A system and method estimate sleep-dependent intracranial fluid transport using a T1-registered or atlas-scaled probabilistic cerebrospinal-fluid-flow prior atlas calibrated by FAST-EIT or FAST-ECS measurements. The atlas represents high-resolution probabilities of cerebrospinal-fluid spaces, perivascular routes, parenchymal extracellular-space compartments, and outflow pathways, together with driver-specific response maps for cardiac pulsatility, respiration, vasomotor oscillations, sleep stage, slow oscillations, REM phasic activity, arousal, posture, and transcranial stimulation. The atlas is nonlinearly registered to a subject's T1 MRI or scaled to a subject's head model. EEG, impedance, and physiological recordings are acquired during sleep, and an electrical forward model predicts boundary impedance from atlas-derived extracellular-space and conductivity changes. Bayesian super-resolution fits the high-resolution fluid-transport prior to lower-native-resolution FAST-EIT/FAST-ECS data to estimate posterior maps of CSF flow, perivascular exchange, extracellular-space expansion, parenchymal transport, and glymphatic clearance capacity. Multi-night assessments, including four-night WISP or high-density EEG/EIT recordings, may be combined with dried blood or dried plasma spot biomarkers to classify clearance-reserve stage and validate amyloid, tau, glial, neurodegenerative, synucleinopathy, or mixed-pathology risk. Stimulation or sham challenge may be used to compute target engagement and model adequacy.