The guidance is based on core strategies described in research papers:
More information on these papers and more can be found below. I've also gotten helpful information from books and conversations with experts.
I am not an academic, though I have an academic background, so I enjoy reading research papers. If you find yourself compelled to push past the invisible rope separating reasonable parent advocacy from professional authority, you may find yourself reading research papers too.
If you are in this position, here are a few things to keep in mind.
First, papers are not symbols of fixed authority. They are a conversation between researchers. Imagine a group of people trying to solve a problem together. They scatter off in different directions to try out different approaches, until one yells out, hey guys! I found something. Others run over and try to build on that, adding to the conversation. Newer papers aren’t always better, or more correct. Sometimes the conversation just drifts into a different direction, and it can be hard to get people excited about backtracking.
Second, research is hard, and research in this area is particularly hard. It’s hard to get enough kids to participate, hard to do the same thing the same way with each one, hard to measure outcomes objectively, hard to get statistical significance. You’ll find papers trying to combine the results of many smaller papers to make sense of it all.
And you’ll find that papers cite papers that cite papers all the way down to some primary study that may not support the original assertion quite as confidently as later papers proclaim. It’s like a game of telephone. For example, clinicians may say that “all the new research” indicates that modeling is better than physical prompting. But what is really happening is that there have been more recent papers talking about modeling, and there was this one study where three children with unspecified developmental delays were taught how to lace beads using video models. Or introduction after introduction will state that verbal ability correlates strongly with life success and well-being, when in fact the original paper found that verbal ability only correlates strongly with academic success, and in fact ALL autistic individuals tend to have many challenges in life. Always go back to primary sources.
Third, academic researchers tend to be dedicated people with a strong interest in helping the people they study. They are not the “ivory tower”; they are not people trying to cudgel you with authority or hide inconvenient truths to “get grant money”. They are people putting in the hard work to find hard answers to hard problems. And if you as a parent approach them, you may find yourself pleasantly surprised at how accessible they are — willing to send papers, give you the “scoop” on what the current research shows, and generally be respectful and supportive of you in your journey.
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. This is a foundational paper establishing 4 categories of play: simple, relational (what Kasari renames “Combination”), functional, and symbolic. They found that autistic children had a hard time with functional and symbolic play, and posited that it was a more general impairment of “the ability to form and manipulate symbols”, critical to language development. The transition from the sensorimotor stage to the preoperational stage is particularly hard for autistic children.
Lifter, K., Sulzer-Azaroff, B., Anderson, S. R., & Cowdery, G. E. (1993). Teaching play activities to preschool children with disabilities: The importance of developmental considerations. Journal of Early Intervention, 17(2), 139-159. Can play be taught to autistic children? Yes! But it’s important to find the child’s “play level”, and teach the next level up, not several levels ahead. When they do the former, children progress; when they try the latter, they do not.
Bondy, A. S., & Frost, L. A. (1994). The Picture Exchange Communication System. Focus on Autistic Behavior, 9(3), 1-19. This is the seminal paper on the PECS method. Bondy and Frost had developed the method starting in 1985, working with preschoolers at the Delaware Autism Program. Back then, AAC devices were bulky and prohibitively expensive. Smartphones and tablets did not exist. This was a real breakthrough — a portable, inexpensive system plus teaching method designed to promote SPONTANEOUS requesting by children.
Gillon, G. T. (2000). The efficacy of phonological awareness intervention for children with spoken language impairment. Language, Speech, and Hearing Services in Schools, 31(2), 126-141. If you can’t talk, learning to read is premature, right? Wrong! Teaching children with a speech delay basic phonological principles got them quickly to a similar reading level as non-speech delayed children — and their ability to pronounce words improved too. This is now one of the main treatment methods for childhood apraxia of speech.
