In finite stochastic population models, a small isolated group always eventually goes extinct from demographic noise alone —
random die-offs eventually wipe out at least one species with no way to recover.
The paper's finding: simply coupling that
vulnerable patch to a much larger stable patch — one that sustains a traveling activity front — supplies a steady trickle of
immigrants across the border. No extra fitness advantage, no rule changes: that flux alone suppresses the finite-size extinction
instability that would otherwise doom it. Source:
arXiv:2511.10807 — "Invading activity fronts stabilize
excitable systems against stochastic extinction" (Distefano, Shabani & Täuber, 2025).
Isolated corner walled off
A small corner patch, always cut off from the refuge — no reinforcements can cross the border.
warming up…
Refuge-connected corner connected
The same size patch, same noise — but the border is open, so the front can invade and reseed it.
warming up…
Empirical extinction probability — same noise, same patch size, only connectivity differs
Isolated patch—
0 trials so far
Refuge-connected patch—
0 trials so far
Species A
Species B
Species C
cyclic dominance: A beats B → B beats C → C beats A
dashed line = vulnerable/refuge border · a trial ends when a species is absent from the patch long enough to be unrecoverable, or when it survives the trial window