Table 4 Estimated parameter values and the associated 95% confidence intervals [σ, σ+] derived from the profile likelihood for subject 1.

From: Bridging the gap between measurements and modelling: a cardiovascular functional avatar

Parameter

Estimated value

PL-based confidence interval

Identifiability

σ

σ+

Pulmonary venous system

Ppu

12

11.07

12.92

Identifiable

Heart parameters

Left atrium (LA)

Ks,LA

1.19·10−8

−∞

Structurally non-identifiable

Emin,LA

0.12

0.09

0.14

Identifiable

Emax,LA

0.15

0.15

0.2

Identifiable

V0,LA

3.22

Practically non-identifiable

m1,LA

1.29

Practically non-identifiable

m2,LA

8.30

Practically non-identifiable

α1,LA

0.11

Practically non-identifiable

α2,LA

0.15

 

Practically non-identifiable

Rvisc,LA

8.46·10−5

Structurally non-identifiable

OnsetLA

0.75

0.75

0.85

Identifiable

Left ventricle (LV)

Ks,LV

2.09·10−9

Structurally non-identifiable

Emin,LV

0.10

0.06

0.14

Identifiable

m1,LV

1.66

1.48

1.82

Identifiable

m2,LV

35.89

27.54

49.29

Identifiable

α1,LV

0.41

0.30

0.50

Identifiable

α2,LV

0.48

0.38

0.58

Identifiable

Rvisc,LV

 

Structurally non-identifiable

OnsetLV

−0.06

−0.07

−0.06

Identifiable

Mitral valve

Rmv

6.26·10−3

Practically non-identifiable

Lmv

5.9·10−4

2.67·10−4

Practically non-identifiable

Aortic valves

Lav

1.64·10−4

1.24·10−3

Practically non-identifiable

Systemic arterial system

Raa

0.08

0.06

0.13a

Practically non-identifiable

Laa

1·10−4

1.46·10−3

Practically non- identifiable

Raav

5·10−3

Structurally non-identifiable

Caa

0.11

0.02

0.21

Identifiable

Cpsa

0.31

0.14

0.31

Identifiable

Cpia

0.06

0.29

Practically non-identifiable

Rda

0.08

0.05

0.13a

Practically non-identifiable

Lda

2.15·10−3

1.17·10−3

Practically non-identifiable

Rdav

5·10−3

Structurally non-identifiable

Cda

0.21

0.07

0.34

Identifiable

  1. aThe confidence boundary is computed outside the parameter boundary used in the optimization. Empty confidence boundaries in one (or both) directions indicate that no crosses with the threshold were found within the parameter boundaries used for the optimization. Parameters with a confidence interval [∞,∞] had a flat profile likelihood for any arbitrary value of the parameter.