Charlop-Christy, M. H., Carpenter, M., Le, L., LeBlanc, L. A., & Kellet, K. (2002). Using the picture exchange communication system (PECS) with children with autism: Assessment of PECS acquisition, speech, social-communicative behavior, and problem behavior. Journal of Applied Behavior Analysis, 35(3), 213-231. The initial paper on PECS found that 76% of children introduced to PECS eventually started talking. But it wasn’t clear this was caused by PECS alone — the children were receiving many different interventions, and were at varying levels of communicative ability. This paper took 3 minimally verbal autistic children and introduced PECS at different times. It found that PECS caused increases in speech and social communicative behaviors, as well as a decrease in problem behaviors.
Blischak, D. M., Lombardino, L. J., & Dyson, A. T. (2003). Use of Speech-Generating Devices: In Support of Natural Speech. Augmentative and Alternative Communication, 19(1), 29-35. The Charlop-Cristy paper showed that AAC did not inhibit speech, it increased it. This paper cited other studies showing similar results, and provided explanations for how AAC could help speech develop: raising the frequency of communication, reducing motor effort, and stimulating brain areas responsible for language.
Anderson, D. K., Lord, C., Risi, S., DiLavore, P. S., Shulman, C., Thurm, A., Welch, K., & Pickles, A. (2007). Patterns of growth in verbal abilities among children with autism spectrum disorder. Journal of Consulting and Clinical Psychology, 75(4), 594-604. This was a longitudinal study including 130 children referred for early intervention who were diagnosed at an early age (younger than age 2) with either autism or pervasive developmental disorder-not otherwise specified (the PPD-NOS category was later reclassified as autism). They tracked the children’s overall verbal ability at ages 2, 3, 5, and 9 years old. I took the liberty of replotting data from the paper in the figure below. Two aspects jump out: (1) ~40% of the children at age 9 remained at a minimal verbal level, with either no words at all or only a few single words — it wasn’t just a delay; these children NEVER progressed beyond the expressive level of an 18 month old child. (2) their progress before age 5 determined the SLOPE of future verbal gains; those who learned slowly continued to learn slowly, those who learned quickly continued to learn quickly. Gains compounded.
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. Many intervention papers follow the same format: (1) most children with this development delay can’t do this thing; (2) what if we taught them to do this thing; (3) wow! Now they can do this thing. Plus many other things we didn’t expect. In this study, Binger noted that many AAC users got stuck at single-word requesting — then taught 5 preschool aged children to make multi-word sentences using activity boards in play contexts. Amazingly, just a few 15 minute sessions was enough to get 4 of 5 of the children to produce multi-word sentences on their own, with results persisting at least 2 months after the study ended. This is the research basis for using activity boards (curated vocabulary in play contexts) rather than giving children access to thousands of words from day one.
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. Remember the study by Lifter et al showing that autistic children could be taught play skills? What happens to their language abilities when you do that? Quite a bit. This is the foundational study for Kasari’s JASPER method, and the basis for the guide’s “Improve Gestures and Play” and “An Intensive Push” sections. They worked with 58 children between the ages 3-4 enrolled in a UCLA early intervention program. These children were receiving 30 hrs/week of ABA through the program — they split the children into 3 groups: (1) control = no extra treatment; (2) joint attention = focus on improving gestures; (3) symbolic play = focus on improving play ability. They then did 30 minutes of intervention per day for 6 weeks. The result was striking. There was on average a YEAR boost in language ability 12 months after the program ended for both the joint attention and symbolic play groups relative to the control group. Both of the interventions had a similar outcome — the researchers hypothesized that the main thing that mattered was entering a state of “joint engagement” with the child. Think about it. All the children were getting 30 hrs of ABA per week. But that little extra push on gestures and play made all the difference in terms of verbal communication ability, with gains that continued to compound after the study ended. And from the Anderson paper, you can see the difference that year boost will continue to make in a child’s life.
