Table 5.

Summary of notations used to refer to comparative strategies evaluated in this work.

Symbol	Description
Classification Methods
${\mathcal{C}}_{CAS}$	Cascaded Classifier
${\mathcal{C}}_{OSC}$	One Shot Classifier
${\mathcal{C}}_{OVA}$	One-vs-All Classifier
${\mathcal{C}}_{BIN}$	Binary Classifier
Data Fusion Methods
${ℱ}_{SMVCCA}$	Supervised Multiview CCA data fusion approach
${ℱ}_{PCA}$	Principal component analysis of concatenated features as baseline fusion approach
Modality Selection
${\mathcal{K}}_{ALL}$	Multimodal dataset comprising all modalities considered in this study
${\mathcal{K}}_{SEL}$	Multimodal dataset comprising select subset of all available modalities
${\mathcal{K}}_{MRI}$	Unimodal dataset containing quantitative attributes extracted from MRI
${\mathcal{K}}_{PET}$	Unimodal dataset containing quantitative attributes extracted from PET
${\mathcal{K}}_{CSF}$	Unimodal dataset containing proteomic measurements from CSF
${\mathcal{K}}_{PP}$	Unimodal dataset containing plasma proteomic data
${\mathcal{K}}_{APOE}$	Unimodal dataset containing APOE data
${\mathcal{K}}_{ADAS}$	Unimodal dataset containing ADAS-Cog scores
Comparative Methods
${\mathscr{H}}_{MRI}$ (${\mathcal{C}}_{CAS}$ + ${\mathcal{K}}_{MRI}$)	Cascaded classification of single modality MR data
${\mathscr{H}}_{PET}$ (${\mathcal{C}}_{CAS}$ + ${\mathcal{K}}_{PET}$)	Cascaded classification of single modality PET data
${\mathscr{H}}_{CSF}$ (${\mathcal{C}}_{CAS}$ + ${\mathcal{K}}_{CSF}$)	Cascaded classification of single modality CSF data
${\mathscr{H}}_{PP}$ (${\mathcal{C}}_{CAS}$ + ${\mathcal{K}}_{PP}$)	Cascaded classification of single modality Plasma Proteomics data
${\mathscr{H}}_{APOE}$ (${\mathcal{C}}_{CAS}$ + ${\mathcal{K}}_{APOE}$)	Cascaded classification of single modality APOE data
${\mathscr{H}}_{ADAS}$ (${\mathcal{C}}_{CAS}$ + ${\mathcal{K}}_{ADAS}$)	Cascaded classification of single modality ADAS-Cog data
${\mathscr{H}}_{CAMCCO}$ (${\mathcal{C}}_{CAS}$ + ${ℱ}_{SMVCCA}$ + ${\mathcal{K}}_{SEL}$)	Classifier resulting from CaMCCo framework, which is comprised of cascaded classifier, sMVCCA data fusion and modality selection
${\mathscr{H}}_{ALL}$ (${\mathcal{C}}_{CAS}$ + ${ℱ}_{SMVCCA}$ + ${\mathcal{K}}_{ALL}$)	Cascaded classifier in combination with sMVCCA based data fusion method to combine all modalities
${\mathscr{H}}_{PCAL}$ (${\mathcal{C}}_{CAS}$ + ${ℱ}_{PCA}$ + ${\mathcal{K}}_{ALL}$)	Cascaded classifier with PCA reduced representation of data concatenated from all modalities
${\mathscr{H}}_{PCA}$ (${\mathcal{C}}_{CAS}$ + ${ℱ}_{PCA}$ + ${\mathcal{K}}_{SEL}$)	Cascaded classifier with PCA reduced representation of data concatenated from selected subset of modalities
${\mathscr{H}}_{OVA}$ (${\mathcal{C}}_{OVA}$ + ${ℱ}_{SMVCCA}$ + ${\mathcal{K}}_{SEL}$)	One-vs-all classifier constructed from sMVCCA fused data from selected modalities.
${\mathscr{H}}_{OSC}$ (${\mathcal{C}}_{OSC}$ + ${ℱ}_{SMVCCA}$ + ${\mathcal{K}}_{SEL}$)	One shot classifier constructed from sMVCCA fused data from selected modalities.
${\mathscr{H}}_{BIN}$ (${\mathcal{C}}_{BIN}$ + ${ℱ}_{SMVCCA}$ + ${\mathcal{K}}_{SEL}$)	Binary classifier constructed from sMVCCA fused data from selected modalities.