Skip to content

Traumatic brain injury pilot: scoping

Status: built (2026-10-01); see Outcome at the end. This tests the injury granularity rule in design decisions §3f (design-decisions.md) on one case before it is applied more widely.

Why traumatic brain injury

Two sequela entries already exist and already model parts of the acute injury cascade inside themselves:

  • Post-Traumatic_Epilepsy has a root Traumatic Brain Injury node (TISSUE scale) and downstream nodes for blood-brain barrier disruption with albumin extravasation, neuroinflammation and inflammasome activation, reactive astrogliosis with impaired glutamate homeostasis, and GABAergic interneuron loss. It conforms to epilepsy_excitation_inhibition_imbalance for the epileptogenic part only.
  • Chronic_Traumatic_Encephalopathy has a root Repetitive Head Trauma node, then perivascular hyperphosphorylated tau and TDP-43 pathology, and conforms to tdp43_proteinopathy.

If a traumatic brain injury entry and the modules below are added, both existing entries should be able to conform to those modules at their injury-cascade nodes. Whether that works without distorting either entry is what the pilot tests.

Ontology coverage (checked 2026-10-01)

Need Finding
disease_term MONDO:0858950 traumatic brain injury. It sits under MONDO:0021178 injury and also under MONDO:0005560 brain disorder, so it is reachable from MONDO:0000001 disease and validates as a DiseaseTerm without a schema change.
Parent and sibling terms MONDO:0043510 brain injury (also under disease); MONDO:0800482 head injury (under injury only, so not admissible as a disease_term today). No MONDO term for concussion or mild traumatic brain injury was found (concussion search returned only post-traumatic epilepsy).
Related entries already in the KB MONDO:0043264 post-traumatic epilepsy; MONDO:0043512 traumatic encephalopathy (bound by the CTE entry).
Human exposure term (ECTO) None. l~injur, l~trauma, l~concuss, l~impact, l~crush, l~collision, l~acceleration, l~force all returned nothing. ECTO:9001937 exposure to explosive exists but is a chemical-exposure class, not blast overpressure.
Experimental condition (XCO) XCO:0000968 experimental traumatic brain injury, XCO:0000969 lateral fluid percussion injury. Usable for animal-model context only.
Existing modules to reuse glutamate_excitotoxicity (glutamatergic stimulation, calcium overload, mitochondrial dysfunction, excitotoxic neuronal death); neuroinflammation_glial_activation (danger signal release, glial activation, cytokine release, bystander injury); mitochondrial_dysfunction; tdp43_proteinopathy; epilepsy_excitation_inhibition_imbalance. No module covers blood-brain barrier breakdown in the CNS, diffuse axonal injury, or tau pathology.

Proposed shape

Entry: Traumatic_Brain_Injury

  • disease_term: MONDO:0858950.
  • environmental: the mechanical injury, unbound, with the ECTO searches above recorded in notes.
  • progression: primary injury (minutes), secondary injury (hours to weeks), chronic phase (months to years). These are phases, not subtypes.
  • No has_subtypes: split. Severity is a gradient of one exposure and is recorded in the environmental record and in progression:. Focal contusion and diffuse axonal injury have different tissue mechanisms but usually co-occur in the same patient, so they are two branches of the pathograph, not two strata of patients.
  • Root pathophysiology node: the mechanical event at TISSUE scale with UBERON locations, matching the shape Post-Traumatic_Epilepsy already uses. It has two downstream branches: focal contusion (vascular disruption, haemorrhage, local necrosis) and diffuse axonal injury.
  • computational_models: one finite-element head-impact model, if a published one can be cited, linked to the axonal-injury node with model_scale and a PROXY_QUANTITY divergence (tissue strain standing in for axonal damage). This is the only part of the pilot that tests the biomechanics half of §3f.

Modules: small and reusable

The cascade is split across several small modules rather than one trauma-specific module, so each can be reused outside trauma. Candidate chain, each step to be checked against the literature before a node is written:

Step Home
Primary mechanical tissue disruption The entry's own root node (trauma-specific, not a module)
Blood-brain barrier breakdown New module blood_brain_barrier_breakdown. Reusable for ischaemic stroke, sepsis-associated encephalopathy, and the existing Post-Traumatic_Epilepsy barrier node
Diffuse axonal injury New module focal_axonal_injury_wallerian_degeneration. Neither peripheral_axonal_degeneration nor corticospinal_tract_axonopathy could be widened: both model length-dependent dying-back of long axons, not acute injury at one point followed by degeneration of the distal axon
Ionic flux, glutamate release, calcium overload, excitotoxic death Existing glutamate_excitotoxicity
Mitochondrial failure Existing mitochondrial_dysfunction
Neuroinflammation Existing neuroinflammation_glial_activation

The create-module skill applies to the two new modules, including its guidance on stating complementarity with sibling modules in the module description.

Conformance changes to existing entries

  • Post-Traumatic_Epilepsy: its blood-brain barrier node conforms to blood_brain_barrier_breakdown and its neuroinflammation node to neuroinflammation_glial_activation; its epileptogenesis nodes keep their existing conformance.
  • Chronic_Traumatic_Encephalopathy: repetitive mild injury is mostly axonal, so a link from its trauma node to the axonal injury module is the candidate. Its tau node stays unconformed until a tau module exists.

Decisions (2026-10-01)

  1. The §3f rule is accepted as the working rule.
  2. MONDO:0021178 injury is added to the DiseaseTerm and DiseaseOrSubtypeTerm roots. This does not affect the pilot, since MONDO:0858950 was already admissible, but it makes burn, frostbite and fracture entries possible.
  3. No subtype split, by severity or by lesion type (see above).
  4. Small reusable modules, not one secondary-injury module.
  5. No ECTO term request. The exposure stays unbound with the searches in notes.

Build order

  1. blood_brain_barrier_breakdown and focal_axonal_injury_wallerian_degeneration modules.
  2. Traumatic_Brain_Injury entry conforming to them and to the existing modules.
  3. Conformance edits to Post-Traumatic_Epilepsy and Chronic_Traumatic_Encephalopathy.
  4. The finite-element model link, if a citable model is found.

Outcome (2026-10-01)

Built as planned, with three departures.

  • Post-Traumatic_Epilepsy's inflammation node conforms to the existing nlrp3_inflammasome_activation module rather than neuroinflammation_glial_activation: the node is about IL-1 beta and NLRP3, which is that module's central step. Its barrier node conforms to the barrier module's extravasation node, not its central node, because it is about albumin entering the neuropil and acting on astrocytes.
  • Chronic_Traumatic_Encephalopathy was not conformed to the axonal module. None of its nodes describes axonal injury; its mechanism is tau and TDP-43 pathology. Instead its trauma-to-tau edge gained the finite-element evidence that impact strain is greatest at the sulcal depths where CTE pathology sits. Conforming it properly needs an axonal node in the entry, or a tau module.
  • Brain swelling is not drawn from barrier disruption in Traumatic_Brain_Injury. The human diffusion MRI studies cited there find post-traumatic swelling predominantly cellular, and the barrier module scopes itself to vasogenic oedema.

The biomechanics part of §3f worked without a schema change beyond one enum value: the schema had no model type for a finite-element model, so BIOMECHANICAL was added to ComputationalModelTypeEnum. The head model links to the mechanical root node at equal scale and to diffuse axonal injury with a PROXY_QUANTITY divergence (strain standing in for axonal damage), which is the pattern §3f predicted.