Speech Motor Chaining: A Structured Approach to Articulation Therapy
How this evidence-based method helps children move from accurate practice to confident, automatic speech
How this evidence-based method helps children move from accurate practice to confident, automatic speech
There's a frustrating pattern that SLPs know all too well: a child produces their target sound beautifully in the therapy room (clear /r/s, crisp /s/s, perfect placement) and then walks out the door and it's like the session never happened. The sound disappears into everyday conversation.
This gap between what a child can do with support and what they actually do independently is one of the central challenges of articulation therapy. It's the difference between skill acquisition (those in-session wins) and skill learning (the lasting changes that stick).
Speech Motor Chaining (SMC) is an approach specifically designed to bridge that gap. In speaking with speech-language pathologists, we've heard how useful motor chaining is for helping children not just produce sounds correctly, but retain and generalize them. Here's how it works, what the research shows, and how to apply it.
Speech Motor Chaining was developed by Jonathan Preston and colleagues at Syracuse University over a series of studies, and their tutorial lays out the full procedure (Preston, Leece, and Storto, 2019). ASHA’s practice portal on speech sound disorders lists it among treatment approaches, describing it as based on motor learning theories and focused on repeated motor practice of speech targets.
The core idea is that speech is fundamentally a motor skill. When a child learns to produce a sound correctly, they're not just learning where to put their tongue. They're building a movement pattern that needs to become automatic enough to run in the background of conversation. SMC structures practice to make that happen.
SMC explicitly addresses two different stages:
Some children show acquisition (better accuracy during sessions) without learning (poor generalization or retention). SMC starts with conditions that favor acquisition and then shifts, step by step, toward conditions that favor learning.
SMC moves children through progressively more complex speech targets within what's called a "chain." Each chain is built around one short sound sequence containing the target, called a variant (for example, the "ar" in car), and that sequence stays the same all the way up the chain. The five levels are:
The basic sound sequence: at least one consonant and one vowel, such as consonant plus vowel, vowel plus consonant, or a small cluster. For our example, "ar."
A one-syllable word that begins or ends with the sequence and has both a starting and an ending consonant: "car."
A word of two or more syllables that contains the one-syllable word. Building from "car," this might be "cartoon" or "carpool."
Short phrases of 2 to 5 words containing the target word: "watch a cartoon." This introduces connected speech while keeping the load manageable.
The child makes up a new sentence using the target word. Generating language takes more planning than repeating, which pushes the motor skill toward automaticity.
Targets are practiced in blocks of six attempts. To move to the next level, a child needs at least 5 out of 6 correct productions in a block. If they don't get there, they switch to a chain for a different variant of the same sound and start again at the syllable level, rather than drilling the same context repeatedly. A chain is considered mastered when a child completes all five levels successfully on two separate occasions.
If you know the traditional articulation hierarchy, which moves a child up one level at a time and often only after reaching a set accuracy within a session, motor chaining compresses that climb: the task can change after every block of six, so a child who is succeeding can go from syllables to sentences in about two minutes (Preston, Leece, and Storto, 2019). A child who is struggling can stay at the syllable level for the whole session.
One of SMC's distinguishing features is its structured approach to feedback. Motor learning principles suggest that the type and frequency of feedback affect whether a skill is only acquired in the session or actually learned (Maas et al., 2008). SMC uses three kinds:
Specific feedback about how the child produced the sound: the movement itself. "You kept your tongue pulled back that time" or "Your tongue tip touched your teeth."
Simple feedback about whether the sound was accurate or not. "That was correct" or "Not quite that time." This is feedback about the outcome, not the process.
The child rates their own production before hearing the clinician's feedback, on three of the six attempts in each block. "How did that one sound to you?"
This is where the motor learning science gets practical. At the syllable level, feedback is frequent and specific, with more KP to help the child understand what a correct production feels like. As the child moves up to phrases and sentences, feedback becomes less frequent and less specific.
In a six-attempt block at the syllable level, the tutorial calls for KP plus KR on four attempts, KR only on one, and no feedback on one. At the sentence level this flips: no feedback on three attempts, KR only on two, and KP plus KR on just one, giving the child space to rely on their own monitoring.
