Per-subject baseline monitoring
Within-subject reference distributions built from continuous physiological signal, replacing population thresholds as the basis for detecting change.
Capability statement
Operator state, measured against the operator.
Selected, trained and conditioned personnel are the population for whom a general reference range is least meaningful. Intuitive Pathos builds physiological monitoring that establishes each operator’s own baseline and reports departure from it — in both directions, weighted equally.
The problem
Reference ranges assume an average operator. There isn’t one.
Physiological thresholds in general use are derived from general populations. Operators in high-consequence roles are, by selection and by training, not drawn from that distribution — and the specific adaptations that make them effective are exactly the ones that make a population range misleading for them.
The result is an instrument that under-detects in the people it is most important to monitor, and over-flags in ways that erode trust in the system until it is ignored.
The harder problem
Degradation arrives quietly.
Fatigue, thermal strain, hypoxia, dehydration and accumulated load rarely announce themselves as a spike. They present as flattening — reduced variability, slower recovery, a narrowing of normal response — the critical-slowing-down signature set out on the evidence page. Conventional monitoring is tuned to the loud direction.
The operator is also, reliably, the last person to recognize it in themselves. Self-report is the least available signal exactly when it matters most.
Core competencies
What we do.
Early regime-change detection
Memory coefficient, critical slowing down, λ* criticality and sliding-window regime detection — established dynamical-systems early-warning methods applied to operator state. See the lineage.
Symmetric state reporting
Elevated and suppressed departures weighted identically, so degradation that presents as flattening is not systematically under-detected.
Sensor-agnostic integration
The approach depends on continuous signal, not on specific hardware. It is designed to work with sensing a program already fields.
Architecturally enforced data boundaries
What may leave a deployment and what may never leave it is expressed in the type system, not in configuration. There is no setting that enables the prohibited case.
Auditable fairness properties
Equity claims are inspectable at the interface level rather than asserted in documentation — relevant to emerging requirements on algorithmic accountability in federal systems.
Data posture
Sovereignty as a structural property.
For defense deployments, the question is never whether a vendor promises to keep data in place. It is whether the system is capable of moving it. We designed for the second question.
Per-subject baselines and identifiable physiological data do not cross the federation boundary. That constraint is enforced in the typed federation contract shared by client and server, which means it is not a deployment option, an environment variable, or a default a configuration error could reverse.
Deployment topology — including fully disconnected and air-gapped operation — is a program-level decision we expect to accommodate rather than negotiate.
Scope discipline
Pathos Guardian is a non-clinical system. It does not diagnose, does not generate a medical record, and does not make fitness-for-duty or employment determinations.
Those boundaries are architectural: the clinical decision logic is not present in the product’s dependency graph.
Trust with the monitored
A monitoring system that operators believe can be used against them will be defeated by the operators. What the system can emit about an individual is deliberately limited, and we regard that as a capability requirement rather than an ethical nicety.
Differentiators
Why us.
-
The comparison set is the operator
Not a cohort, not a percentile, not a demographic model. This is a different measurement, not a better-tuned version of the standard one.
-
Clinical origin, non-clinical product
The methods come from acute-care monitoring, developed by a practicing neuroscience nurse. The defense product is deliberately scoped to stay out of clinical territory.
-
Constraints that are verifiable
Our data and fairness guarantees are inspectable at the interface level. A technical evaluator can check them rather than take them on assurance.
-
Dual-use substrate
The same core is under development for clinical, veterinary and consumer settings, so the defense line benefits from evidence and hardening generated outside the defense budget.
Contracting
Program offices, primes and research partners.
We can provide a formatted capability statement, current registration details, and technical materials suited to a specific solicitation or teaming discussion.
Pathos Guardian is in development and is not currently available for operational deployment. Nothing on this page constitutes a representation of past performance.