Pickett, E., Pullara, O., O'Grady, J., & Gordon, B. (2009). Speech acquisition in older nonverbal individuals with autism: A review of features, methods, and prognosis. Cognitive and Behavioral Neurology, 22(1), 1-21.This is a hopeful paper, in a glass half-full glass half-empty kind of way. Parents (and clinicians) often believe that if a child doesn’t talk by age 5, they will never talk. But in this paper, they find 64 papers or books reporting 78 cases where children have begun to talk at age 5 or older. Here is the key data:
| Age of Speech Acquisition | Number of Children |
| 5 | 22 |
| 6 | 22 |
| 7 | 12 |
| 8 | 6 |
| 9 | 5 |
| 10 | 3 |
| 11 | 5 |
| 12 | 2 |
| 13 | 1 |
It seems that speech is still attainable from ages 5 to 7, with cases trailing off after that. Here’s the glass half-empty part — there must have been 100,000 autistic kids during that time period, and 78 cases were all they found? That’s less than 0.1%. Speech after age 5 is still rare. But here’s the glass half-full part — not all cases get published. Speech after age 5 is possible. The duration of treatment matters, but there’s no connection with the child’s reported IQ (“speech acquisition was not limited to children with high IQs.”) Once they get sounds, they progress quickly to imitated words, then spontaneous words (“once they began speaking, subsequent improvements were often quite rapid.”) The concluding paragraph is cautiously optimistic: “intensive training in basic elements of speech and language might well ignite greater capabilities in some previously nonverbal individuals. The data we have reviewed provide some evidence that this may have already happened in at least some cases.”
Paparella, T., Stickles Goods, K., Freeman, S., & Kasari, C. (2011). The emergence of nonverbal joint attention and requesting skills in young children with autism. Journal of Communication Disorders, 44(6), 569-583. Typically developing children get a whole bunch of gestures by the time they are 12 months old — looking back and forth between a person and an object, reaching out for an item they want, giving an object to someone if they need help, pointing, etc. Pointing is the hardest one; they usually get it by 12 months, then take another 2 months to get to the point where they’re not just pointing to request for items, but pointing to share interests. In autistic children, all this gets delayed — but exactly which ones and by how much? Here researchers videotaped the type of gestures 39 autistic children and 21 typical children used. They grouped the children by their expressive language ability. They found a very interesting pattern: in autistic children, requesting gestures were intact; it was the social (joint attention) gestures that were delayed. The exception was pointing. Both requesting and social pointing were delayed in autistic children. The good news is that requesting reach is NOT delayed, so if you see a requesting reach, you can shape it into a point as a way of teaching pointing. Pointing stood out as a critical skill correlated closely to expressive language development, which is why I emphasize it in the “Improving Gestures” section of the guide.
Wodka, E.L., Mathy, P., & Kalb, L. (2013). Predictors of phrase and fluent speech in children with autism and severe language delay. Pediatrics, 131(4), 1128-1134. Let’s say your child is 4 years old, and can still only say single words. That’s considered a severe language delay, since children usually begin to combine words at the age of 18 months. What are the chances your child will overcome the delay, and be able to combine nouns with verbs, and possibly even create more complex sentences? Anderson’s study would suggest a poor prognosis — see the “minimal” (42%) versus “limited” (26%) trajectories, with children falling into the “minimal” category remaining at the 18 month expressive level by age 9. But this study suggests a more hopeful language outcome. They looked at 535 autistic children who were severely language delayed (no phrase speech by age 4), and discovered that at age 8, 70% of them had acquired phrase speech (2 to 3 words including a verb), and 47% of them had acquired fluent speech (complex sentences). Why the difference from Anderson? It may be that his sample — children referred to early intervention prior to age 2 — may have included more severely impacted children; or that parents in the current study overstated the extent of the language delay at age 4 (a retrospective reporting bias called “forward telescoping”). Regardless, the take home message is that language delays can be overcome! And certainly the chances of this are greater with intervention.
