Neural uncertainty flags difficult dopamine transporter scans
Source check: Source checked. Not independent scientific confirmation.
Finding: Residual-network ensembles classified dopamine-transporter SPECT and flagged uncertain decisions. A separate 645-scan Parkinson progression dataset showed 98.0% binary classification accuracy against clinical diagnosis; the uncertainty module flagged 92.3% of classification errors. Performance varied across camera datasets. This retrospective evaluation does not establish improved patient management or reliable exclusion of neurodegeneration in individual borderline scans.
Technology: Residual convolutional neural-network ensembles
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Separate high-sensitivity and high-specificity learned ensembles flag disagreement around an independently generated image classification.
- Test and result
- Development on 1,250 scans, separate 490-scan development test, 640-scan camera test, and 645-scan external clinical-reference test.
- Limitation / still unknown
- Prospective clinical utility remains unestablished; reference labels differ across cohorts, camera shifts affect performance, and uncertainty is binary rather than a calibrated probability.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10957699/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Primary relevant model Methods, Results and limitations; exact primary identity and current primary PubMed correction/retraction relationships checked..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.