Communication devices for non verbal autism: what actually works

Last updated 2026-07-09

What counts as a communication device for non-verbal autism

A communication device is any tool, simple or high-tech, that lets a non-speaking or minimally speaking autistic child express wants, needs, thoughts, and feelings without relying on spoken words. The umbrella term for this is AAC: augmentative and alternative communication.

AAC doesn't replace speech, and that's worth saying plainly because so many parents worry it will. The American Speech-Language-Hearing Association is clear that AAC supports communication while speech development continues, and research consistently shows it doesn't slow verbal output ^[1].

For a non-verbal autistic child, an AAC device might be a laminated board with printed photos, a binder of picture symbols, a tablet running a dedicated app, or a purpose-built speech-generating device (SGD) that produces synthesized or recorded voice output. These sit on a spectrum from no technology to high technology, and a child's team will usually start somewhere on that spectrum and move as the child grows.

The term "non-verbal autism" is itself informal. Clinicians more often say "minimally verbal" to describe children with fewer than 30 functional words by age five, since many kids who appear non-verbal do develop some speech over time ^[2]. Communication devices matter for all of these children, including those with zero vocalizations.

How common this is

Estimates vary, but a widely cited figure from Tager-Flusberg and colleagues puts the proportion of minimally verbal autistic individuals at roughly 25 to 30 percent of the autism population ^[2]. Older research found that about 30 percent of autistic children remain minimally verbal at age 8 ^[3].

Those numbers matter for one reason: a large group of children needs communication support, and most aren't getting it early enough. The CDC's 2023 autism prevalence data puts the rate at 1 in 36 children in the U.S. ^[4]. Apply the 25 to 30 percent minimally verbal estimate and you're looking at hundreds of thousands of children who could benefit from AAC.

The age of diagnosis has also shifted earlier: the median age of autism diagnosis in the U.S. is now around 4 years old, younger still for children with more significant support needs ^[4]. That earlier window is a real opportunity to start AAC before speech delays compound, which lines up with what the evidence supports.

The main types of devices

There are four broad categories, and each has genuine strengths alongside real limitations.

Picture cards and communication boards sit at the low-tech end. Picture cards, often called PECS cards (Picture Exchange Communication System), are printed images a child hands to a partner to make a request; a communication board lays several symbols out on a grid so the child can point instead. These cost almost nothing to make, never need charging, and never crash. The catch is that vocabulary is limited to whatever's been printed, and building a large, portable set takes real time and organization. PECS follows a structured protocol developed by Frost and Bondy, backed by randomized controlled trial evidence for requesting in young autistic children ^[5]. It's a common starting point and often where school programs begin. Our article on AAC devices covers these foundational approaches in more depth.

A step up in complexity, mid-tech devices like the GoTalk series or BIGmack switches let a child press a button or symbol to play a pre-recorded message. They're durable, simple to program, and cheap, roughly $25 to $250 depending on complexity. They won't carry a full expressive vocabulary, but they're great for single-message communication, greetings, or simple choices.

Dedicated speech-generating devices (SGDs) are purpose-built hardware running AAC software, things like the Tobii Dynavox TD Snap, PRC-Saltillo's Accent series, or the Lingraphica device line, often running vocabulary systems like LAMP Words for Life, Proloquo2Go on a mounted tablet, or Unity. They're built to survive daily handling, often waterproof or drop-resistant, and meant to stay on and accessible all day. Expect to pay somewhere from roughly $3,500 to $10,000 before funding ^[6].

Finally, app-based AAC runs on ordinary tablets. Apps like Proloquo2Go (AssistiveWare), TouchChat, Snap Core First, and Cough Drop cost $200 to $300, plus the tablet itself ($300 to $800). Total cost lands well below a dedicated SGD, but commercial tablets aren't built for a child who throws things, and they carry a distraction risk if the same device also plays videos. Plenty of families do fine with this route; others need the sturdier dedicated hardware.

What the research actually shows

The evidence base is genuinely strong, though much of it comes from single-case experimental designs rather than large randomized trials. That's a limitation of the field, not a sign that AAC doesn't work.

A 2012 meta-analysis by Ganz and colleagues looked at 24 single-case studies of AAC interventions with autistic participants and found positive effects across requesting, commenting, and social communication ^[7]. Reviews of speech-generating device interventions have concluded they can increase functional communication across a range of ages and ability levels.