Kasari, C., Kaiser, A., Goods, K., Nietfeld, J., Mathy, P., Landa, R., Murphy, S., & Almirall, D. (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. This study extended JASPER to a hard group of children — 61 autistic children ages 5-8 who remained minimally verbal (<20 spoken words, with little to no phrase speech) even after early intervention and ~2 years of behavioral intervention. And it introduced one extra element: an AAC device. Would it help these children gain speech? Or would it hold them back? They did one hour sessions 2x/week over six months (in some cases where they child didn’t improve much, they increased to 3x/week halfway though). They measured before and after the number of times a child spontaneously communicated during a 20 minute play session. No prompting allowed! Just responsive play. The results were striking. With JASPER, it went from ~30 to ~40 communicative utterances. But with JASPER + AAC, it went from ~30 to ~60 communicative utterances — and 90% via natural speech, not through the device. AAC turbocharged JASPER, and the earlier in the study it was introduced, the better. This is the basis of the “An Intensive Push” section.
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. JASPER is great, but when most parents start AAC, they don’t want to focus on play. They want to focus on something very practical — finding out what their child wants. And that means teaching requesting. Under these circumstances, can speech improve? YES. In this study, 19 minimally verbal autistic children (~2.5 years old) were taught PECS over a period of 6 months (2 hour sessions, 5x/week). The average number of spoken words skyrocketed from 5 before treatment to 89 after treatment, with ~78% of children responding well (>10 spoken words). This is the basis of the “Find Out What Your Child Wants” section. Want to develop language, and get to phrase and fluent speech? That’s what the next sections are for.
Orr, E., & Geva, R. (2015). Symbolic play and language development. Infant Behavior and Development, 38, 147-161. Symbolic play and language development are closely related. In symbolic play, you’re substituting one object for another. In language, you’re substituting words for meanings. But what leads to what? In this study, the researchers followed 14 infants from 6-18 months and found that early symbolic play came first, followed by talking and language. This supports why symbolic play interventions like JASPER are so effective at improving language outcomes.
Pierce, K., Gazestani, V. H., Bacon, E., Barnes, C. C., Cha, D., Nalabolu, S., Lopez, L., Moore, A., Pence-Stophaeros, S., & Courchesne, E. (2019). Evaluation of the diagnostic stability of the early autism spectrum disorder phenotype in the general population starting at 12 months. JAMA Pediatrics, 173(6), 578–587. This is THE paper supporting early diagnosis of autism. It asked a simple question: if a toddler is diagnosed with autism at an early age, and reevaluated later, how often do they “keep” that diagnosis? They examined records of 2,241 toddlers who had been diagnosed between 12 to 36 months. They found that autism had a diagnostic “stability coefficient” that went from 0.5 at 12 months to 0.82 at 14 months. It then rose gradually in a gentle plateau for further months. Children who “lost their diagnosis” mostly were still found to have ASD features; only 1.8% became typically developing. The conclusion was that ASD could be reliably diagnosed at 14 months. I first read this paper when my son was 15 months old, and was showing signs of autism. I wrote to Pierce excited about the possibility of early diagnosis and intervention for my son, but said I was in Los Angeles. She acknowledged a rivalry between San Diego and Los Angeles, but acceded that UCLA might be able to provide help. I reached out to UCLA and they dryly noted that they had “some” expertise in autism diagnosis. My son received his ASD diagnosis from them a month later. It was very clear cut. Subsequent early intervention went so well that I wondered at his 3 y/o reevaluation whether he might lose the diagnosis. Nope! Still plenty autistic. I do not know what he would have been like without the early intervention, but I feel he would have been a quite different child. If you have suspicions of ASD, plenty of well-meaning people will tell you to wait and see how it plays out. Pediatricians. Family. Regional Center said they wouldn’t consider a diagnosis before 24 months. In many countries, doctors won’t diagnose before 36 months. This is outdated and burns precious time. I still remember a conversation I had with a home organizer I wanted to hire to reconfigure my living room to be more friendly for my son. She couldn’t bring herself to talk about the job. Instead she talked about her own autistic son, and how she wished she could have gotten him diagnosed much earlier. No parent ever regrets having their child evaluated early, and plenty regret having waited too long. If you’re on the fence about it, remember that the best time to plant a tree was 20 years ago. The second best time is now.
