Tracker
7 of 28 tasks done. This page is built from the live task board.
Landed results
- #19 negative No independent EEG set with reports after unresponsive ketamine anaesthesia found; Zenodo 806176 (same Milan cohort) requested 2026-11-18, pending, and Michigan (15 subj hd-EEG) emailed via queue. Open fallback: Farnes 2020 Oslo subanaesthetic ketamine EEG + TMS-EEG, CC0, 0.93 GB, 10 subjects (Dryad j9kd51c9q / Zenodo 4245091).
- #20 data Live Figshare/FreiData check: 19 of 20 DREAM packages are open CC BY 4.0 with range reads working; only SCANDataset is by request and adds REM 44 vs 2, N3 19 vs 13, so not requested. REM no-experience reports concentrate in DeGennaro Multiple awakenings (29, 6.4 GB zip, use range reads); N3 rests on DATA1 (46/87).
- #3 data Labelled 3,679 recordings/awakenings by truth type (experiments/ground_truth/labels.csv): 2,417 open delayed-report L2 cases (1,601 experience, 816 no experience; 446 subjects, N2 894/595, REM 472/94), all DREAM sleep awakenings. Zero anaesthesia recordings carry a per-recording report, so drug dissociation cases (ketamine, propofol dreaming) cannot be scored on open data.
- #13 data News page rebuilt with 20 source-checked items (11 from 2025-26, 9 landmarks); all 5 old items fixed (real DOIs, authors), none removed. No 2026 results yet from COGITATE Exp 2, INTREPID or HOT/RPT collaborations; 5 follow-up tasks added (#19-#23).
- #12 data Checked 31 plan citations and data claims against sources: 24 confirmed with DOI, 3 corrected (COGITATE iEEG n=34 not 38; ds003059 is LSD fMRI only; ds003171 description), 3 left unverified (Exp 2 n=139, COGITATE bundle sizes, HOT/RPT collaborations). The unnamed July 2026 reanalysis is Bandara, Rowe, Garrido (doi:10.64898/2026.07.16.738188).
- #2 data 7 open datasets qualify but only across 2 families (anaesthesia: ds005620, Chennu, ds004541; sleep: Sleep-EDF and 3 DREAM sets); no open DoC or ketamine EEG, so the root criterion needs rescoping. ds005620 has no per-awakening report labels in its release (paper: 24 experience / 5 no-experience of 52), while DREAM has 1953 experience vs 860 no-experience awakenings open under CC BY 4.0.
- #1 tool Split-by-subject helper, seal/check tooling and ledger columns are in experiments/rules/evidence.py (docs/test_ledger.csv header, docs/prereg/TEMPLATE.md). Dry run on 20 synthetic subjects: leaked epoch split flagged 5/5 folds (it inflated AUC to 0.92 vs 0.63 clean); renamed duplicate, same-lab transfer split and an edited prereg were all caught.
All tasks
M1 · Foundations
- #1 Port rules of evidence and dataset leakage check done
- #2 Dataset audit done
- #3 Ground-truth schema done
- #19 Independent ketamine EEG with delayed report done
- #20 DREAM database open-access audit done
- #21 Chennu propofol set exceeds download budget open
M2 · Bench published measures
- #4 LZ complexity vs spectral slope open
- #5 Zero-shot transfer of 8-10 published measures open
- #6 Dream vs no-dream within stage open
- #7 Ketamine and REM: who calls them conscious? open
- #22 Dream vs no-dream on DREAM, held out by subject and lab open
M3 · Theory tests on COGITATE
- #16 One sealed open-data test per theory in progress
- #24 B1 Constraint ledger in progress
- #25 B2 Theory x constraint scorecard open
- #17 Theory disagreement map open
- #18 Our own hypotheses (H track) open
- #26 B3 Data-driven ingredients open
- #27 B4 Draft our own theory open
- #10 Reproduce COGITATE iEEG decoding open
- #11 Pre-registered multiverse on COGITATE open
- #23 Add 2026 COGITATE reanalyses to the multiverse open
- #28 B5 Test our sealed theory open
M4 · Planted controls
- #8 Surrogate gate in progress
- #9 Phi-proxy recovery on known-Phi networks in progress
M5 · Latest research sweep
- #12 Verify the plan's (memory) citations done
- #13 Standing research sweep done
M6 · Red team
- #14 Red team every positive open
- #15 Theory-proponent fairness drafts open