What happened
The study recorded one night of polysomnography in 116 adults: 30 cognitively healthy without amyloid, 45 healthy with amyloid (preclinical) and 41 with mild impairment or mild Alzheimer's dementia. Slow-wave power was already reduced in the preclinical group, and in early Alzheimer's it was 15–21% lower than in healthy people without amyloid.
The shape of the slow wave was the clearest change along the continuum.
Why it matters
If a change in deep sleep shows up before symptoms, it could become an early signal — cheaper and less invasive than a brain scan or a spinal tap. But one night of sleep in 116 people is a long way from a diagnostic test.
What comes next
What is needed is a larger study with repeated nights, to see whether the slow-wave shape predicts who goes on to develop symptoms. This is an early signal, not a diagnostic test.
Small study, one night of sleep, a biomarker not a diagnosis. Different from Monday's story 9 (slow waves and tau in Berkeley) — different team, different cohort and different marker (wave shape, not tau). If published, make clear it is an early signal, not a diagnostic test.