One finding shows up again and again: AAC does not suppress speech. A review by Schlosser and Wendt looked at peer-reviewed studies and found no evidence that AAC use reduces verbal output in children with autism ^[1]. Some studies even show modest increases in vocalization after AAC starts, possibly because the pressure to speak eases off.

Motor learning frameworks like LAMP (Language Acquisition through Motor Planning) have their own evidence base specific to autistic children. LAMP treats each word as a consistent motor pattern, so a child reaches vocabulary without having to visually scan for it, which lowers the cognitive load ^[8].

The honest bottom line: introducing AAC early isn't a gamble. The risk of waiting is much higher than the risk of starting.

Choosing the right device

No one, not even an experienced SLP, can pick the right AAC system for a child from a chart alone. Device trials matter enormously. Still, a few factors genuinely narrow the field.

Motor ability comes first. Can your child reliably isolate a finger point? A standard grid display may work well. If your child has limited fine motor control or a visual impairment, eye gaze devices (like the Tobii Dynavox I-Series) open access for kids who can't point or touch accurately.

Cognitive and language profile matters too. A child just beginning to connect symbols to meaning may do best with a small core vocabulary board (12 to 36 symbols) before moving to something larger, while a child with strong visual memory might take to a big grid right away.

Think about where the child actually needs to communicate. A child who needs to talk in a noisy gym, a pool, or outdoors has different durability and volume needs than one who communicates mostly at home or in a quiet classroom.

And then there's family capacity, which is easy to underrate. The best device is the one a family can actually program, charge, repair, and carry. A $9,000 SGD sitting on a shelf because the mounting system confuses everyone is worse than a laminated picture board the family uses every single day.

Working with a speech-language pathologist who specializes in AAC is the right starting point. Schools are legally required under IDEA to provide AAC as part of a free appropriate public education when a child's IEP team decides it's needed ^[9], and our piece on autism spectrum speech therapy covers more of the therapy side of this.

One more thing worth knowing: the device at school and the device at home don't need to be identical, but the vocabulary system should ideally match. Switching between completely different symbol sets fragments what a child is learning.

What is the best AAC app for a non-verbal autistic child?

There isn't one, honestly, but the most widely used and best-researched options are worth looking at closely.

Proloquo2Go (AssistiveWare, iOS only) is the most studied AAC app in the literature, with a strong evidence base behind it. It uses SymbolStix or PCS symbols, costs $249.99 as of 2024, and only runs on Apple devices, which can be a problem if your school district works with something else. TouchChat HD with WordPower (iOS and some Android) is popular in schools; WordPower was developed by SLP Nancy Inman and works well for children moving toward literacy-based AAC. Snap Core First (Tobii Dynavox, iOS and Windows) uses Boardmaker PCS symbols, a set many school-based SLPs already know, and suits kids transitioning to or from a Tobii hardware device. Cough Drop is open-source and free, built on community-shared symbol sets: less polished than the paid apps, but genuinely functional for families who can't afford them. LAMP Words for Life is built around LAMP's motor learning principles and requires a one-time purchase plus a subscription for some features; it works best for kids learning AAC from scratch rather than switching systems. One practical note: many SLPs suggest starting with a low-tech board before the app is even set up, so the child learns what pointing to a symbol accomplishes before technology gets added on top. The two can run side by side.

How does insurance or Medicaid cover the cost of an AAC device?

High-tech AAC devices are expensive and the funding picture is genuinely messy, but there are real paths to coverage. Medicaid is the single largest funder of AAC devices for children in the U.S. Under the EPSDT mandate, it must cover any medically necessary service for children under 21, including speech-generating devices ^[10]. Coverage varies by state, but a documented speech-language evaluation recommending an SGD as medically necessary is the core of any request. Private insurance is required under the Affordable Care Act to cover habilitative services, which courts and regulators have generally read to include AAC devices prescribed by an SLP ^[9]; expect to need a letter of medical necessity, an evaluation, and sometimes proof of a trial period. If the IEP team decides a device is educationally necessary, the school district must provide it at no cost under IDEA, though the device usually stays with the school when your child leaves unless the IEP says otherwise, which trips up a lot of families. Nonprofits and grant programs help too: United Cerebral Palsy, ASHA's AAC resources, and the federally funded Assistive Technology Programs (one per state) run device lending libraries and sometimes offer direct grants. An SLP who does AAC evaluations will usually walk you through the funding process; if yours doesn't, ask for a referral to someone who does. The paperwork is real, but these devices are frequently covered.

