Table 4.

Node-pair connections (i.e. edges) that explained a significant portion of variance in RSA and SCL in the regression analyses^a

	Edge category	Edge label	Zero-order correlation	β	R2 change	p value of F change
RSA
Positive correlation	Left vAI-left ACC	2–5	0.41	0.48	0.22	0.002
	Left vAI-right ACCb	2–16	0.33	0.51	0.14	0.01
	Right vAI-right ACC	13–15	0.29	0.83	0.12	0.03
Negative correlation	Left ACC-left hypothalamus	7–9	−0.57	−0.56	0.20	0.01
	Left ACC-right hypothalamus	7–19	−0.57	−0.52	0.30	0.001
	Right ACC-right hypothalamusb	16–19	−0.33	−0.37	0.12	0.01
	Right vAI1-right vAI2	12–13	−0.32	−0.61	0.09	0.04
	Right ACC-right hypothalamusc	17–19	−0.30	−0.43	0.06	0.04
	Right vAI-right hypothalamusb,c	13–19	−0.26	0.11	0.21	0.01
	Right vAI-right ACCb	12–17	−0.25	−0.32	0.06	0.01
	Left ACC-right amygdalab	6–18	−0.22	−0.42	0.09	0.02
SCL
Positive correlation	Left ACC-right hypothalamus	6–19	0.31	0.65	0.14	0.02
	Right ACC-right amygdala	16–18	0.23	0.38	0.08	0.04
Negative correlation	Left dAI-PAG	1–21	−0.37	−0.54	0.25	0.003
	Right vAI1-right vAI2	12–13	−0.31	−0.39	0.10	0.03
	Right ACC-left ACC	17–5	−0.30	−0.40	0.12	0.01
	Right dAI-PAG	11–21	−0.28	−0.71	0.25	0.003
	Left aMCC-right hypothalamus	4–19	−0.20	−0.59	0.16	0.02

^aForward-entry hierarchical regression analyses in 15 patients with bvFTD and 13 HCs (controlling for age, sex, education, and diagnosis) were used to determine which edges were significant predictors of baseline PANS and SANS. Edges with positive and negative associations with RSA and SCL entered the models, suggesting that a combination of strong and weak functional connectivity in the ANS network is necessary for resting physiological activity.

^bEdge that lost significance when controlling for mean resting respiration period and interbeat interval.

^cEdges that lost significance when controlling for the mean gray matter volume of the nodes in that edge.