01One framework, three questions
What demands arise? How does the system regulate? What happens to function over time? TAP brings these questions together as an educational synthesis of existing research, with explicit assumptions and candidate questions for future study.
This is a conceptual framework developed by The Allostasis Project. It is not a clinical diagnosis, validated biomarker or empirically validated predictive model. Its originality lies in this presentation and synthesis; the underlying scientific concepts remain attributed to their sources.
Predictive regulation is one influential account of allostasis, not a universally settled definition of the construct. TAP uses it as a theoretical starting point.[1][7]
02Read the relationships
The map places prediction and adjustment inside a feedback loop. Current state and reserve may influence how a challenge is met; behavior may also change the environment. Recovery and accumulating costs can coexist over different timescales. Arrows organize this synthesis rather than certify causal effects.
CURRENT STATE & RESERVE
Domain-specific capacity may influence the response.
- DEMANDS & CONTEXT
What is required?
Timing · duration · predictability - PREDICTION
What is expected?
Prior experience · current cues - PHYSIOLOGY & BEHAVIOR
How does the system adjust?
Coordination · action - FUNCTION OVER TIME
What happens to function?
Maintain · adapt · recover
Feedback informs subsequent adjustment. Behavior can change exposure.
ACROSS REPEATED CHALLENGES
Allostatic load concerns cumulative physiological costs. Reserve is an influence, not a resilience score.
Recovery and resilience are interpreted from a specified system, challenge and outcome. Conceptual connections—not inevitable stages or validated causal effects.
031. Demands and context
Describe the demand’s type, timing, duration and predictability, together with the surrounding conditions. A planned task and an unexpected task can be comparable in workload while differing in advance information. This is a proposed comparison, not a claim that equal task labels imply equal exposure.
The focus on anticipated requirements comes from predictive-regulation theory; specifying the system and challenge is also central to the NIH resilience framework.[1][4]
042. Prediction, physiology and behavior
Sterling’s account motivates including anticipation, coordinated adjustment and learning.[1] TAP represents these together with actions that may change exposure. The environment is not merely a background variable: newer allostatic proposals explicitly include reciprocal organism–environment relationships.[5]
Current state and domain-specific reserve are possible contributors to this process. A measured capacity in one task is not a whole-body battery or a guarantee of resilience under another challenge.[6]
053. Function over time
A demand, a regulatory response and a functional outcome are separate observations. A larger physiological response can support performance; a smaller signal does not automatically mean better function. The direction of a useful change must be defined for the outcome and setting.
The NIH framework and physical-resilience literature motivate following maintenance, decline and recovery of a specified system.[4][6] TAP therefore asks for a reference condition and observation window rather than treating return to an arbitrary flat line as the only successful result.
06The next challenge meets a changed system
Allostatic-load theory motivates attention to cumulative costs, while Reactive Scope considers changing response ranges.[2][3] Neither account makes every demand harmful, nor establishes that incomplete recovery in one variable equals measured allostatic load.
TAP’s synthesis keeps longer-term capacity and costs distinct from a single short response. Training, illness, resources and repeated demands may lead to different changes. Any claimed direction, timescale or causal relation needs evidence in the particular setting.
07Explore a short task
Imagine a brief cycling task that starts at time 30 and stops at time 60. In this mathematical illustration, demand, regulatory activation and task function use separate, dimensionless scales. Accurate advance notice, an unexpected start, a cancelled task and persistent activation expose different assumptions.
EDUCATIONAL SCENARIO · v1.0
A task begins at 30 and ends at 60. The response follows the demand; their initial mismatch affects the constructed function outcome.
The cue starts preparation at 20. A smaller shortfall at task onset is exchanged for a pre-task excess penalty. This trade-off is an assumption of the equation.
The cue starts preparation at 20, but the task never happens. The response starts to subside at 30. Anticipation incurs a penalty without a task benefit in this example.
The task is unchanged. Only the response’s decay time is longer. This does not measure tissue repair, allostatic load or an individual’s resilience.
Shared time axis · arbitrary units. Dashed markers show the scheduled start and end; no task occurs in “False alarm”.
Constructed curves, not measured physiology. Demand, response and function use separate scales; compare their timing and shape, not their absolute magnitudes. Read the assumptions ↓
08What the illustration assumes
Demand is a fixed pulse of 0.65 during the task. Regulatory activation moves exponentially toward the required level. An advance cue starts preparation at time 20; in the cancelled-task scenario no task occurs and the cue is withdrawn at time 30. These times and values are arbitrary.
