White matter damage due to vascular, tau, and TDP-43 pathologies and its relevance to cognition

Table 2 Final parsimonious models evaluating the utility of neuroimaging measures in predicting cognitive performance

Variable	Estimate (s.e.)	p value	Partial R²
MCSA Global Cognition (Model R² = 0.546)
Intercept	0.48 (1.08)	0.66
Age	− 0.06 (0.009)	< 0.001	0.121
Male	− 0.23 (0.10)	0.031	0.014
Education	0.17 (0.02)	< 0.001	0.178
Visit Number	0.11 (0.02)	< 0.001	0.090
Amyloid	− 1.23 (0.24)	< 0.001	0.074
Genu ISOVF	− 9.16 (2.20)	< 0.001	0.049
Genu NDI	4.65 (1.35)	< 0.001	0.034
MCSA MMSE (Model R² = 0.282)
Intercept	23.26 (1.35)	< 0.001
Education	0.19 (0.03)	< 0.001	0.092
Amyloid	− 1.78 (0.36)	< 0.001	0.066
Genu ISOVF	− 11.91 (3.12)	< 0.001	0.041
Genu NDI	8.01 (2.09)	< 0.001	0.041
ADRC MMSE (Model R² = 0.209)
Intercept	4.15 (9.45)	0.66
Tau	− 5.99 (2.41)	0.016	0.096
ITWM NDI	45.08 (19.40)	0.024	0.085

ISOVF, isotropic volume fraction; CGH, parahippocampal cingulum; ITWM, inferior temporal white matter, NDI, neurite density index, MMSE, mini mental state examination. The initial had all NODDI measures but only these variables
Significant predictors of cognition are shown in the parsimonious models. The models included all of these as potential predictors: age, male, education, visit number, amyloid, tau, genu ISOVF, genu NDI, ITWM NDI, and ITWM CGH

ISSN: 2051-5960