Bryan Johnson drops sleep architecture data on co-sleeping vs solo beds:

Actigraphy on 55 couples shows ~6 partner-triggered wake events per night. 1 in 5 awakenings cascade from partner movement. You only sleep through half your partner's restless phases. Study didn't isolate solo sleep as control.

Self-reported survey (n=1000) correlates bed-sharing with less insomnia and fatigue, but it's cross-sectional garbage—no causality, just happier people choosing to share beds.

Small polysomnography study (n=12 couples) found 10% REM boost when co-sleeping, less fragmented REM cycles, stronger sleep-stage synchronization between partners. Trade-off: more limb movement noise.

Sex-specific split in n=10 study: women's actigraphy and subjective ratings both tanked with a partner present. Men reported better sleep subjectively, but objective metrics unclear.

Circadian sync matters more than proximity: 46 couples showed tighter sleep-wake alignment = lower nocturnal blood pressure (especially women) and reduced systemic inflammation, independent of actual bed-sharing frequency.

TL;DR: polysomnography hints at REM gains, but sample sizes are laughably small (10-12 couples). Most data is correlational noise. No universal answer—optimize for your own sleep fragmentation tolerance and REM architecture.