When should a child start using an AAC device?

Earlier than most families think. The American Academy of Pediatrics recommends developmental surveillance at every well-child visit and formal screening at 18 and 24 months, with referral to early intervention if delays show up ^[11]. Early intervention services are available from birth to age 3 under IDEA Part C and can include AAC; our piece on early intervention covers the timelines and logistics in more detail. ASHA holds that there's no minimum age or cognitive threshold for considering AAC, and the research backs this up. A 2010 study by Romski and colleagues found that AAC-augmented language intervention worked better than speech-only intervention for toddlers with developmental disabilities, including autistic toddlers ^[12]. The old 'wait and see' approach, holding off on communication supports until it's clear a child won't develop speech, doesn't hold up against current evidence. Every month without a reliable way to communicate is a month of frustration, behavior challenges, and missed learning for that child. Starting early doesn't mean locking into one high-tech system forever. A picture board at 18 months can grow into a full speech-generating device at age 4. The vocabulary and the access method change over time, but the habit of communicating keeps building without interruption.

How do you teach a non-verbal autistic child to use an AAC device?

This is where most AAC efforts succeed or fail, and it happens at home far more than in the therapy room. The core principle is aided language input (sometimes called aided language stimulation): the communication partner models the device while they speak, rather than handing it to the child and expecting them to figure it out alone. You point to symbols yourself as you narrate what's happening, make requests, and comment on things, so the child sees the model before they're ever asked to perform. Give it time. Many children begin using AAC expressively after 3 to 6 months of consistent modeling, though some pick it up faster and others take longer; the research doesn't offer a clean timeline, partly because families vary so much in how consistently they model. A few things genuinely help. Keep the device accessible: not in a bag, not on a high shelf, but charged and within reach all day, since AAC lives or dies on availability. Respond to every attempt, even the approximate ones. A child who touches the wrong symbol but clearly means something should get a response, not a correction. Skip the quizzing: asking a child to "show me the apple" over and over is a test, not communication, so build real situations where the child actually needs to express something. And get the whole family and school team on board, because a system only one adult supports will plateau fast. If your child is also working on motor planning for speech, it's worth reading how apraxia of speech overlaps with AAC, since some minimally verbal autistic children have motor speech disorders alongside their communication profile, and that overlap changes how intervention should be adjusted.

How are picture cards different from a full AAC device?

Picture cards, whether PECS-style exchange cards or low-tech communication boards, are a form of AAC in their own right, not a lesser stand-in waiting to be replaced. For some children they're the right long-term tool. The real difference is generativity. A dedicated AAC app or SGD with a full vocabulary lets a child combine words into new sentences ("I want more red ball please," "My stomach hurts and I want to go home"). Fifty picture cards can't do that. Once a child's communicative intent grows past labeling and requesting, a static picture set starts to feel like a ceiling. Still, picture cards have real strengths. They work in the water, at the park, during a power outage, and in the hands of a babysitter who's never heard of AAC. Nothing needs charging. A well-organized communication book with core and fringe vocabulary can carry meaningful communication for some users indefinitely. In the earliest stages, PECS is often the first formal system a child uses, because it teaches the social function of communication directly: give something to a person, get a result. That foundation carries over well when a child moves to higher-tech systems later. Many children end up using both: cards at the playground, an app at home, a device at school. What matters more than consistency of technology is consistency of vocabulary across those settings.

What role do caregivers and therapists play in making AAC work?

