SSTIM is an open vocabulary for describing sensory stimulation: the deliberate, parameter-specified delivery of auditory, visual, tactile and cross-modal stimuli, and the records and evidence that accompany it. It is published as an OWL 2 ontology with a multilingual SKOS vocabulary, SHACL shapes and four conformance profiles, at the persistent namespace https://w3id.org/sstim.

This report describes what the vocabulary covers, how its modules and profiles are organised, what an implementer must do to conform to a profile, and how its identifiers are versioned, cited and mapped to external vocabularies. It is a description of a published artifact rather than a standard: the normative content is the RDF itself, which this document points at rather than reproduces.

This is a working draft of a Community Group Report. It is published to invite review of the vocabulary's scope, structure and external mappings by people who describe stimulation protocols and by people who build the software that delivers them.

The vocabulary it describes is versioned and archived independently of this document. Cite a release, never this draft, when an implementation depends on exact term semantics.

Introduction

Sensory stimulation is described today in prose, per study and per product. Two descriptions of the same 40 Hz auditory protocol, written by two groups, are usually not comparable, and neither is machine-readable. What varies is not only terminology but which parameters are considered worth stating: a carrier frequency without a modulation depth, a session length without a delivery route, an outcome without the evidence tier of the claim behind it.

SSTIM exists so that such a description can be written once, exchanged, and read back without loss. It names the techniques, the stimuli and their parameters, the media and body placements through which they are delivered, the exposure and safety boundaries that constrain them, the session record that results, and the evidence assessments attached to claims about them.

SSTIM describes stimulation. It does not assert that any stimulation produces any effect. Claims of effect are modelled explicitly, as evidence assessments carrying a tier, a direction and a source, so that a consumer can see what is asserted and on whose authority.

Scope

In scope:

• stimulation techniques, from binaural beats and photic driving to clinical neuromodulation methods such as transcranial direct current stimulation and vagus nerve stimulation;
• stimuli, their parameters, and the signal shapes and patterns that define them;
• delivery media, device capabilities and body placements;
• exposure profiles, comfort and safety boundaries, including photosensitivity limits;
• session records: their timeline, clock, events and observations;
• evidence assessments, their propositions, directions, tiers and study designs;
• the ecosystem layer: agents, qualified relationships and their provenance.

Out of scope, deliberately:

• clinical indications, diagnoses and treatment recommendations;
• personal or participant data of any kind;
• an event-type vocabulary that would duplicate HED, which SSTIM instead maps to;
• audio or video encoding formats, which SSTIM references rather than redefines.

Namespace and identifiers

The vocabulary's namespace is https://w3id.org/sstim, which content-negotiates: an RDF request returns the namespace catalogue in Turtle, JSON-LD or RDF/XML, and an HTML request returns a page describing the namespace. Term IRIs are hash IRIs with human-readable local names, for example https://w3id.org/sstim#SensoryStimulation and https://w3id.org/sstim/vocab#techBinauralBeats.

Four term namespaces are in use, each dereferenceable:

An implementer should resolve a term rather than construct it from a prefix registration alone. The prefix sstim is registered at prefix.cc for https://w3id.org/sstim#.

Modules and profiles

SSTIM is one suite with a synchronised release version, distributed as modules. A module owns the terms for one concern and declares its direct dependencies. A profile names an exact, transitively closed consumer contract. A consumer implements a profile, not the whole graph.

The machine-readable bill of materials is the suite manifest at https://w3id.org/sstim/manifest, which declares every module, its dependencies, its profile membership, its checksums and its publication locations. Where this document and the manifest disagree, the manifest is authoritative.

Conformance

A graph conforms to a profile when it validates against that profile's SHACL shape closure and uses only terms the profile's closure defines. Conformance is per profile: a Core-conformant graph need not carry evidence or session terms, and a consumer of Core must not require them.

The Core shape package is deliberately weak. A stimulus channel and a target may be omitted; what is constrained is what an asserted value must be. This is so that a description that knows less is still valid rather than unrepresentable.

Every released profile ships a positive fixture and at least one executable competency query. Every profile with a non-empty shape closure also ships out-of-scope and adversarial fixtures, so that a claim of conformance can be reproduced rather than trusted.

External mappings

SSTIM maps its terms to Wikidata, MeSH, UBERON and SNOMED CT, conservatively and in one direction, using the SKOS mapping relations. Two rules govern them.

First, an identifier is verified at the authority that mints it before it enters the graph, and re-verified continuously; label similarity has never been sufficient evidence, and three published mappings were withdrawn for being label-alike and semantically wrong.

Second, containment is stated rather than flattened. Where an external concept is broader or narrower, the mapping says so with skos:broadMatch or skos:narrowMatch rather than defaulting to skos:closeMatch. Every mapping carries an annotated axiom naming its source, its verification date and the reason for its predicate.

Mappings to SNOMED CT reference identifiers only and carry no SNOMED content, which is a licensing constraint rather than an editorial one.

Versioning and citation

The live development line is mutable and must not be cited. Each release is frozen, byte-identical, archived with a DOI, and identified by a versioned IRI of the form https://w3id.org/sstim/0.17.0, which is the current release. A concept DOI names the continuing project across releases; a version DOI pins one release. The vocabulary's persistent identifier resolves to the latest release rather than to the working line, so dereferencing https://w3id.org/sstim hands a consumer a graph that can be cited.

The vocabulary is published under CC BY 4.0. That licence covers the RDF and its documentation; the copyright statement on this report is the Community Group one above, which is a different thing and is governed by the Contributor License Agreement.

An implementation that depends on exact term semantics should import a versioned profile entry point, whose closure names its exact versioned siblings, rather than the mutable latest one.