Multi-dimensional persistent feature analysis identifies connectivity patterns of resting-state brain networks in Alzheimer’s disease

Abstract

Objective. The characterization of functional brain network is crucial to understanding the neural mechanisms associated with Alzheimer’s disease (AD) and mild cognitive impairment (MCI). Some studies have shown that graph theoretical analysis could reveal changes of the disease-related brain networks by thresholding edge weights. But the choice of threshold depends on ambiguous cognitive conditions, which leads to the lack of interpretability. Recently, persistent homology (PH) was proposed to record the persistence of topological features of networks across every possible thresholds, reporting a higher sensitivity than graph theoretical features in detecting network-level biomarkers of AD. However, most research on PH focused on zero-dimensional features (persistence of connected components) reflecting the intrinsic topology of the brain network, rather than one-dimensional features (persistence of cycles) with an interesting neurobiological communication pattern. Our aim is to explore the multi-dimensional persistent features of brain networks in the AD and MCI patients, and further to capture valuable brain connectivity patterns. Approach. We characterized the change rate of the connected component numbers across graph filtration using the functional derivative curves, and examined the persistence landscapes that vectorize the persistence of cycle structures. After that, the multi-dimensional persistent features were validated in disease identification using a K-nearest neighbor algorithm. Furthermore, a connectivity pattern mining framework was designed to capture the disease-specific brain structures. Main results. We found that the multi-dimensional persistent features can identify statistical group differences, quantify subject-level distances, and yield disease-specific connectivity patterns. Relatively high classification accuracies were received when compared with graph theoretical features. Significance. This work represents a conceptual bridge linking complex brain network analysis and computational topology. Our results can be beneficial for providing a complementary objective opinion to the clinical diagnosis of neurodegenerative diseases.

Development of the Functional Connectome Topology in Adolescence: evidence from Topological Data Analysis

“…Regarding the average Betti curves, the sample intercept for the B 0 curve showed an inverse sigmoid pattern that replicates previous findings in functional connectivity fMRI (Liang and Wang, 2017; Gracia-Tabuenca et al, 2020; Li et al, 2021) and PET (Lee et al, 2012) studies. Furthermore, the average B 0 curve of the randomized data reached the single component faster than the observed data, evincing a less segregated network.…”

Development of the Functional Connectome Topology in Adolescence: Evidence from Topological Data Analysis

“…Regarding the average Betti curves, the sample intercept for the B 0 curve showed an inverse sigmoid pattern that replicates previous findings in functional connectivity fMRI ( Liang and Wang, 2017 ; Gracia-Tabuenca et al, 2020 ; Li et al, 2021 ) and PET ( H. Lee et al, 2012 ) studies. Furthermore, the average B 0 curve of the randomized data reached the single component faster than the observed data, evincing a less distributed network.…”

“…In terms of the functional organization of the brain, previous cross-sectional studies have shown that the adolescent period is characterized by an increase in modularity and specialization ( Fair et al, 2009 ; Satterthwaite et al, 2013a ; Gu et al, 2015 ), with prominent effects in frontal and parietal systems, along with executive performance ( Marek et al, 2015 ; Gracia-Tabuenca et al, 2021 ). However, as far as we are concerned, TDA in human connectomes has mainly been applied to neuropsychiatric disorders ( H. Lee et al, 2012 , 2017 ; Gracia-Tabuenca et al, 2020 ; Li et al, 2021 ) but not to characterize the typical development. There is still a huge degree of incertitude in this field because of the great variability between samples, sexes, and cultures ( Sawyer et al, 2018 ), with special emphasis on the fact that some individuals have faster or slower pubertal development even when they have the same chronological age ( Blakemore et al, 2010 ; Vijayakumar et al, 2018 ).…”

Multi-scale discriminative regions analysis in FDG-PET imaging for early diagnosis of Alzheimer’s disease

“…Improvements within medical imaging have created new opportunities regarding both diagnosing and understanding many types of neurodegenerative diseases [8,9], one of which is AD. Since various neuroimaging technologies can differentiate neuropathological alterations, they have been widely used for AD diagnosis [10][11][12][13][14][15]. Fluorodeoxyglucose positron emission tomography (FDG-PET) imaging is one of the effective functional biomarkers for AD diagnosis by indicating glucose metabolism activity and distribution.…”