Early Wondering Minds

Autism and Early Literacy Instruction Adaptations

Different autism profiles need different reading instruction, not one-size-fits-all methods.

· 9 min read
Cover illustration for “Autism and Early Literacy Instruction Adaptations”
Learning Differences · October 23, 2026 · 9 min read · 1,914 words

Early literacy instruction for autistic children needs to fit the child’s reading profile. Some children can read words but struggle to understand them. Others need help with both. Effective adaptations preserve systematic teaching while changing how children participate, show what they know, and learn to connect print with meaning.

Why autistic children do not share a single reading profile

Parents and teachers may expect autism to produce a predictable reading difficulty. The research shows no such single pattern. Some autistic children decode words fluently while struggling profoundly with meaning, a profile sometimes called hyperlexia in informal usage. Others face barriers in both decoding and comprehension. That difference changes what instruction needs to do, because a lesson aimed at reading words cannot, by itself, solve a problem with understanding them.

Most structured literacy programs are built to help children who struggle with decoding. That makes them useful for one part of the picture, but an incomplete answer for children whose word reading is already strong. Fluent decoding can hide how much help a child needs with comprehension, just as trouble speaking can hide what a child knows about reading. Neither performance should stand in for a full assessment. Instruction needs to distinguish what a child can read from what the child understands.

A concurrent intellectual disability, present in roughly a third of children with ASD, can complicate both learning and assessment. The literacy needs of minimally verbal children, who use very few or no functional spoken words, are especially likely to be overlooked. Their speech makes the way adults assess their instructional needs more important.

How underestimating competence compounds the instructional problem

A reading assessment that demands spoken answers measures more than reading. For a nonspeaking child, it also measures whether the child can use a response format that is inaccessible. The score can show less understanding than the child has, yet still guide how much teaching the child gets. Low scores then become a reason to put less into teaching. An assessment flaw becomes an instructional barrier.

Yael Kimhi (2025) describes research in which nonspeaking children who could not score on a traditional word recognition test performed near the top of the scoring range when an adapted digital version allowed touchscreen responses. The response format changed what children could demonstrate. That finding undermines the assumption that failure to answer aloud proves failure to recognize words. The test must reach the learner before its results can guide teaching.

Kimhi identifies the belief that speech must come before reading as a persistent reason nonspeaking learners are excluded from literacy curricula. Leaving these children out makes the original measurement problem worse because they lose the chance to learn more skills through teaching. Kimhi sees literacy as a basic right that helps people make their own choices, take part in society, and communicate on their own. The rights issue follows directly from the method. A tool that cannot register competence should not become a gatekeeper to learning.

What the Science of Reading offers autistic learners — and where it stops short

Systematic phonics is the right foundation for autistic learners, including children who use AAC. That foundation includes phonemic awareness, letter-sound correspondences, decoding, and encoding. These are not skills reserved for children who speak. Audiovisual integration research shows that decoding and language comprehension jointly link audiovisual integration with reading comprehension across diagnostic groups. Autistic children do not need a separate route that bypasses those foundations.

Structured teaching also offers a practical advantage. Lessons that follow a predictable pattern and teach skills in a clear order make it easier to work out what unclear social and instructional cues mean. Children can engage with the reading task rather than having to work out what the adult expects. Applying the Science of Reading well therefore supports systematic instruction, but it does not require identical delivery for every learner.

The gap is in what comes with that foundation. Evidence-based comprehension supports designed specifically for nonspeaking and minimally verbal learners remain critically underdeveloped, even as phonics-first instruction has become the established standard. More phonics cannot fill every reading gap. A child who decodes well needs teaching directed at meaning, while a child who cannot answer aloud needs another way to participate. Neither requirement challenges systematic teaching. Both are conditions for making it useful.

Adaptations that make systematic instruction reachable for children who do not speak

Systematic phonics can reach minimally verbal learners when the response format does not require speech and the materials support nonverbal participation. Touchscreen selection, AAC symbol choice, and visual-spatial matching can remove the oral-response bottleneck without changing the phonics target. The skill stays. The route to answering changes. This distinction prevents an access adaptation from being mistaken for easier instruction.

Savaldi-Harussi and Robinson-Harshoshanim (2026) used adapted Hebrew vowel-based instruction with nonverbal response options for minimally verbal adolescents with autism and intellectual disability. Two of three participants made significant gains in vowel identification and CV syllable recognition, and teachers found they could use the approach and wanted to keep using it. Word-level gains were limited, however. The findings support accessible phonics teaching while pointing to the need for contextual and semantic support beyond decoding.

Caron et al. (2026) used the ALL app with three minimally verbal or nonspeaking children who used high-tech AAC, whose parents were trained to teach them. All three progressed in letter-sound correspondences, sound blending, and sight words, with large-to-very-large effect sizes. Sound blending also generalized beyond the training context. AAC served as a literacy scaffold, allowing responses while supporting connections among symbols and sounds.

Visual modeling offers another way into decoding. A 2025 study examined augmented reality that placed video models of decoding over printed word sheets for first-grade students with developmental delays and autism. More broadly, multimodal instruction deliberately engages visual, auditory, kinesthetic, and tactile channels.

