Table 5 Pharmacological treatment response prediction
Author | Features | Algorithm | Validation |
|---|---|---|---|
Jaworska et al. [72] | EEG and eLORETA | Random forests | Tenfold cross-validation |
Browning et al. [108] | Initial QIDS-R and face-based emotional recognition task (FERT) | Linear SVM | External validation on unseen data |
Pei et al. [109] | EEG and genetic markers | Linear SVM | Leave-one-out cross-validation |
Chang et al. [16] | MRI and genetic markers | Artificial neural network | Holdout set and k-fold cross-validation for hyperparamater tuning |
Tian et al. [105] | fMRI | Linear support vector machine | Leave-one-out cross-validation |
Carrillo et al. [10] | Speech data | Gaussian Naive Bayes | Sevenfold cross-validation |
Lin et al. [110] | Genetic markers | Multilayer feedforward neural network | 10 iterations of tenfold cross-validation |
Mumtaz et al. [111] | EEG | Logistic regression | 100 iterations of tenfold cross-validation |
Chekroud et al. [112] | Sociodemographic, questionnaires (such as HAMD), clinical information | Gradient boosting machine | 10 iterations of tenfold cross-validation and externally validated on unseen data |
Patel et al. [113] | Demographic and neuroimaging | Alternating decision trees | Leave-one-out cross-validation |
Khodayari-Rostamabad et al. [15] | Pretreatment EEG | Mixture of factor analysis | 100 iterations of leave N out cross-validation |