The research behind Visual Choices

How do you get an autistic individual to communicate more? There is no single approach that works for every child. However, there are some methods considered broadly effective. Visual Choices is built on research from psychology and speech-language pathology. Here are the studies that shaped it, in roughly the order a child moves through the app.

Making a choice is harder than it looks

Choosing between two pictures seems simple. Yet when we survey parents, about a quarter of them indicate that their child cannot make a choice between two items on a screen.

Share of children who can't yet choose between two items on a screen, by ageBar chart. Ages 0 to 3: 33% of 1,043. Ages 4 to 7: 19% of 976. Ages 8 to 11: 13% of 211. Ages 12 to 15: 6% of 87. Ages 16 to 18: 6% of 17. Ages 19 and up: 19% of 86.0%10%20%30%40%Can't yet make a choice33%0–3n=1,04319%4–7n=97613%8–11n=2116%12–15n=876%16–18n=1719%19+n=86Child's age (years)
Parents who said their child can't yet choose between two items on a screen, at the start of a Visual Choices trial. n is the number of parents in each age group.

Why the difficulty? There are several things a child needs to know. Pictures represent objects in the real world. Tapping a picture means that you want the object. Getting one choice excludes getting the other.

And most importantly, there's the basic idea of communication: if you want something, you get it by directing some request to another person, rather than just getting it yourself.

All of these ideas tend to be picked up quickly by neurotypical individuals. But autistic individuals may need explicit teaching and practice to learn choice making, one of the most basic units of social communication.

Picture exchange: small steps to requesting, and to speech

The Picture Exchange Communication System (PECS) offers a system for teaching choice making. It proceeds in phases. In Phase 1, the child learns to exchange a picture card for an object. In Phase 2, they learn to find a person to hand the card to, even if the person and object are far removed. In Phase 3a, they learn to choose between an item they like and one they don't. In Phase 3b, they learn to choose between two objects they like.

Some families and clinicians find the phases rigid, but each one teaches something a child needs: that communicating gets results, that it takes another person, and that different pictures mean different things.

The Visual Choices progression of single item to binary choices to whole board selections is based on these initial PECS phases.

Importantly, children taught with PECS learn to talk just as much as children given explicit speech training. In a randomized trial, 39 minimally verbal toddlers, about 2½ years old, got six months of either PECS or Pivotal Response Training (PRT), a play-based method that targets speech directly.1 The speech improvements were similar.

Average spoken words before and after picture-based and speech-based interventionLine chart. Picture exchange group: 5 words at start, 89 at the end of treatment, 130 three months later. Speech-focused group: 12, 83, then 113.020406080100120140Spoken words (average)StartEnd (6 months)3 months laterPictures (PECS): 130Speech (PRT): 113
Average spoken words, reported by parents. Individual children varied widely, and the study had no untreated comparison group. Schreibman & Stahmer (2014), Table 2.

This is counterintuitive to most people. Why would exchanging picture cards boost speech as much as practicing words directly? The key is that speech is a skill at the top of a pyramid. Teaching prerequisite communication and language skills first often pushes speech to emerge later.

Activity boards put the words for a game on one page

Once a child can request, the next step is combining words into phrases. For typically developing children, this happens at around 18 months of age, when children begin combining nouns and verbs, nouns and adjectives, and more.

These words don't need to be spoken. What a child does need is to know and access about 50 to 100 words, with a mix of different parts of speech.2 This is an ideal use for AAC, which can summon up words at the touch of a button and give your child a jump-start on language.

An article by Cathy Binger and colleagues emphasizes the value of AAC for this kind of language development, and dives into the pros and cons of different systems.3

Robust AAC is great for building vocabulary. It has tens of thousands of words organized into folders, and children can explore words freely.

However, there are two issues with robust AAC. First, it tends to emphasize high-frequency “core words,” e.g. he, she, with, to, up, down. These words are abstract and hard to use for language building. Second, it tends to spread “fringe” nouns, verbs, and adjectives across different pages.

