al., 2011; Saint‐Aubert et al., 2014; Tolboom et al., 2010; Trzepacz et al., 2014; Tzen et al., 2014; Vandenberghe et al., 2010; Villemagne et al., 2011, 2019; Alvarez et al., 2018; Ben Bouallegue et al., 2017; Chen et al., 2016; Dukart et al., 2016; Iaccarino et al., 2017; La Joie et al., 2019; Mielke et al., 2018; Oliveira et al., 2018; Ottoy et al., 2019; Park et al., 2019; Schreiber et al., 2015; Seo et al., 2017; Takahashi et al., 2017; Villeneuve et al., 2015; Wang, Chen, et al., 2016; Wang, Yi, et al., 2016; Xu et al., 2016; Zhang et al., 2017; Zwan et al., 2016, 2021; Chanisa et al., 2021; Kitajima et al., 2021; Lesman‐Segev et al., 2021; Peretti et al., 2019). The specific process of literature screening is detailed in Figure 1.…”

Mindboggling morphometry of human brains

“…Sulcal width has been used to differentiate between groups with mild cognitive impairment [50] and global and local gyrification indices computed from sulci have been used to characterize schizophrenia [51] and earlyonset vs. intermediate-onset bipolar disorder as well as bipolar and unipolar depression [52][53][54]. More abstract shape measures such as Zernike moments (see below) have been used in patient classification, such as to distinguish cases of dementia [55].…”

“…Geometric moments of 3-D coordinates have been used to construct shape descriptors for human brain morphometry [109] because of desirable characteristics such as invariance to rotation, symmetry, and scale, and they can be computed for any topology. Zernike moments [110] have also been applied to human brain morphometry for classifying dementia patients [55] and confer several advantages over geometric moments. They form a set of orthogonal descriptors, where each descriptor contains independent information about the structure, allowing the original shape to be reconstructed from the moments.…”