Why parent delivery at home is becoming a viable instructional channel

Teaching at school alone cannot give autistic children all the repetition and consistency they usually need. Trained caregivers can extend that instruction into the home, where children spend much of their early years. The Caron study used virtual parent training and remote progress monitoring, showing a model that did not require a specialist to attend every session in person.

Veronica Fleury’s FSU / IES-funded project (2024–2028), funded by the National Center for Special Education Research, is developing adaptive shared reading for preschoolers with ASD. It is designed specifically for caregiver delivery at home. Dialogic reading and print referencing are first-stage options, with added support planned for children who do not respond to the initial approach. The project is development and testing work.

Home delivery also depends on conditions that a program cannot assume are already in place. Caregivers need brief, clear training, and books and materials need to be selected for accessibility. How children feel matters too, because they often resist reading after repeated failure or embarrassment. Reading to a pet, choosing between two books, and keeping sessions short and consistent can lower the stakes. These adjustments support participation while preserving learning.

What comprehension instruction for autistic learners actually requires

Comprehension needs its own strand of instruction from the start. For autistic children who read words well but understand little, explicit vocabulary teaching is a direct way to address the gap. When children have fewer chances for incidental language learning, their grasp of word meanings can lag behind their word recognition. Reading instruction should therefore include repeated vocabulary teaching tied to the text from the start. Recognizing a word is only part of knowing it.

Understanding a story also involves connections the page does not spell out. A character’s action may imply an unstated feeling, or an event may depend on something described on a previous page. Those links can require explicit teaching for autistic learners rather than an expectation that they will emerge unaided. Shared reading provides a way to teach them. In dialogic reading, the adult asks open questions, builds on the child’s responses, and connects the text to meaning.

The Fleury project is designed in part to test this approach at preschool level, where the evidence is still emerging. Its rationale fits a broader, multicomponent model that combines phonics, whole-word recognition, vocabulary, and comprehension. How much of each to teach depends on the child’s profile. Caron and colleagues also stress teaching foundational and related literacy skills at the same time, with adaptations and meaningful shared experiences that help children want to learn and take part.

What technology can and cannot do in this instructional picture

Technology earns a place when it removes a specific barrier. Text-to-speech can provide access, adaptive practice can target a child’s current skill level, and AAC-compatible response formats can allow nonverbal participation. Repeatable digital practice can also reduce dependence on an adult’s ability to deliver the same task patiently, while adapted materials can reduce the work of retrofitting standard ones. But a useful function is not proof of effectiveness. Some tools have clinical trial support, while most apps used with autistic children have not been tested.

AI-driven adaptive instruction has promise because a fixed program cannot fully respond to this population’s varied strengths and gaps. A system that tracks what a child can and cannot do, then adjusts teaching moment by moment, fits the need for individualization. Instructional responsiveness makes the difference. An evolving model of the learner offers something different from a digitized worksheet or a repeating quiz.

That promise still depends on access. A system that listens to a child’s voice in real time may offer useful adaptation for children who can use that response channel, but voice dependence does not solve access for nonspeaking learners. Human expertise also remains essential. AI cannot replace the relationship that motivates many young autistic children, observe the full behavioral and sensory context a skilled teacher reads, or make the complex judgments about AAC integration that a speech-language pathologist brings.

What parents and teachers should actually look for when choosing or designing instruction

Parents and teachers should first ask which skills this child needs to develop and which response formats make those skills visible. For nonspeaking or minimally verbal children, parents and teachers should press for an adapted assessment before accepting a claim that the child has low ability. Oral performance alone is not enough evidence. The assessment should guide the choice of instruction.

A suitable approach should have systematic, explicit phonics as its foundation, with response options that do not require speech where needed. It should teach decoding and comprehension together, with explicit vocabulary woven throughout instruction. Multimodal delivery should have a clear instructional purpose. The approach should also change as the child responds, rather than continue through a fixed script regardless of progress. A strong starting design is not enough if it cannot adapt.

Caregiver involvement should include actual training and useful progress information. Parents need to know which skills are developing, which are stalling, and what instruction attempted in each session. That visibility makes home reinforcement more focused and helps families flag an unproductive approach before weeks are lost. Short, consistent daily practice is preferable to long, infrequent sessions. For a child who resists reading, early sessions may first need to weaken the link between reading and failure before pursuing mastery.

These criteria give families something concrete to ask for. Individualization is not a luxury or a special-education niche, but the ordinary requirement that teaching respond to the learner. Accessible response formats, caregiver training, and responsive tools make that requirement more achievable for families who know what to look for.

Sources

  1. Introduction
  2. Effectiveness of adapted vowel-based literacy instruction in minimally verbal adolescents with autism and intellectual and developmental disability - ScienceDirect
  3. Frontiers
  4. Improving Early Literacy Skills for Autistic Children Who Use High-Tech AAC: Effects of a Parent-Delivered Intervention - PubMed
  5. Improving Early Literacy Skills for Autistic Children Who Use High-Tech AAC: Effects of a Parent-Delivered Intervention
  6. Developing an Emergent Literacy Adaptive Intervention for Preschoolers with Autism
  7. TYPE Perspective PUBLISHED 11 December 2025 DOI 10.3389/feduc.2025.1722796

More in Learning Differences