Neurospectrum beats all baselines classifying OCD from fMRI (0.70 accuracy)
Pseudomanifold · x · 2026-11-09
Neurospectrum thread (12/n): on resting-state fMRI from 27 unmedicated OCD patients and 31 healthy controls, scanned before/after a decision-making task, Neurospectrum outperformed all baselines in OCD-vs-control classification with subjects held out per fold — 0.70±0.02 fMRI accuracy vs. 0.65 for ICA and 0.63 for dynamic functional connectivity. It also led on Kuramoto synthetic data (MSE 0.0041) and mouse visual cortex stimulus decoding (0.60).
Related event: Neurospectrum Detects OCD Brain Trajectory Changes from fMRI(2 posts)→
More from Research
- Neural network output layers explained: classification vs regression — HarperSCarroll · 2026-11-09
- Ropedia W0 trains robot world models on 10,000 hours of human video, lifting LIBERO-Plus by 6.76 points — jiqizhixin · 2026-11-09
- Mathematician rebuts AI proof critics: demanding Lean verification is a double standard — AlexKontorovich · 2026-11-09
- RobPicker in Nature Methods: meta-learning fixes noisy labels, protein detection 0.24 to 0.47 — bravo_abad · 2026-11-09
- Google's ScientistTwo AI writes expert-level papers end-to-end and self-reviews — 2C_ornot2C · 2026-11-09
- GAP: experimental gated block merges attention and MLP, cutting ~20% params — WAMFT · 2026-11-09