“[W]hen children only have access to complex solutions that require the child to navigate to various screens to locate vocabulary, expressive grammar development is necessarily delayed due to challenges in locating words.”

Binger, Harrington & Kent-Walsh (2024)3

For language development, activity boards are an appealing addition. They hold the words for one activity on a single page, e.g. pig, giraffe, ride, throw, eat, car, cookie. You are probably thinking of some fun ways these words can be combined already!

Visual Choices makes it easy to create activity boards by taking pictures of toys, then saying the names of a few verbs and adjectives. The app automatically brings in the needed symbols. When children practice word combinations through AAC, they build up their language quickly.

“Adding activity scene displays with vocabulary focused on particular activities eliminates navigational demands, with many children readily increasing their expressive word diversity and grammar skills.”

Binger, Harrington & Kent-Walsh (2024)3

Word combinations in a few short play sessions

Binger and Light noticed that most AAC users only produced single-symbol messages. They wondered if they could be taught to produce longer messages.

They worked with five preschoolers, ages 3 to 5, who used AAC because of conditions such as Down syndrome or childhood apraxia of speech.4 None were autistic. During play, an adult tapped symbols on the child's board while saying a full sentence, then encouraged the child to tap symbols as well.

Four of the five children began combining symbols on their own. Two reached the study's goal, at least 12 combinations in each of three sessions in a row, after just three and five 15-minute sessions.

Word combinations per session before and after aided AAC modeling, for five preschoolersDot chart. Valerie 5 to 11. Timmy 2 to 17.3. Robyn 1 to 3.6. Nathan 0.7 to 9.3. Richard 0.3 to 5.05101520Different word combinations per 15-minute sessionBeforeAfterValerie11Timmy17.3Robyn3.6Nathan9.3Richard5
Average number of different word combinations per 15-minute play session. Robyn, the one child who didn't reach the goal, is shown for her first intervention phase. Binger & Light (2007), Table 2.

Teaching play builds language

Many autistic children find play itself hard. They may line toys up or repeat one action, rather than build a game with another person. JASPER (Joint Attention, Symbolic Play, Engagement, and Regulation), developed by Connie Kasari's group at UCLA, teaches play and joint attention directly, following a developmental sequence.5

In an early trial, 58 autistic children aged 3 and 4 got 30 minutes a day of joint attention or play instruction for 5 to 6 weeks, or no added instruction.6 A year later, children in both instruction groups had gained nearly twice as much expressive language as the control group.

Expressive language age over one year after a six-week play or joint attention interventionLine chart, approximate values. Play group: about 22 months at start, 39 at one year. Joint attention group: about 21 to 36. Control group: about 20 to 29.010203040Expressive language age (months)Start6 weeks6 months12 monthsPlayJoint attentionControl
Expressive language age, in months. Approximate values read from Figure 2 of Kasari et al. (2008).

Why? Play keeps a child jointly engaged with another person for longer, so they hear and practice more language. And symbolic play draws on the same ability language does: letting one thing stand for another.7

Play and AAC together

The same group then asked what happens when AAC activity boards are added. In a trial with 61 minimally verbal autistic children, ages 5 to 8, every child got play-based therapy, and half also used a speech-generating device during play.8

These children were older than the preschoolers above, and had remained minimally speaking despite years of intervention. At their age, most clinicians would be focused on getting them proficient with robust AAC, rather than expecting their speech to improve.

Progress took longer than with the preschoolers: 48 hours of therapy over 24 weeks, rather than a few sessions. But by week 24, the children using AAC activity boards in combination with play were producing about 50% more spontaneous communication than the children doing the play-based therapy only.

