Multiclass characterization of frontotemporal dementia variants via multimodal brain network computational inference

Abstract

Characterizing a particular neurodegenerative condition against others possible diseases remains a challenge along clinical, biomarker, and neuroscientific levels. This is the particular case of frontotemporal dementia (FTD) variants, where their specific characterization requires high levels of expertise and multidisciplinary teams to subtlety distinguish among similar physiopathological processes. Here, we used a computational approach of multimodal brain networks to address simultaneous multiclass classification of 298 subjects (one group against all others), including five FTD variants: behavioral variant FTD, corticobasal syndrome, nonfluent variant primary progressive aphasia, progressive supranuclear palsy, semantic variant primary progressive aphasia; and healthy controls. Fourteen machine learning classifiers were trained with functional and structural connectivity calculated through different methods. Due to the large number of variables, dimensionality was reduced employing statistical comparisons and progressive elimination to assess feature stability under nested cross validation. The machine learning performance was measured through the area under the receiver operating characteristic curves, reaching 0.81 on average, with a standard deviation of 0.09. Furthermore, the contributions of demographic and cognitive data were also assessed via multifeatured classifiers. An accurate simultaneous multiclass classification of each FTD variant against other variants and controls was obtained, based on the selection of an optimum set of features. The classifiers incorporating brain’s network and cognitive assessment increased performance metrics. Multimodal classifiers evidenced specific variants’ compromise, across modalities and methods through the feature importance analysis. If replicated and validated, this approach may potentially help to support clinical decision tools aimed to detect specific affectations in the context of overlapping diseases.

Model-based whole-brain perturbational landscape of neurodegenerative diseases

“…Finally, despite our choice of the AAL parcellation being based on its widespread adoption in dementia research ( Agosta et al, 2013 ; Lord et al, 2016 ; Reyes et al, 2018 ; Sedeño et al, 2017 ; Whitwell et al, 2011 ), future research may require a systematic exploration of other brain parcellations applied to the dementia population. In that direction, to compare the differences across parcellations, a recent study by Gonzalez et al used the AAL and HCP atlas ( Glasser et al, 2016 ) parcellations on a dementia subsample (graph connectivity and graph multifeature in both modalities) and did not find significant differences ( Gonzalez-Gomez et al, 2023 ).…”

Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI

“…We used structural features from T1-weighted MRI in the current study to derive differential features for detecting subtypes within FTD. Extension of current work could involve additional neuroimaging modalities such as diffusion tensor imaging (DTI) ( Torso et al, 2020 ) or functional MRI (fMRI) ( Gonzalez-Gomez et al, 2023 ). Another future direction for dealing with limited features would be to use a self-supervised approach as a feature extractor, to be trained on larger datasets, to extract disease-agnostic generalized neuroimaging features in lower dimensions, and then train a using the low-dimension representation space ( Krishnan et al, 2022 ; Tang et al, 2022 ; Huang et al, 2023 ).…”