Roadmap
Each mark on the tape below is one published output, placed by year. Teal marks are first-author work. The same device runs through the site, because it is the same idea the research is built on: a record written in order, read back at the end.
Four research systems, one question carried through all of them. Each step produced the problem the next had to solve.
2012 – 2015 · St Petersburg
Neuronal differentiation and the pathways controlling it: p53, GSK3β, ERK1/2 and the MAPK cascade. The finding that set everything after it — the same cascade that drives a cell to differentiate also restrains what it secretes.
assets/img/roadmap-01.jpg2015 – 2017 · St Petersburg
Hippocampal neural stem cells from seizure-prone rats proliferated less and differentiated faster than controls, pointing to aberrant neurogenesis as a cause rather than a consequence of epilepsy.
assets/img/roadmap-02.jpg2017 – 2018 · Oulu and Oslo
Wnt signalling in central nervous system development, on funding secured independently, adding CRISPR editing, imaging and iPSC-derived organoid culture. A short mission in Oslo added cerebral organoid and spinal cord production.
assets/video/roadmap-03.mp42018 – 2022 · Oslo
Four preclinical programmes on mechanism, developmental safety and distribution to target tissue. Drugs reach the developing brain within minutes at clinically relevant concentrations, and penetration depends on developmental stage. Twice, a single endpoint pointed the wrong way.
assets/video/roadmap-04.mp4Three aims in order, each with the criterion that decides whether the next one starts. These are fixed before the data are seen.
Three recording channels in four human iPSC lines. The test is whether the record returns a glucocorticoid schedule that was set in advance.
Gate: half of receptor-expressing cells record a 24-hour pulse · two pulses ten days apart ordered correctly in 80% · eightfold dynamic range · under 5% false positives · still writing at day 150 — full criteriaA time-indexed and a state-indexed model compared on out-of-sample prediction in the same cells, lineage and cell type held constant. Two exposures: one acute and receptor-mediated, one chronic and undatable.
Gate: pre-specified model comparison, dose anchored to real human internal exposure — detailIntervene inside the identified state and the exposure-associated phenotype should fall. Intervene outside it and nothing should change.
Gate: the asymmetry holds, and differentiation, cell proportions and survival are preserved — detail