For handlers with epilepsy and other seizure disorders. The certified discipline is seizure response: trained assistance during and after a seizure. Pre-seizure alerting is recorded as documented alert behaviour when an individual dog demonstrably shows it (§9.4) — a meaningful minority of seizure-response dogs develop it — and an emerging scent-based alert-training protocol is provided in §9.7 for teams who wish to pursue it with honest expectations.
9.1 Certifiable tasks (at least 3 required; ⚠ chains 100% in assessment)
| Task | Trigger | Standard |
|---|---|---|
| Remain-and-guard | Event (trained on staged episodes) | Dog stays with the seizing handler in a safe passive position (beside, not on top of), does not lick the face or interfere with the airway, remains until the episode ends |
| Summon help ⚠ | Event or cue | Finds designated household member, attention behaviour, leads/returns |
| Alert-device activation ⚠ | Event or cue | Presses adapted seizure-alert button / pulls cord |
| Retrieve medication/phone | Cue (post-ictal) or chained | Labeled rescue-med pouch or phone to hand/floor beside handler |
| Post-ictal DPT / grounding | Cue | Deep pressure once the handler is recovering and invites it; calm release |
| Post-ictal block & guide | Cue | Passive blocking during disoriented wandering; guides to a seat on a short known route |
| Fetch a blanket / designated item | Cue | For post-ictal care routines |
Deliberately not certified: any task requiring the dog to physically reposition, restrain, or "cushion" a seizing person — the safety evidence does not support it, and it risks both dog and handler.
9.2 Candidate suitability
- Exceptionally sound-and-motion-stable dogs: a convulsive seizure is intense stimulus. Candidates are screened with recorded-seizure video/audio exposure and staged-motion simulations before task training begins.
- Initiative and environmental confidence: several chains must run with no handler direction.
- Dogs showing fear or arousal escalation around simulated episodes are redirected to a different discipline — this is a welfare gate, not a failure of the dog.
9.3 Training methodology requirements
- Event recognition is built on scripted staged episodes (a trained adult simulates the motor pattern per script). Scripts are standardized so the dog keys on the event class, not one person's acting.
- Chains proofed with the "handler" non-verbal throughout (a seizing person cannot cue).
- Third-party and device chains trained to completion criterion with confirmation behaviour (dog returns to handler after summoning/pressing).
- Gradual intensity progression protects the dog; sessions end on success; stress signals end the session.
9.4 Documented alert behaviour ("with documented alert")
Identical evidentiary bar to §7.3:
- ≥ 10 corroborated pre-seizure alerts over ≥ 8 weeks, logged (date/time, alert behaviour, the subsequent witnessed/recorded seizure, interval).
- Corroboration: witnessed episode, seizure-diary entry by another person, or wearable/EEG data where available.
- When on file, the registry and letters record "Seizure Response Service Dog with documented alert behaviour." Without it, no alert claim is made anywhere.
9.5 Task test
- Each certified task: 10 distributed trials ≥ 9/10 across 2+ environment classes; ⚠ chains 5/5.
- Staged full-episode drill ⚠: 3 scripted seizures (per §9.3 scripts, non-verbal "handler"): the dog executes its complete certified sequence — e.g., remain-and-guard → summon/device → post-ictal tasks — 3/3. At least one drill includes a second person present as a distraction; at least one is conducted away from home.
- Remain-and-guard is additionally scored for welfare: the dog must stay task-focused without distress escalation (whining/panic invalidates the trial).
9.6 Maintenance
Weekly chain refreshers; monthly staged drill (logged); alert log maintained if the §9.4 designation is held (reviewed at re-assessment; the designation lapses if the behaviour stops).
9.7 Appendix — emerging protocol: scent-based seizure alert (in-training designation only)
Published research indicates seizures produce a characteristic odour profile that trained dogs can discriminate. Service Paws supports teams pursuing scent-based alert with honest expectations: this is an emerging methodology, and Service Paws does not certify a scent-alert claim in v1.0 — a dog in this program may carry the file designation "scent-alert program: in training" only.
Protocol requirements to hold the in-training designation:
- Sample collection: sweat swabs (forehead/neck/palms) taken by a caregiver during or immediately after a seizure, sealed and frozen per the Chapter 6 handling standard; matched baseline samples from calm periods as controls.
- Imprint & discriminate exactly per the Chapter 6 methodology (blanks, distractors, handler-blind lineups).
- Progress bar: the dog may be presented for a Chapter 6-style blind discrimination test (15 presentations, 7 hot) — a pass is recorded in the file as "scent discrimination demonstrated (laboratory conditions)".
- No live-alert claim is recorded unless/until the dog independently meets the §9.4 corroborated live log — at which point it is the §9.4 designation that appears, with the scent program noted as its training basis.
This appendix will be revisited in future versions as evidence and field results accumulate.
9.8 Records for this chapter
Task sheets; staged-drill scoresheet + video when overseen remotely; §9.4 log if designation claimed; §9.7 program records if applicable. On pass, Service Paws records the task testing date.