Early-warning signals, explained
In many complex systems, recovery from small disturbances slows before a larger shift. That slowing can show up as rising autocorrelation and variance. Here is what the idea can, and cannot, tell you about one person.
Critical slowing down
As a system approaches a tipping point, it takes longer to bounce back from small knocks. Researchers call this critical slowing down, and it has been studied in ecosystems, climate and other systems (Scheffer et al., 2009). Its traces are rising lag-1 autocorrelation and rising variance.
How it is measured
Standard methods detrend a series, compute the indicators in a rolling window, test their trend, and compare the result with surrogate data (Dakos et al., 2012). The State-Change Signal follows that procedure.
What group studies found
At the group level, researchers found signs of critical slowing down in mood measures before changes in depression (van de Leemput et al., 2014). ATS-1 tags this Model, with mixed evidence.
Why one person's signals are uncertain
A 2024 review found that, at the level of one person, the signals appeared before fewer than half of changes, and in about 25% of people whose state did not change (Helmich et al., 2024). That is why ATS-1 reports a graded statement, never an alert, and why every output carries the disclosure. The false-alarm calculator shows what it means for a group.
A research estimate
ATS-1 is for research and general wellness only. It does not diagnose, treat, prevent or predict any disease, including any mental health condition. The State-Change Signal is a research estimate of change in self-reported wellbeing patterns, not a prediction of any health condition.
Questions
Can early-warning signals tell me a change is coming?
Not reliably for one person. They are research estimates with substantial uncertainty.
Sources
Keep reading
Take the next step.
What early-warning signals are, why recovery slows near a shift, what group studies found, and why one person's signals are uncertain.