Spontaneous communicative utterances with and without AAC added to play-based therapyLine chart. Both groups start near 30 utterances. With AAC: 54 at week 12, 62 at week 24, 53 at week 36 follow-up. Without AAC: 35, 40, 45.010203040506070Spontaneous communication (20 min)StartWeek 12Week 24Week 36With AACWithout AAC
Spontaneous communicative utterances in a 20-minute sample. Weeks 1 to 24 were therapy; week 36 was a follow-up. Model-estimated means from Kasari et al. (2014), Table 2.

And remarkably, over 90% of their social-communicative utterances were spoken, not tapped. The Kasari group has video after video of once minimally speaking children now talking spontaneously and cheerfully during play.

Visual Choices did not exist during the time of their study, but Kasari's group now uses it in their JASPER intervention studies. They like how easy it makes it to program activity boards relevant to their play activities.

We've received federal funding to develop an AAC app called Visual Play, with knowledge built in about play development, aiming to make programming and using AACs during play even easier. You can watch our talks at aacforplay.com.

Sign language

Signs are another route. In a study of 82 autistic toddlers in Italy, about 22 months old with no spoken words, half were taught to request with PECS and half with manual signs, alongside 15 hours a week of ABA therapy.9 The PECS group learned 20 requests sooner. But the signing group moved from those requests to saying the words sooner.

The researchers believe that imitating and learning signs is more effortful than picture exchange, so signs take children longer to learn. At the same time, signing back and forth might stimulate social communication growth to a greater extent, leading to faster speech acquisition.

This is all speculative, but it fits with our general understanding of what's needed to get autistic children communicating. That's why Visual Choices includes hundreds of sign language videos: language practice that can continue even when no device is available.

What this adds up to

AAC works best as a teaching tool, not just a way to ask for things. Start where the child is, with a single photo, then a choice. Use activity boards during play to build word combinations. Teach play itself. That's the order Visual Choices is built around.

It's not just about the app. It's about what parents and clinicians do with the app, together with the child. That's why we include many guide videos on how to work with your child, as well as a comprehensive guide.

References

  1. Schreibman, L., & Stahmer, A. C. (2014). A randomized trial comparison of the effects of verbal and pictorial naturalistic communication strategies on spoken language for young children with autism. Journal of Autism and Developmental Disorders, 44(5), 1244–1251.
  2. Bates, E., & Goodman, J. C. (1997). On the inseparability of grammar and the lexicon: Evidence from acquisition, aphasia and real-time processing. Language and Cognitive Processes, 12(5–6), 507–584.
  3. Binger, C., Harrington, N., & Kent-Walsh, J. (2024). Applying a developmental model to preliterate aided language learning. American Journal of Speech-Language Pathology, 33(1), 33–50.
  4. Binger, C., & Light, J. (2007). The effect of aided AAC modeling on the expression of multi-symbol messages by preschoolers who use AAC. Augmentative and Alternative Communication, 23(1), 30–43.
  5. Kasari, C., Gulsrud, A. C., Shire, S. Y., & Strawbridge, C. (2021). The JASPER model for children with autism: Promoting joint attention, symbolic play, engagement, and regulation. Guilford Press.
  6. Kasari, C., Paparella, T., Freeman, S., & Jahromi, L. B. (2008). Language outcome in autism: Randomized comparison of joint attention and play interventions. Journal of Consulting and Clinical Psychology, 76(1), 125–137.
  7. Ungerer, J. A., & Sigman, M. (1981). Symbolic play and language comprehension in autistic children. Journal of the American Academy of Child Psychiatry, 20(2), 318–337.
  8. Kasari, C., Kaiser, A., Goods, K., et al. (2014). Communication interventions for minimally verbal children with autism: A sequential multiple assignment randomized trial. Journal of the American Academy of Child & Adolescent Psychiatry, 53(6), 635–646.
  9. Frolli, A., Ciotola, S., Esposito, C., et al. (2022). AAC and autism: Manual signs and PECS, a comparison. Behavioral Sciences, 12(10), 359.

More in our full bibliography.