Predicting Couple Therapy Outcomes Using Deep Neural Networks on Pre-Treatment Assessments
Bibliographic record
Abstract
The objective of this study was to determine whether deep neural networks can accurately predict couple therapy outcomes using only pre-treatment psychological and relational assessments. This quantitative predictive study analyzed pre-treatment data from 176 Canadian couples seeking therapy for relational distress, communication difficulties, or emotional disconnection. Participants completed standardized intake measures including relationship satisfaction, communication patterns, emotional symptoms, attachment orientations, and dyadic demographic variables. All data were preprocessed, normalized, and transformed into dyadic-level and discrepancy-level features. A deep neural network was developed using TensorFlow/Keras, optimized via hyperparameter tuning, and evaluated against baseline machine-learning models. The dataset was split into training, validation, and test subsets using a couple-level 70/15/15 partition to preserve dyadic independence. The deep neural network demonstrated superior predictive accuracy compared to random forest, support vector regression, and linear regression models, achieving an R² of .71 on the test set. SHAP analyses revealed that relationship satisfaction, demand–withdraw patterns, attachment avoidance, partner stress discrepancy, and constructive communication were the strongest predictors of therapy outcomes. Inferential patterns indicated significant non-linear interactions between emotional symptoms and communication variables, with higher improvement predicted for couples displaying lower avoidance, greater baseline cohesion, and smaller dyadic emotional discrepancies. Predicted-versus-actual outcome comparisons showed strong convergence, with minimal dispersion around the diagonal line of fit. Deep neural networks offer a powerful method for predicting couple therapy outcomes using intake assessments, capturing the complex non-linear dynamics inherent in relational functioning. By identifying key pre-treatment predictors such as satisfaction, communication patterns, attachment profiles, and dyadic discrepancies, these models can support personalized treatment planning and enhance clinical decision-making. The findings underscore the promise of computational approaches in advancing precision-based psychological interventions for couples.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".