Strand, E. A. (2020). Dynamic Temporal and Tactile Cueing: A Treatment Strategy for Childhood Apraxia of Speech. American Journal of Speech-Language Pathology, 29(1), 30-48. This paper describes DTTC, a main treatment method for childhood apraxia of speech. The idea is to approach speech as a sequence of lip, tongue, and jaw movements that can be taught in much the same way as you might teach a child to tie their shoes, or put on a shirt. The therapist guides the child by saying the word together with the child; touching the child’s face; pointing at features on their own face; or by saying the word at super slow speed. Eventually the child is able to say the words properly on their own. It’s an intensive therapy (3-5x/week) requiring a lot of repetition and attention.
Rogers, S. J., Yoder, P., Estes, A., Warren, Z., McEachin, J., Munson, J., Rocha, M., Greenson, J., Wallace, L., Gardner, E., Dawson, G., Sugar, C. A., Hellemann, G., & Whelan, F. (2021). A multisite randomized controlled trial comparing the effects of intervention intensity and intervention style on outcomes for young children with autism. Journal of the American Academy of Child & Adolescent Psychiatry, 60(6), 710-722. Once I knew my son was autistic, I bought Dawson’s “An Early Start for Your Child” book, based on her Early Start Denver Model (ESDM) intervention. The strategies she provided in her book worked well, and gave me something to do while waiting for therapy to start. I wrote to her with questions about ABA therapy — what kind did she recommend? and how many hours per week? The companies I reached out to were all pushing 30-40 hours per week, which seemed insane. She replied right away saying they had just published a paper about this. The paper examined the outcomes of children who received two years of ESDM versus traditional ABA at two intensity levels (15 versus 25 hours per week). They found no difference in outcomes between ESDM versus traditional ABA, and no difference in outcome between 15 versus 25 hours per week. She emphasized that I shouldn’t be overly concerned with an intervention’s philosophy — instead I should be focusing on the quality of the delivery: “are the therapists well trained and supervised, do they track children’s progress and make adjustments when needed, are the therapists collaborative and incorporate the values and needs of the child and family, and so on.” This guidance proved invaluable. I was able to find companies willing to provide fewer hours of therapy, and also to push for well-trained and well-supervised therapists (there was — as she hinted — plenty of variability there). The JASPER model takes this to an extreme, delivering impressive outcomes with only 2-3 hrs/wk of therapy — but it does require highly trained therapists. When it comes to autism therapy, quality matters more than quantity.
Frolli, A., Ciotola, S., Esposito, C., Fraschetti, S., Ricci, M.C., Cerciello, F., Russo, M.G. (2022). AAC and autism: Manual signs and PECS, a comparison. Behavioral Sciences, 12(10), 359. For autistic children, learning either sign language or PECS often leads to talking. But which is a better bridge? The researchers worked with 82 nonverbal (no speech at all) autistic toddlers, ages 20–24 months. Half were taught sign language. Half were taught picture cards (PECS). The PECS group mastered 20 picture requests in 23 weeks, then took another 29 weeks before speaking those words (52 weeks total). The sign group took 39 weeks (longer!) to master the signs, but — amazingly — began speaking those same words at nearly the same time! The researchers hypothesize that the social imitation and motor effort needed to learn signs activated language centers in the brain so strongly that once the signs were learned, speech came along for the ride.
Chenausky, K. V., Norton, A. C., Tager-Flusberg, H., & Schlaug, G. (2022). Auditory-motor mapping training: Testing an intonation-based spoken language treatment for minimally verbal children with autism spectrum disorder. Annals of the New York Academy of Sciences, 1515(1), 266-275. Music helps talking? In this study, 14 minimally verbal (<20 words) autistic children between 5-10 years old were randomly assigned to one of two treatments. In the Speech Repetition Training group, children said words together with a therapist, then tried to repeat them back. In the Auditory Motor Mapping Training group, they did the same thing, but the words were sung and the child and therapist banged tuned drums together in synchrony. Four weeks after completing 25 sessions, the music and drumming group improved their syllable approximations by 12 percentage points compared to only 3 percentage points for the speech repetition group. The researchers hypothesized that singing and drumming together helped build neural connections between motor movements and sounds, aiding speech development.
