{
"risk_flags": ["physio_spike_no_verbal_ack"],
"arousal_event": "41:12",
"verbal_ack": "none_in_window"
}Say/body gap — HR spike, words didn’t acknowledge it. Clinician can dismiss.
The client said “fine.” The body disagreed.
Heart-rate and vocal signals, aligned to the session timeline — FHIR-ready evidence for behavioral-health platforms.
Not a diagnostic. Clinician decides. Post-session only.
Published method Session Physiology definition and science foundation — public as of July 2026. Definition Science
SeroState turns wearable physiology into structured session documentation, journey tracking, and EHR-ready reports — post-session, clinician-reviewed, audit-ready.
Session topic · workplace stress
Session-dense heart rate and voice from the encounter (EDA when present) — the captured signals that become the record.
Physiology and words on one timeline. Say/body gaps with surrounding context — reviewable evidence, not automated alerts.
Chart destination · FHIR R4 → partner EHR
QOAX Clinical Session Report and payload for write-back. The clinician reviews and always decides.
Transcript and topic from the session — “No, I’m fine.” at 41:12.
Heart-rate trace with an evidence window on the same timeline as the words.
Reviewable fields and FHIR R4 into the chart. The clinician always decides.
SeroState is the Session Physiology Platform for behavioral-health platforms and EHRs: post-session heart-rate and vocal signals, time-aligned to the encounter, returned as FHIR R4 evidence via the QOAX API. Not a scribe. Not a wellness dashboard. Definition → · Company →
Deteriorating patients can go unseen until crisis when the chart holds words alone. Independent context: therapists caught about 1 in 8 deteriorating clients (Østergård et al., 2024; not a SeroState study).
AI scribes transcribe and summarize what was said. SeroState fuses the same session audio with wearable physiology — then flags where the body’s response didn’t match the words.
During the session, wearable physiology and vocal features are logged densely enough to align with conversation — not a day-long wellness average. Nothing changes about how the clinician works.
Post-session, SeroState time-aligns physiological and vocal signals against the session timeline.
Detects say/body gap events — physiological spikes with no verbal acknowledgment in the surrounding window.
Delivers a structured, FHIR-compatible payload: flags, evidence segments, confidence and signal-quality fields. SeroState is an intelligence layer, not a diagnostic device. The clinician reviews the trace and always decides.
Primary path — add Session Physiology to your stack.
One API integration adds a physiology evidence layer beside the note — FHIR R4 for the chart. Built for EHR and telehealth platforms, not a standalone clinician dashboard.
Secondary path — objective session measures.
Session-time-aligned physiological and vocal data via API alongside PHQ-9 and GAD-7. Protocols and ethics owned by the partner.
Program integrity — documentation-layer evidence.
Encounter-level physiology beside the note for Medicare/Medicaid program integrity evaluation — not fraud scoring, not a CMS endorsement claim.
More use cases
Session physiology measures the body’s signals during a behavioral-health session — heart rate first, vocal patterns when available — time-aligned to what was said. Self-report is what a client shares; physiology shows how their nervous system actually responded.
Where the two diverge, SeroState calls that the say/body gap: a reviewable flag with the trace and confidence behind it.
Illustrative shape only. confidence and signal_quality appear as computed fields in the post-session payload — no sample scores shown here.
Your organization holds the PHI. We don’t store it.
Architected for HIPAA-aligned deployments. BAAs for pilot partners.
Flags are reviewable evidence. Clinicians retain full judgment.
FAQ
We are a behavioral health evidence API company. SeroState fuses post-session wearable and vocal biomarkers into FHIR R4 session physiology evidence for EHR platforms. Full overview →
Blinded comparison against expert clinical judgment, reported by population subgroup. Prospective trial in design; we pre-register and publish before making accuracy claims. Scientific foundation →
SeroState is not an AI that talks to patients — we give clinicians physiological evidence to review. A 2026 Nature Medicine audit found concerning chatbot behavior was widespread across 810 simulated conversations — reduced in newer models, but still present. Full answer →
Primary path: platforms and EHRs evaluating Session Physiology. Research, payer/TPA, and trial roles welcome — we route the conversation accordingly.
Where do you fit?