Fine-tuned model from the PsyNamic project (code, webapp): Outcomes.
This is the model deployed in the PsyNamic webapp after the revision (Nov 2026). It supersedes STRIDE-lab/scibert-outcomes-20250221 (tag submission-v1, initial submission). All current models: PsyNamic collection.
michiyasunaga/BioLinkBERT-base
Outcomes (multi-label: several labels per abstract possible)
Lists the outcomes measured or reported in the study (e.g., fMRI, physiological functions).
| ID | Label | Definition |
|---|---|---|
| 0 | The ability to drive a car | – |
| 1 | Suicide | including suicidal ideation or related questionnaires. |
| 2 | Functional MRI | Also called fMRI. |
| 3 | MRI | E.g., arterial spin labelling. If fMRI is used, classify it under both MRI and fMRI. |
| 4 | PET, SPECT | Or similar imaging approaches. |
| 5 | Physiological Functions | E.g., heart rate, blood pressure, body temperature, pupillary diameter, BMI. |
| 6 | Electroencephalography (EEG) | Includes MEG. |
| 7 | Mental Functions | E.g., ability to recognize faces, form language, mindfulness, memory, measures of mood or depression, impulsivity. |
| 8 | Psychedelic experience | E.g., mystical experience questionnaire or visual scale for psychedelic experience. |
| 9 | Soluble biomarker | E.g., proteins in blood/CSF, substances in body compartments like hair, pharmacokinetic measures, or genotyping. |
| 10 | Surveys | E.g., questionnaires on recreational psychedelic use; excludes acquisition of patient data in the clinic, mostly refers to online surveys. |
| 11 | Interviews | E.g., person-to-person interviews. |
| 12 | Neurotoxicity | E.g., direct measures like MR-spectroscopy or patient cell assays; excludes indirect measures such as MDMA adverse events. |
| 13 | Emergency department visits | E.g., studies assessing reasons for emergency department visits of users of psychedelic substances. |
Multi-label: a sigmoid is applied per label, and every label with probability ≥ 0.29 is predicted. This is the threshold used in the PsyNamic webapp. It was tuned on the validation set with a pre-specified rule: start at the threshold with maximal weighted F1, then lower it as long as weighted precision stays within 0.05 of its value there.
Five base models (BiomedBERT-abstract, BioBERT, BioLinkBERT, ClinicalBERT, SciBERT) were fine-tuned with identical hyperparameters. The checkpoint is the one with the lowest validation loss (early stopping), and the base model was selected on the validation set (weighted F1). The test set was used once, after selection.
Held-out test set (94 abstracts), tuned threshold as in the webapp, 95% bootstrap CIs (1,000 BCa resamples).
| Metric (test) | Value [95% CI] |
|---|---|
| Weighted F1 | 0.827 [0.778–0.866] |
| Weighted precision | 0.827 [0.758–0.870] |
| Weighted recall | 0.847 [0.796–0.891] |
import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification
repo = "STRIDE-lab/biolinkbert-outcomes-20261107"
tokenizer = AutoTokenizer.from_pretrained(repo)
model = AutoModelForSequenceClassification.from_pretrained(repo)
inputs = tokenizer("<title>. <abstract>", truncation=True, max_length=512, return_tensors="pt")
logits = model(**inputs).logits
probs = torch.sigmoid(logits)[0]
pred = [model.config.id2label[i] for i, p in enumerate(probs) if p >= 0.29]
{
"data": "<private_dataset>",
"task": "Outcomes",
"model": "biolinkbert",
"cross_val": false,
"batch_size": 8,
"learning_rate": 5e-05,
"weight_decay": 0.01,
"lr_scheduler": "linear",
"warmup_ratio": 0.1,
"epochs": 30,
"dropout": 0.1,
"early_stopping_patience": 5,
"gradient_clipping": 1.0,
"max_length": 512,
"threshold": 0.5,
"is_multilabel": true
}
threshold (0.5) is only the cut-off used to compute the validation metrics during training. It does not affect the model weights or the checkpoint choice (lowest validation loss). Predictions in the webapp and the test performance above use the tuned threshold 0.29 (see Prediction rule).