Studies that could test the theory.

Is there a digital ventilation mismatch?

Use a within-person comparison of high-arousal digital engagement, low-arousal digital engagement, high-arousal non-digital engagement, and a low-arousal non-digital control. Match posture, timing, light, duration, recent activity, speech, and prior sleep as closely as the design allows.

Measure minute ventilation, oxygen consumption, carbon dioxide production, and end-tidal carbon dioxide together. Verify that the arousal manipulation worked using measures independent of the respiratory outcome. The primary question is whether ventilation exceeds metabolic requirement, not simply whether breathing becomes faster.

If high-arousal non-digital engagement reproduces the signature, the mechanism may be sedentary arousal more broadly, with digital engagement as an important modern exposure. If the ventilatory rise is fully explained by metabolism, the proposed mismatch loses its starting point.

Does a repeated daytime effect reach sleep?

Repeated exposure and recovery measurements should separate altered chemoreflex responsiveness from biochemical adaptation. Measure the time course rather than testing one morning blood sample and assuming it represents the entire pathway.

The two-tier prediction is that neural changes may appear at a dose where morning bicarbonate has not measurably shifted, while sustained habitual overbreathing may additionally recruit acid-base adaptation. Overnight carbon dioxide, loop gain or reserve estimates, arousal burden, and sleep continuity then test whether either change matters at night. Ren & Robbins Krapf et al.

Does baseline stability moderate the evening effect?

Apply matched high- and low-arousal pre-bed challenges across a continuous range of baseline respiratory stability. Pre-specify an interaction between the exposure and measured baseline respiratory variables. Keep light and sleep opportunity matched so the respiratory prediction receives its own test.

Repeat the challenge within the same people after an intervention that measurably changes respiratory stability. The strongest version documents an ongoing acute arousal response while the sleep consequence diminishes. This distinguishes disappearance of the stimulus response from an increased ability to absorb its respiratory consequence.

Does stabilization improve restorative throughput?

At adequate sleep opportunity, measure stable non-REM continuity, high-frequency cardiopulmonary coupling, slow-wave activity, REM access, and next-day performance. The primary endpoint is the quality and functional yield of sleep, not whether participants can be made to sleep fewer hours.

Changes in sleep duration can be recorded, but they should not substitute for functional outcomes or be interpreted as beneficial by default. A person may initially sleep longer as continuity improves; the throughput hypothesis should be tested on the outcomes it actually specifies. Thomas et al. Tasali et al.

Test the sleep-mediated emotional sequence

Follow the sequence from respiratory change to non-REM continuity and REM access, then to emotional outcomes. Baseline anxiety, psychological content, circadian timing, medication, and other relevant variables need explicit treatment in the causal design.

The revised core does not require an emotional effect that remains after all sleep mediation has been removed. It predicts an emotional consequence carried through sleep. Direct amygdala mechanisms belong to a separate experimental question.

What would change the model?

Verified arousal with no ventilation–metabolism mismatch would challenge the upstream proposal. Repeated acute mismatch with no persistent respiratory or neural change and no nocturnal consequence would challenge the carryover claim. No exposure-by-stability interaction across an adequately measured range would weaken the latest moderation prediction.

Failure of a downstream extension would not automatically erase every component finding. The point of the staged program is to locate the failing link. Outcomes and analysis plans should be specified before examining results, so “arousal load” cannot be redefined afterward to rescue the hypothesis. Read the seven formal predictions in the manuscript.