Task function is defined as 1 − 0.7 × max(0, demand − response) − 0.22 × max(0, response − demand). This invented relationship penalizes both shortfall and excess. It is not an estimated physiological law. Changing it could change the apparent benefit of anticipation.
No fatigue, tissue repair, learning, reserve depletion or allostatic-load index is estimated. Persistent activation changes only its settling time. The example explains distinctions; it cannot establish which response is healthiest or predict a person’s recovery.
09What a study would need to observe
Specify the population, system, challenge and outcome before collecting data. Record a suitable reference period, the timing and size of exposure, repeated measures of the chosen response and function, and contextual factors such as sleep, medication and task familiarity. Different outcomes may require different schedules.[4][6]
For a short-task study, externally recorded workload, a physiological signal and an independently measured performance outcome should remain distinct. Using the same signal as both predictor and definition of success risks a circular result. Reserve should be represented by a justified capacity measure, not inferred from the shape of the same curve.
10Questions that could test the framework
Candidate question 1: with task demands held comparable, does accurate advance information alter the relation between regulatory response and function? A false-cue condition could test whether preparation carries a cost when a predicted demand does not occur. No such experiment is reported here.
Candidate question 2: does a prior response trajectory add out-of-sample predictive information about performance at a subsequent challenge, beyond workload, baseline function and measured capacity? The proposal is weakened if it adds no reliable information beyond those simpler predictors.
These are TAP research proposals, not findings. A study would need an appropriate protocol, prespecified analyses, confounder handling, uncertainty estimates and ethical review before recruiting participants. Observational prediction alone would not identify a causal recovery mechanism.
11Sources, synthesis and limits
The source theories support particular definitions and arguments; they do not jointly validate this whole diagram. TAP contributes the three-part organization, the explicit separation of demand/response/function and the illustrative implementation. Those are the objects being documented and versioned.
The map is TAP-FIG-0001 v2.1. This framework is TAP-MDL-0001 v0.3. Scientific review and methodological review are tracked separately in the editorial record. Neither an identifier nor a citation to NIH implies endorsement or completed validation.
12References
- Allostasis: a model of predictive regulation
Peter Sterling
Physiology & Behavior · 2012
Study record & citations - Stress and the individual. Mechanisms leading to disease
Bruce S. McEwen · Eliot Stellar
Archives of Internal Medicine · 1993
Study record & citations - The Reactive Scope Model — A new model integrating homeostasis, allostasis, and stress
L. Michael Romero · Molly J. Dickens · Nicole E. Cyr
Hormones and Behavior · 2009
Study record & citations - Conceptualizing a resilience research framework at The National Institutes of Health
LaVerne Brown · Barbara Cohen · Rebecca Costello · Olga Brazhnik · Zorina Galis
Stress and Health · 2023
Study record & citations - Resilience phenotypes derived from an active inference account of allostasis
Laura A. Harrison · Antonio J. Gracias · Karl J. Friston · J. Galen Buckwalter
Frontiers in Behavioral Neuroscience · 2025
Study record & citations - Physical Resilience in Older Adults: Systematic Review and Development of an Emerging Construct
Heather E. Whitson · Wei Duan-Porter · Kenneth E. Schmader · Miriam C. Morey · Harvey J. Cohen · Cathleen S. Colón-Emeric
The Journals of Gerontology: Series A · 2016
Study record & citations - Clarifying the roles of homeostasis and allostasis in physiological regulation
Ramsay DS · Woods SC
Psychological review · 2014
Study record & citations
Follow the connections.
Reproduce this edition
The downloadable TypeScript source preserves all four scenarios and their equations for this edition.
Download model specification ↗Editorial & review record
Record for v0.3 · Updated
Editing & copyediting
No completion recordEvidence verification
No completion recordScientific review
Not yet scientifically reviewedMethods review
No completion recordAI assistance
AI-assistedTAP Research Agent
Conceptual synthesis and source-linked educational preparation; not completed human scientific or methodological review.
OpenAI Codex (GPT-6) — Research assistance · Drafting and bilingual editing · Source-to-claim checks — Human checks have not been recorded.
v0.3
Editorial and interest disclosures
Founded and supported by Resilio. Authorship, scientific review, interests and updates are disclosed here. References do not establish product effectiveness.
Read the editorial standardsRevision history
A content update does not mean a new scientific review has been completed.
- v0.2
Clarified reserve as a possible contributor and specified the primary figure edition; model equations are unchanged.
- v0.3
Added explicit model status, primary theoretical influences and the distinction between predictive regulation and other accounts of allostasis; illustrative equations are unchanged.