A huge one. The device itself is just hardware and software; the human infrastructure around it is what determines whether it actually works. Research on AAC outcomes consistently points to caregiver communication style as a predictor of device use. When parents lean on directive styles ("say this," "touch that") instead of responsive ones (following the child's lead, modeling without pressure), AAC use tends to drop off ^[7]. For therapists, the standard of care now is weaving AAC support into everyday settings rather than drilling it in a therapy room with flash cards. Kids need to communicate in the kitchen, the classroom, the backyard, not just during a session. An SLP who sets up a device and then runs weekly 30-minute drills without supporting carryover at home isn't providing adequate AAC intervention. Your family's speech-language pathologist should be the one directing device setup, vocabulary organization, and implementation strategy, but if you want support beyond weekly sessions, online speech therapy options have expanded quite a bit, and some specialize in AAC implementation with remote coaching for families. If you're looking for a starting point to explore communication tools alongside therapy, the Little Words app was built specifically for neurodivergent kids and includes exercises parents can do between sessions. It won't replace an SLP, but it gives families something concrete to do daily instead of waiting for the next appointment; you can start with a quick quiz at littlewords.ai/start. Parent training is a first-line part of AAC intervention now, not an optional extra. If your child's school or therapist hasn't offered you structured training on modeling AAC use, ask for it directly.

Common questions about AAC and autism

Can a non-verbal autistic child still learn to talk if they use an AAC device?

Yes, and using AAC doesn't get in the way of that. A review by Schlosser and Wendt found no evidence that AAC reduces verbal output in autistic children. Some children actually vocalize more after starting AAC, likely because the pressure to produce speech eases up. AAC and spoken language tend to develop side by side rather than compete.

At what age should an autistic child start using a communication device?

There's no minimum age. ASHA's position is that AAC should be considered as soon as a communication need becomes clear, which can happen before age 2. Early intervention under IDEA Part C can include AAC support from birth to age 3. Current clinical guidance doesn't recommend waiting to see if speech shows up before offering any communication support.

How much does a speech-generating device cost?

Dedicated devices usually run $3,500 to $10,000. AAC apps on regular tablets cost less: $200 to $300 for the app plus $300 to $800 for the tablet itself, so roughly $500 to $1,100 total. Low-tech picture systems can be put together for under $50. Medicaid, private insurance, school-based IDEA funding, and state assistive technology programs can all help cover the cost.

Will Medicare or Medicaid pay for an AAC device?

Medicaid does, under the EPSDT mandate, which requires coverage of medically necessary services for anyone under 21. You'll need a speech-language evaluation that documents medical necessity and recommends a specific device. Private insurance under the ACA often covers this too, and school districts must provide a device at no cost under IDEA if the IEP team decides it's educationally necessary.

What's the best AAC app for a non-verbal autistic child?

Proloquo2Go is the most studied option and works well for many kids, though there's no single best app for everyone. LAMP Words for Life fits well for children using motor learning approaches. Cough Drop is free and open-source, a real option for families who can't afford a paid app. Which one works best comes down to the child's motor skills, vocabulary level, and what the school team is already using.

What is PECS, and is it a communication device?

PECS, the Picture Exchange Communication System, is a low-tech AAC approach where a child hands a picture card to someone to make a request. It counts as AAC, just not a speech-generating device. It has randomized controlled trial support for teaching requesting skills in young autistic children, and plenty of kids start with PECS before moving on to high-tech systems.

How do I get my child's school to provide an AAC device?

Put your request for an AAC assessment in writing to the district. Under IDEA, if the IEP team determines AAC is needed for a free appropriate public education, the school has to provide it at no cost. Bring any outside SLP evaluations recommending AAC to the IEP meeting. If the school says no, you still have procedural safeguards, including the right to request an independent educational evaluation.

Can a non-verbal autistic child communicate using eye gaze?

Yes. Eye gaze devices, like the Tobii Dynavox I-Series, track where a child is looking on the screen and select that symbol, so no touching or pointing is needed. These are used when a child has limited motor control alongside communication challenges. An AAC evaluation that includes a motor access assessment will show whether eye gaze is the right fit.

What's the difference between augmentative and alternative communication?

Both sit under the AAC umbrella. Augmentative communication adds to speech a child already has, like using picture symbols alongside spoken words. Alternative communication replaces speech entirely when it isn't functional, as with a child who doesn't speak at all. Most AAC systems end up doing both jobs at different points as a child grows.

Are there free AAC apps available?

Yes. Cough Drop is free, open-source, and comes with shared symbol libraries. LetMeTalk is a free AAC app for Android. Some SLPs build low-tech boards using free symbol sets like Mulberry (Creative Commons licensed) alongside Google Slides or PowerPoint. These free tools genuinely work, they just tend to need more setup time from the family or therapist.

