A Low-Burden Behavioral Measurement Architecture

Human evidence is most useful when its collection respects the people who provide it. A mission can pair a limited set of repeated measures with operational events, food-system activity, sleep, and environmental conditions, without recasting a demonstration as continuous survey work. This report sets out a way to explain change while safeguarding crew time and the appropriate handling of personal data.

For Gastronaut, the architecture describes an ORCA data interface for a human-research setting. The platform can record system events and food exposures. NASA and its research partners can govern behavioral and clinical measures. A synchronized timeline could connect those records while retaining the line between system telemetry and private human data. The practical question is how to produce explanatory evidence without asking more of a crew than the study requires.

The design problem

The neurocognitive synthesis contains 553 analysable studies from 5,382 unique records, including 95 human-spaceflight and 298 human ground-analog studies. The review retrieved no full texts. Its counts therefore describe abstract reporting.

Repeated assessment has been used during exposure. Thirty-five of 95 human-flight studies and 146 of 298 human-analog studies sampled during exposure. The same evidence identifies a design concern. Among 104 human-tier records carrying a cognitive outcome, 12 abstracts named practice effects, parallel forms, or a separately learned baseline. Full texts may address those matters, but this review cannot confirm that.

Dev and colleagues provide a useful illustration of the needed separation. Twenty-five astronauts completed NASA’s Cognition Battery across preflight, early-flight, late-flight, early-postflight, and late-postflight phases. Scores were corrected for practice and stimulus-set difficulty. At the group level, performance was stable across much of the mission. Selected early-flight domains slowed, and cognitive scores showed no clear association with sleep or alertness ratings (Dev et al. 2024). The finding supports examining domain, timepoint, and individual variation rather than drawing a conclusion from one group summary.

The corpus does not establish one common mission-phase pattern. Studies from Antarctic stations, ground isolation, Mir, the ISS, and other settings reported mid-mission troughs, end-of-mission changes, winter effects, and null results. Latitude, altitude, season, light, confinement, and mission design can shift together. A “third quarter” label is therefore not a substitute for measured conditions. The architecture should carry mission context from the outset, rather than add it as an explanation after analysis.

Confidence is high that low-burden repeated assessment can be conducted during missions and analogs. Confidence is moderate that operational and environmental context can explain part of the variation. Confidence is low in a universal time-course model across settings.

Four linked records

The architecture uses four records on a shared timeline, each with a distinct purpose and access boundary. That division serves analysis and also protects the trust needed for collection.

ORCA state, alarms, automation changes, interventions, maintenance, faults, and recovery belong in the system record. The record can identify equipment and task without including clinical information.

Harvest, preparation, offering, consumption, waste, composition, caffeine, and other relevant intake make up the food record. Participant linkage should use research identifiers under the approved protocol.

Brief measures of workload, sleep, fatigue, mood, acceptability, perceived restoration, and cognition form the human record. The instrument set should be prespecified and brief enough to support adherence. Clinical and behavioral data remain within the human-research governance structure.

Mission phase, schedule, light, noise, temperature, humidity, carbon dioxide, exercise, medication, and other declared confounders form the context record. This record allows investigators to consider whether an apparent food-system effect accompanied another change.

A governed linkage layer can join the records for analysis. Access controls can give engineers task and system data, food researchers exposure data, and authorized investigators coded human outcomes. Personnel-management use should remain outside the research purpose unless a separate authority and consent process establishes otherwise.

Keep the protocol light

A lunar demonstration can combine scheduled measures with measures prompted by events. A short baseline period establishes individual variation. Brief daily measures can record sleep, fatigue, mood, and workload. Cognitive testing can occur at selected mission phases with parallel forms or a validated practice correction. Prompts can follow a harvest, system fault, recovery, or extended manual task. The protocol then places measurement where it can illuminate an operating event.

Measurement frequency should follow the causal question. A study of restoration after plant interaction needs a comparator interval and a time window. Workload measurement should follow a defined task. A food-and-cognition question requires linked intake and testing records, while temporal proximity does not establish cause.

Missingness can also be retained as information. A crew member may omit a survey because mission work takes precedence. The protocol can record that circumstance, limit reminders, and avoid treating nonresponse as a performance deficiency.

Validate in stages

Gastronaut can start with the system and food records in ground operations. ORCA remains at approximately TRL 3 to 4 with no flight or lunar cycles. Its 1,042-cycle internal record may show which fields are available and which require prospective collection. It does not establish human outcomes.

A ground crew study can test timing, burden, privacy, and data linkage. A controlled analog can introduce behavioral measures under an approved research protocol. At each stage, NASA can decide whether the architecture is useful, whether the burden is acceptable, and whether flight collection is warranted. Those decisions prevent an unresolved measurement question from moving forward by assumption.

The lasting asset is a traceable account of the system, the food, the environment, and the person, with uncertainty and access boundaries preserved. It gives a later reviewer a basis for learning while leaving crew time and trust intact.

References

Dev, Sheena I., et al. “Cognitive Performance in ISS Astronauts on 6-Month Low Earth Orbit Missions.” Frontiers in Physiology, 2024. https://doi.org/10.3389/fphys.2024.1451269.

Gastronaut. Neurocognitive and Behavioral Health: Evidence Synthesis. Report GAS-B4-NEU-20260822, evidence version 22 Aug. 2026. Research synthesis.

Gastronaut. ORCA Public-Safe System and Evidence Baseline. Evidence version 23 Aug. 2026. Company technical record.