Abstract:<section class="article-section__content"><h3 class="article-section__sub-title section1"> Background</h3><p>Due to the biological heterogeneity, 60%–70% of patients with major depressive disorder (MDD) do not respond to or achieve remission from first‐line antidepressants. Predicting neuroimaging biomarkers for early antidepressant treatment could guide initial antidepressant therapy.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Purpose</h3><p>To assess for neuroimaging biomarkers for antidepressant selection in early antidepressant treatment.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Study Type</h3><p>Prospective.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Subjects</h3><p>A total of 85 MDD patients from the major site and 33 MDD patients from an out‐of‐sample test site.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Field Strength/Sequence</h3><p>A 3.0 T, T1‐weighted imaging using a magnetization‐prepared rapid acquisition gradient‐echo sequence and diffusion tensor imaging (DTI) using an echo‐planar sequence.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Assessment</h3><p>Baseline DTI data of patients who achieved early improvement after 2‐weeks of antidepressant treatment (selective serotonin reuptake inhibitors [SSRI] or serotonin‐norepinephrine reuptake inhibitors [SNRI]) were analyzed. An ensemble model was constructed using data from the major site and then applied to assess the early response of patients at the out‐of‐sample test site.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Statistical Tests</h3><p>Support vector machine combined with leave‐one‐out cross‐validation were applied to construct the whole model from individual base models from different brain regions. Discriminative biomarkers were evaluated by calculating the changes in sensitivity and specificity obtained when removing a single base model from the whole model, the base model being removed changing in each run.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Results</h3><p>Training performance over MDD patients at the major site achieved 75% accuracy while performance with accuracy of 70% was achieved in the out‐of‐sample test site. Assessing sensitivity and specificity changes following the removal of single base models from the prominent model highlighted the functions of two neural circuitries: SSRI‐related emotion regulation circuitry, centered on the hippocampus (sensitivity changes: 10%) and amygdala (sensitivity changes: 11%); and SNRI‐related emotion and reward circuitry, centered on the putamen (specificity changes: 8%) and orbital part of superior frontal gyrus (specificity changes: 12%).</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Data Conclusion</h3><p>These findings support future research on clinical antidepressant selection for MDD.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Evidence Level</h3><p>1</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Technical Efficacy</h3><p>Stage 2</p></section>

Predictive modeling of antidepressant efficacy based on cognitive neuropsychological theory

Predicting Escitalopram Monotherapy Response in Depression: the Role of Anterior Cingulate Cortex

Machine learning to predict clinical remission in depressed patients after acute phase selective serotonin reuptake inhibitor treatment

Predicting Neuroimaging Biomarkers for Antidepressant Selection in Early Treatment of Depression

Connectome-based models can predict early symptom improvement in major depressive disorder

Prediction of Early Antidepressant Efficacy in Patients with Major Depressive Disorder Based on Multidimensional Features of Rs-Fmri and P11 Gene DNA Methylation

Predicting Effective Treatment Selections for Depressive Patients to Achieve Early Improvement According to Their Subtype Affiliations

Optimizing Antidepressant Efficacy: Generalizable Multimodal Neuroimaging Biomarkers for Prediction of Treatment Response

Prediction of Anxious Depression Using Multimodal Neuroimaging and Machine Learning

Development of a multivariate prediction model for antidepressant resistant depression using reward-related predictors

Distinct MRI-based functional and structural connectivity for antidepressant response prediction in major depressive disorder

Individual deviations from normative electroencephalographic connectivity predict antidepressant response

Ensemble Learning for Early‐Response Prediction of Antidepressant Treatment in Major Depressive Disorder

Predicting Treatment Selections for Individuals with Major Depressive Disorder According to Functional Connectivity Subgroups

Predictive utility of emotional regulation abilities for assessing cognitive improvement in depression

An Electroencephalographic Signature Predicts Antidepressant Response in Major Depression

Combining machine learning algorithms for prediction of antidepressant treatment response

Influence and interaction of genetic, cognitive, neuroendocrine and personalistic markers to antidepressant response in Chinese patients with major depression

Resting-State fMRI in Predicting Response to Treatment With SSRIs in First-Episode, Drug-Naive Patients With Major Depressive Disorder

Combination of Spontaneous Regional Brain Activity and HTR1A/1B DNA Methylation to Predict Early Responses to Antidepressant Treatments in MDD