Skip to main content
. 2018 Apr 25;10:111. doi: 10.3389/fnagi.2018.00111

Table 3.

Accuracy, balanced accuracy, and sensitivities [%] for all diagnostic groups, using different subsets of the data sources.

Feature set Acc. Bal. Acc. Sens. CN Sens. AD Sens. FTLD Sens. DLB Sens. VaD
NP 62.3 57.3 83.1 61.9 48.9 46.8 45.8
CSF 51.2 40.6 40.7 72.2 35.9 12.8 41.7
VMRI 45.8 54.5 68.6 26.9 57.6 36.2 83.3
AMRI 66.3 66.1 78.8 63.7 68.5 31.9 87.5
NP and CSF 67.1 59.7 83.1 69.5 55.4 53.2 37.5
NP and VMRI 72.2 74.0 90.7 64.6 67.4 68.1 79.2
NP and AMRI 78.0 77.1 91.5 76.2 70.7 63.8 83.3
CSF and VMRI 63.9 62.3 63.6 69.5 58.7 40.4 79.2
CSF and AMRI 71.2 72.5 79.7 68.2 71.7 55.3 87.5
VMRI and AMRI 68.3 70.0 77.1 64.6 69.6 51.1 87.5
NP, CSF, and VMRI 75.8 73.7 89.0 73.1 71.7 68.1 66.7
NP, CSF, and AMRI 83.3 82.9 92.4 83.0 75.0 76.6 87.5
NP, VMRI, and AMRI 77.2 77.8 89.0 75.8 66.3 70.2 87.5
CSF, VMRI, and AMRI 71.0 74.5 78.0 67.3 67.4 68.1 91.7
All 81.5 82.3 89.0 80.3 76.1 74.5 91.7

NP, neuropsychological tests; CSF, cerebrospinal fluid based biomarkers; VMRI, visual MRI ratings; AMRI, automatic MRI quantifications; Sens., sensitivity for each diagnostic group, i.e., the proportion of patients that are correctly classified in that group; CN, control.