EXPLAIN · MEDIATOR
09Which measured variables are plausible temporal intermediates?
Sleep or EEG physiology, autonomic measures, pain, immune or mucosal signals, endocrine state, energy availability, intake, bowel state, medication pharmacology, and interoception are competing or interacting candidates. A one-time abnormal biomarker or parallel change is not mediation.
POPULATION / COMPARATOR
Who or what should be compared?
Repeated within-person windows with time-ordered candidate intermediates, outcomes, and competing variables; confirmation requires a separate set of windows or participants and, where justified, a matched mechanistic model.
STUDY DESIGN
How could the question be tested?
Choose measurements from the event phenotype; establish temporal order; sample baseline, exposure, functional change, and recovery; model competing mediators and confounders; and connect human signals to allele-aware cellular tests where biologically justified.
EVIDENCE GATE
What would support the proposed link?
the candidate precedes the functional outcome at the study's temporal resolution, explains incremental within-person variance, reproduces, shows a coherent dose or time relation, and remains associated after competing variables are modeled. Concurrent change remains a correlate or biomarker candidate until higher-resolution ordering or an appropriate intervention, correction, or rescue test.
STOP RULE
What would weaken it?
the measurement is cross-sectional or one-off, follows the outcome, lacks a baseline or recovery value, loses association after confounder control, or is being called a mediator because it is abnormal.