Table 2.
Representative ML applications in psychiatry based on neuroimaging and clinical data
| Application | Learning category | ML method | Mental disorder | Data type | Reference |
|---|---|---|---|---|---|
| Diagnosis | supervised classification, deep | dynamic GCN | ADHD | rs-fMRI + phenotypic data | Zhao et al.71 |
| supervised classification | ensemble learning | ADHD | multi-modal | Luo et al.72 | |
| supervised classification, deep | GCN | ASD | task fMRI | Li et al.73 | |
| supervised classification, deep | ensemble learning + GCN | ASD | rs-fMRI | Khosla et al.74 | |
| supervised classification | PCA + LASSO | bipolar | dMRI + cognitive data | Wu et al.75 | |
| supervised classification | RVM | PTSD | rs-fMRI | Zhu et al.76 | |
| supervised classification, deep | ICA + LSTM | schizophrenia | fMRI | Yan et al.77 | |
| supervised classification | SVM | schizophrenia | sMRI | Mikolas et al.78 | |
| supervised classification, deep | CNN | depression | rs-EEG | Uyulan et al.79 | |
| supervised classification, deep | autoencoder + DNN, SVM, random forest | ASD | rs-fMRI | Heinsfeld et al.80 | |
| semi-supervised classification | GNN | ASD | rs-fMRI + phenotypic data | Parisot et al.81 | |
| unsupervised, subtyping | normative modeling + clustering | PTSD | rs-fMRI | Maron-Katz et al.82 | |
| unsupervised, subtyping | CCA + hierarchical clustering | depression | rs-fMRI | Drysdale et al.15 | |
| unsupervised, subtyping | sparse K-means | PTSD, depression | rs-EEG | Zhang et al.42 | |
| unsupervised, subtyping | latent class analysis | ADHD | task fMRI | Lecei et al.83 | |
| supervised, transdiagnostic | normative modeling + GP regression | multiple disorders | rs-fMRI | Parkes et al.84 | |
| unsupervised, transdiagnostic | sparse CCA | multiple disorders | rs-fMRI | Xia et al.51 | |
| supervised, transdiagnostic | PLS | multiple disorders | rs-fMRI | Kebets et al.52 | |
| Prognosis | supervised classification | GP classifier | depression | task fMRI | Schmaal et al.85 |
| supervised classification | LASSO | psychosis | rs-EEG | Ramyead et al.86 | |
| supervised classification | SVM | psychosis, depression | multi-modal | Koutsouleris et al.87 | |
| supervised classification, deep | DNN | PTSD | rs-fMRI/task fMRI | Sheynin et al.88 | |
| supervised classification | SVM | schizophrenia | sMRI | Nieuwenhuis et al.89 | |
| supervised classification, deep | SVM, random forest, DNN | schizophrenia | task fMRI | Smucny et al.90 | |
| supervised regression | LASSO | substance use | MRI/task fMRI | Bertocci et al.91 | |
| supervised regression, deep | SVR + LSTM | PTSD | MEG | Zhang et al.92 | |
| Treatment prediction | supervised classification | SVM | ADHD | sMRI | Chang et al.93 |
| supervised classification | SVM | psychosis | sMRI | Koutsouleris et al.94 | |
| supervised classification | GP classifier | PTSD | MRI/rs-fMRI | Zhutovsky et al.95 | |
| supervised classification | SVM | schizophrenia | rs-fMRI | Cao et al.96 | |
| supervised classification | SVM | depression | rs-EEG | Zhdanov et al.97 | |
| supervised classification | SVM + GP classifier | depression | sMRI | Redlich et al.98 | |
| supervised regression | latent space learning | depression | rs-EEG | Wu et al.99 | |
| supervised regression | RVM | depression | task fMRI | Fonzo et al.100 | |
| supervised regression | MVPA | ASD | task fMRI | Yang et al.101 | |
| supervised regression | LASSO | anxiety | rs-fMRI | Reggente et al.102 | |
| Readmission | supervised classification | SVM | depression | multi-modal | Cearns et al.67 |
| supervised classification | growth mixture modeling | depression | clinical data | Gueorguieva et al.68 | |
| supervised classification | decision tree | bipolar | EHR | Edgcomb et al.103 | |
| supervised classification | ensemble learning | substance use | phenotypic data | Morel et al.104 |