An ecosystem with low species turnover looks stable. The same species persist year after year, the community composition barely changes. A brain with strong episodic memory at age 85 looks healthy — the person remembers as well as someone thirty years younger, no sign of decline. Both systems present an appearance of robustness. Both appearances can be misleading.
Nwankwo and Rossberg, analyzing community data spanning a century, found that species turnover in ecosystems worldwide has slowed by a third since the 1970s. The mechanism is not stability — it is depletion. Regional species pools have shrunk. When a species declines in a local community, there are fewer candidates available to replace it. The turnover rate drops not because the community is healthy but because it has fewer spare parts. The engine runs, but the parts bin is empty.
Disouky, Lazarov, and colleagues, profiling 356,000 individual cell nuclei from 38 human brains, found that SuperAgers — adults over 80 with the memory of 50-year-olds — produce two to two-and-a-half times more new hippocampal neurons than their peers. The strongest predictor of cognitive trajectory was not gene expression but chromatin accessibility: whether the DNA was physically open or closed to transcription. SuperAgers maintained open chromatin in neurogenic lineages. Alzheimer's brains showed progressive chromatin restriction — the regulatory architecture shutting down before the genes themselves went silent.
The structural parallel: in both cases, the surface metric is a poor discriminant. An ecosystem with unchanging species composition and a brain with preserved memory performance look the same from the outside regardless of whether the system beneath them is resilient or depleted. The diagnostic power lives in the reserves — the regional species pool that could respond to a disturbance, the chromatin landscape that could support new neuron production. A resilient system has high reserves and looks stable because nothing needs to change. A depleted system has low reserves and looks stable because nothing can change.
The difference becomes visible only under stress. A resilient ecosystem absorbs a species loss by drawing from the regional pool — the vacancy gets filled, the community reorganizes, function is maintained. A depleted ecosystem absorbs the same loss with silence. The vacancy persists. A resilient brain loses neurons and replaces them from an active neurogenic pipeline. A depleted brain loses neurons and cannot replace them because the chromatin landscape no longer permits the transcription factor programs that would generate replacements.
The measurement implications are unsettling in both domains. Conservation biologists monitoring unchanging communities may conclude the ecosystem is healthy; the species turnover analysis suggests the opposite interpretation. Clinicians seeing preserved memory in an older patient may assume the brain is aging normally; the chromatin data suggests that resilience and early decline can look identical at the surface. In both cases, the conventional metric (community composition, cognitive performance) is a trailing indicator — it tells you what already happened, not what will happen when the system is next challenged.
What connects these two systems is a general principle about redundancy and appearance. Redundancy is invisible in good times. A system with abundant reserves and a system with exhausted reserves produce the same output under normal conditions. The reserves matter only at the boundary — when the system is perturbed and must draw on capacity it either has or doesn't. Measuring the reserve directly (species pool richness, chromatin openness) predicts future resilience. Measuring the output (species composition, memory scores) describes current performance. The two metrics diverge exactly when it matters most, and the divergence is detectable before the crisis arrives — but only if you know to look.