National Institute on Deafness and Other Communication Disorders. (2023). Minimally Verbal/Non-Speaking Individuals With Autism: Research Directions for Interventions to Promote Language and Communication. Virtual Workshop, January 24-25, 2023. 25% of autistic children remain minimally verbal past the age of 5. They are often referred to as the “neglected end” of the spectrum, because they are often left out of research studies and interventions. It’s hard to assess someone who can’t talk. This workshop pulled together leading researchers working on studying these children, as well as parents, therapists, administrators, and other stakeholders. I watched all 8 hours of the recorded webinar. A key finding was that this population is incredibly heterogeneous — they can’t talk for different reasons, and there is no one therapy that works for everyone. There were some contentious discussions circling around ideas of limited cognition and language delays versus high cognition and motor impairments (apraxia), with the latter being more appealing to proponents of facilitated communication. Afterward, I reached out to the program manager, as well as everyone I could from the speaker list, making it clear I wanted to help. Many good conversations came out of this outreach.
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. Binger has been advocating the primacy of language teaching in AAC intervention for the last 20 years. Her approach has been to use activity boards. The primacy of language teaching to overcome language delays has also been a consistent thread through this guide. Three thoughts come into my head when I read this paper, which she sent to me after our conversation: (1) fields have fads (2) sometimes the dog wags the tail, and sometimes the tail wags the dog (3) when change happens, it tends to happen slowly, than all at once. The paper mentions other approaches which have been popular. Noun-first approaches became popular with PECS and its initial focus on requesting. Core-first approaches became popular with the advent of AAC apps, and their need to provide quick navigation to a common set of words. Neither approach is great for language learning — the noun first approaches lack verbs, and the core first approaches lack nouns! Activity boards provide the easiest access to verbs and nouns and are excellent for language learning, but prominent AAC app manufacturers will post blog posts arguing against them (“may not support core word learning and motor planning ...”) Sometimes the tail wags the dog! I feel that the field may change quickly, as more people realize that for autistic children, AAC can be an effective teaching tool to transition to speech or speech-equivalents (typing, sign language) rather than an end in itself. For that application, language-first approaches are critical.
Pope, L., Light, J., & Laubscher, E. (2025). The effect of naturalistic developmental behavioral interventions and aided AAC on the language development of children on the autism spectrum with minimal speech: A systematic review and meta-analysis. Journal of Autism and Developmental Disorders, 55, 3078–3099. Toward the end of the guide, I complained that AAC specialists spent too much time focused on grid sizes and core versus fringe folders, and too little time focused on gestures, play, and language. Janice Light has overseen THE center (Rehabilitation Engineering Research Center on Augmentative and Alternative Communication at Penn State) to develop new AAC technologies over the last decades. This review highlights Kasari’s work in 2014 pointing out that combining AAC with JAPSER (play + gesture intervention) powerfully improved language and speech in minimally verbal autistic children. The review also finds 28 other studies supporting this conclusion: “both aggregate and comparative effect sizes were notably larger when AAC was included within NDBI [Naturalistic Developmental Behavioral Intervention] procedures, as compared to NDBIs without AAC. Results suggest that combining AAC with NDBI procedures may lead to better language outcomes than NDBIs alone for children on the autism spectrum with minimal speech.” Perhaps — to use business jargon — there will be some breaking down of traditional silos, and collaboration across disciplines. The review itself says, please, work with us: “The small representation of AAC research within this NDBI review highlights the substantial disconnect between these two areas of study, both of which are focused on promoting the language development of children on the autism spectrum ... underscoring the critical importance of bridging the gap between these two areas of research and practice.” I am optimistic. Sometimes when change happens, it happens slowly, then all at once.