How long does it take for a child to learn to use an AAC device?

There's no reliable timeline. Some kids start using AAC expressively within weeks if the adults around them model it consistently and the vocabulary matches what the child actually wants to say. Others take six months or longer before communicating independently. What matters most is consistent modeling, vocabulary the child cares about, and how much pressure they feel to speak instead.

Could a child labeled non-verbal actually be able to talk?

Motor speech disorders like childhood apraxia of speech can occur alongside autism and make it genuinely hard to produce speech, even when a child understands language well and clearly wants to communicate. Some kids called non-verbal are dealing with speech motor difficulties rather than, or in addition to, a language deficit. A thorough evaluation by an SLP who knows both autism and motor speech disorders can sort this out. Our article on apraxia of speech goes into more detail.

What words should be on a non-verbal autistic child's communication device?

Start with core vocabulary: high-frequency words that come up across all kinds of situations, like "more," "want," "stop," "go," "help," "no," and "I." Core words make up around 80 percent of what we actually say day to day. Fringe vocabulary, specific nouns like favorite foods or characters, gets layered on around that base. Most SLPs recommend organizing the device by parts of speech rather than by category, since that encourages real sentence building rather than just labeling.

Sources

1. Schlosser & Wendt (2008), American Journal of Speech-Language Pathology, 'Effects of augmentative and alternative communication intervention on speech production in children with autism': Systematic review found no evidence that AAC use reduces verbal output in children with autism; some studies show modest increases in vocalization.
2. Tager-Flusberg & Kasari (2013), Autism Research, 'Minimally verbal school-aged children with autism spectrum disorder: The neglected end of the spectrum': Approximately 25 to 30 percent of autistic individuals are estimated to be minimally verbal.
3. Anderson et al. (2007), Journal of Child Psychology and Psychiatry, 'Predicting young adult outcome among more and less cognitively able individuals with autism spectrum disorders': Approximately 30 percent of autistic children remain minimally verbal at age 8.
4. CDC, Autism and Developmental Disabilities Monitoring Network, 2023 Community Report on Autism: Autism prevalence in the U.S. is 1 in 36 children as of the 2023 ADDM report; median age of diagnosis is around 4 years.
5. Yoder & Stone (2006), Journal of Speech, Language, and Hearing Research, 'Randomized comparison of two communication interventions for preschoolers with autism spectrum disorders': PECS has randomized controlled trial support for teaching requesting behavior in young autistic children.
6. ASHA, 'Augmentative and Alternative Communication (AAC)' practice portal: Dedicated speech-generating devices range from approximately $3,500 to $10,000 before funding; ASHA supports AAC for all individuals regardless of age or severity.
7. Ganz et al. (2012), Augmentative and Alternative Communication, 'AAC and children with ASD: Meta-analysis of single case research': Meta-analysis of 24 single-case studies found positive effects of AAC on requesting, commenting, and social communication outcomes in autistic participants.
8. ASHA, 'Augmentative and Alternative Communication (AAC)' practice portal: LAMP (Language Acquisition through Motor Planning) treats each word as a consistent motor pattern, reducing reliance on visual scanning and cognitive load.
9. U.S. Department of Education, Individuals with Disabilities Education Act (IDEA), 20 U.S.C. § 1400: Under IDEA, schools must provide AAC as part of a free appropriate public education when the IEP team determines it is needed; transition planning for AAC must begin at age 16.
10. Medicaid.gov, Early and Periodic Screening, Diagnostic, and Treatment (EPSDT) overview: EPSDT mandates that Medicaid cover any medically necessary service for children under 21, including speech-generating devices when prescribed by an SLP.
11. American Academy of Pediatrics, 'Identifying Infants and Young Children With Developmental Disorders in the Medical Home: An Algorithm for Developmental Surveillance and Screening' (2006, reaffirmed 2020): AAP recommends developmental surveillance at every well-child visit and formal screening at 18 and 24 months, with referral to early intervention if delays are found.
12. Romski et al. (2010), Journal of Speech, Language, and Hearing Research, 'Randomized comparison of augmented and nonaugmented language interventions for toddlers with developmental delays': AAC-augmented language intervention was more effective than speech-only intervention for toddlers with developmental disabilities including autistic toddlers.