AAC Device Selection: A Feature-Matching Guide

AAC selection is a feature-matching process, not a shopping decision. You profile the person's sensory, motor, language, and cognitive abilities alongside their communication partners and environments, then choose the system whose features line up. Get the order backwards — pick the device first, then try to make the person fit it — and abandonment follows.

The vocabulary you are expected to know

• Unaided AAC uses only the body: gestures, sign, facial expression. Aided AAC uses external materials, from a paper board to a speech-generating device (SGD).
• Low tech (communication boards, PECS books, alphabet boards) versus high tech (dedicated SGDs, tablets running AAC apps, eye-gaze systems).
• Symbol iconicity ranges from transparent (guessable without training) through translucent to opaque. Photographs and line drawings sit at the transparent end; orthography is opaque but the most powerful.
• Access method: direct selection (touch, eye gaze, head pointer, joystick) versus scanning (auto, step, or inverse; linear, row-column, or block).
• Rate enhancement: word prediction, semantic compaction (Minspeak), abbreviation expansion, and pre-stored phrases.

Core vocabulary is the foundation

Roughly 80% of what anyone says comes from a few hundred high-frequency core words — want, go, stop, more, help, that, not, you, do. Fringe vocabulary is the noun-heavy, situation-specific set: pizza, giraffe, bulldozer. Systems built mostly of fringe nouns produce labeling, not conversation. A robust AAC system puts core vocabulary in stable, motor-planned locations and layers fringe around it. Consistent button placement across pages supports motor automaticity — the same principle behind touch typing.

Feature matching, step by step

1. Assess vision and hearing. Field cuts, acuity, and contrast needs determine symbol size, background color, grid density, and whether auditory scanning is required.
2. Assess motor access. Can the person reliably isolate a finger? Is there tremor, spasticity, or fatigue? Test direct selection first — it is faster and less cognitively demanding than scanning. Move to scanning only when direct access is not reliable.
3. Assess language and literacy. A literate user gets a keyboard with word prediction, which is unlimited in what it can say. A pre-literate user gets a symbol system with a literacy instruction plan attached — spelling access is the long-term goal for anyone who can reach it.
4. Assess cognition and attention. Grid size, navigation depth, and page structure follow from working memory and executive function.
5. Map environments and partners. A device that fails in a pool, on a school bus, or with a grandparent who cannot program it will be abandoned. Durability, mounting, battery life, volume, and partner training are selection criteria, not afterthoughts.
6. Trial before you buy. Loaner trials with data collection are the standard, and funders expect documented trial results.

Myths the exam will test

AAC does not suppress speech. The evidence is consistent and it is one of the most reliably keyed facts in this content area: AAC use is associated with equal or increased natural speech production. There are no prerequisites for AAC. Candidacy models requiring a cognitive age, a demonstrated cause-and-effect skill, or a certain receptive language level are outdated; the participation model replaced them. Multimodal is the goal. Speech, gesture, sign, low tech, and high tech coexist — a person needs a backup when the device is charging or wet.

Teaching the system once it arrives

Aided language stimulation — the partner models on the device while speaking — is the primary teaching strategy. Model more than you expect back, model without demanding a response, and model at or slightly above the user's current level. Add wait time; AAC users need dramatically longer to formulate a message, and partners routinely interrupt before the message is complete.

Structure opportunities with communication temptations: sabotage routines, offer choices, give an inadequate portion, present a preferred item in a sealed container. Then honor the message, including rejection and protest. A system that only says "I want" teaches compliance, not communication.

Special populations

• ALS and progressive disease: refer early, before speech is unintelligible. Message banking and voice banking must happen while the voice is still usable, and access will need to migrate — touch to head pointing to eye gaze — as the disease progresses.
• Cerebral palsy: positioning drives access. Seating and mounting evaluation with OT/PT precedes access decisions.
• Autism: AAC supports both requesting and the full range of communicative functions. PECS is a structured requesting protocol with strong early evidence; plan a path from PECS toward a robust core-word system rather than treating it as the endpoint.
• Aphasia: partner-dependent systems, remnant books, and written choice conversation often outperform grid-based SGDs.

Funding and documentation

SGD funding through Medicare, Medicaid, or private insurance requires an SLP evaluation documenting the medical diagnosis, the communication impairment, why less costly options are insufficient, the trial results, and the treatment plan. Writing "patient would benefit from an SGD" is not enough; funders want the feature-match rationale in writing.

Exam angle

Expect vignettes that describe a user's motor and language profile and ask for the appropriate access method or symbol set, plus items testing the no-prerequisites principle and the AAC-does-not-inhibit-speech finding. Practice with AAC practice questions.