From: A review of brain imaging biomarker genomics in Alzheimer’s disease: implementation and perspectives
Feature name | Calculation formula | |
---|---|---|
First-order features | SUVR | \(SUVR_{mean} = \frac{{I_{avg\_ROIC} }}{{I_{avg\_ref} }}\) |
FA | \(\sqrt {\frac{{(\lambda_{1} - \lambda_{2} )^{2} + (\lambda_{1} - \lambda_{3} )^{2} + (\lambda_{2} - \lambda_{3} )^{2} }}{{2(\lambda_{1} + \lambda_{2} + \lambda_{3} )^{2} }}}\) | |
Skewness | \(\sigma^{ - 3} \mathop \sum \limits_{i = 1}^{{N_{g} }} \left( {i - \mu } \right)^{3} p\left( i \right)\) | |
Kurtosis | \(\sigma^{ - 4} \mathop \sum \limits_{i = 1}^{{N_{g} }} [\left( {i - \mu } \right)^{4} p\left( i \right)] - 3\) | |
Variance | \(\mathop \sum \limits_{i = 1}^{{N_{g} }} \left( {i - \mu } \right)^{2} p\left( i \right)\) | |
Other First-order features: cortical thickness; grey matter volume (sMRI features); ALFF, fALFF, ReHo, FC (fMRI signals); MD, radD, axD (DTI diffusion parameters); clustering coefficient, characteristic path length, small-worldness, global efficiency, transitivity, assortativity coefficient, modularity (various network parameters); and so on | ||
High-dimensional features | Energy | \(\mathop \sum \limits_{i = 1}^{{N_{g} }} \mathop \sum \limits_{j = 1}^{{N_{g} }} \left[ {p\left( {i,j} \right)} \right]^{2}\) |
Strength | \(\frac{{\mathop \sum \nolimits_{i = 1}^{{N_{g} }} \mathop \sum \nolimits_{i = 1}^{{N_{g} }} \left( {n_{i} + n_{j} } \right)\left( {i - j} \right)^{2} }}{{\left[ {\varepsilon + \mathop \sum \nolimits_{i = 1}^{{N_{g} }} s\left( i \right)} \right]}},n_{i} \ne 0,n_{j} \ne 0\) | |
Entropy | \(\mathop \sum \limits_{i = 1}^{{N_{g} }} \mathop \sum \limits_{j = 1}^{{N_{g} }} p\left( {i,j} \right)log\left( {p\left( {i,j} \right)} \right)\) | |
GLN | \(\mathop \sum \limits_{i = 1}^{{N_{g} }} (\mathop \sum \limits_{j = 1}^{{N_{r} }} r\left( {i,j} \right))^{2}\) | |
LRHGE | \(\mathop \sum \limits_{i = 1}^{{N_{g} }} \mathop \sum \limits_{j = 1}^{{N_{r} }} i^{2} j^{2} r\left( {i,j} \right)\) | |
GLV | \(\frac{1}{{N_{g} \times N_{r} }}\mathop \sum \limits_{i = 1}^{{N_{g} }} \mathop \sum \limits_{j = 1}^{{N_{r} }} \left( {ir\left( {i,j} \right) - \mathop \sum \limits_{i = 1}^{{N_{g} }} i\mathop \sum \limits_{j = 1}^{{N_{r} }} r\left( {i,j} \right)} \right)^{2}\) | |
Other High-dimensional features are based on other analytical methods |