This fading is intentional. Frequent, detailed feedback is expected to help acquisition, while less frequent, less specific feedback is thought to help longer-term learning, so the learner builds their own error detection instead of depending on the clinician's ear.
SMC was designed for school-age children with speech sound disorders, and the developers report using it with children from age 6 through adults. It has been used with:
Suitable targets are typically sounds the child produces correctly less than half the time at the word level. The developers also note that many 3 and 4 year olds, and older children with significant intellectual disabilities, may not do well with a highly structured drill, though that has not been tested.
The SMC tutorial pulls together six earlier studies of 39 children and adolescents (ages 8 to 21) treated for residual speech errors, mostly on /r/ (Preston, Leece, and Storto, 2019):
Those studies varied: some children had apraxia of speech, treatment ran 14 to 16 hours, and the amount of visual biofeedback varied from roughly 20 to 45 percent of the session. The developers are candid that nearly every child made gains within sessions, but not every child generalized to untrained words, and some may need more hours or a different approach.
Across 36 participants in six studies, sessions averaged 218 practice trials, with 40 to 52 minutes of each session spent on speech practice. Clinicians who trained one-on-one for about three hours delivered the prescribed feedback on more than 80% of trials, and on more than 90% for all but one of the seven clinicians. Two of those seven were graduate students, which suggests the feedback structure can be learned and followed as designed.
Newer trials have looked at scheduling. In a randomized controlled trial of 48 children aged 9 to 17 with apraxia of speech, the same 20 hours of SMC led to more generalization when delivered intensively over five weeks than when spread over ten, and ultrasound biofeedback sped up early learning, although that advantage faded (Preston et al., 2024). And in a small pilot trial with children aged 7 to 9, ten-minute SMC sessions four times a week improved /r/ in syllables over eight weeks, though gains in untrained words were not yet clear (Herbst et al., 2025).
Before chained practice begins, the child needs to produce each variant correctly with support. Use modeling, shaping, and visual and verbal cues for placement until the child has at least three correct productions of each variant you'll be training, with the whole sound sequence correct, not just the target sound. For a child who is not yet stimulable, this can take a whole session: in the tutorial's case example, a 10 year old spent his first session in pre-practice before starting chains.
SMC typically uses four variants for each target sound. Choose variants that:
If a child gets stuck at one level in a chain, they move to a different variant rather than grinding through repeated failures. This keeps motivation high and prevents practicing the error.
SMC sessions are intensive. Most children in the research reached more than 200 trials per session, and that high dose is part of the design. Motor skills require repetition, and speech is no exception. Structure sessions to maximize practice opportunities while keeping the child engaged.
Performance during practice shows acquisition, not learning, so the developers use probes of untrained words as the main measure. They probed about every two sessions in research and suggest every three to four sessions clinically, at the start of a session. When word-level probe accuracy reaches about 80%, they switch to probes of 10 to 15 untrained sentences, and accuracy above 80% at the sentence level may be enough to move on to conversational work.
The principles of motor chaining can extend beyond the therapy room. When assigning home practice, consider these guidelines:
The challenge of articulation therapy has never been teaching children to produce sounds correctly. Most children can do that with enough cueing and support. The challenge is getting those sounds to stick: to become automatic, to generalize to new words and contexts, to show up in the rush of everyday conversation.
Speech Motor Chaining offers a systematic path from correct production to lasting skill. By structuring practice through progressive complexity levels, strategically fading feedback, and building self-monitoring, SMC addresses the full trajectory of motor learning. It's not magic. It's applied motor science.
For children who've been stuck in the gap between "can do in therapy" and "does in real life," motor chaining provides a structured bridge to the other side.
Speech Motor Chaining (SMC) is a structured, motor-based articulation treatment described by Jonathan Preston and colleagues at Syracuse University. Practice is built into chains that move through five levels (syllables, monosyllabic words, multisyllabic words, phrases, and self-generated sentences), and the type and amount of feedback change as the child succeeds, to support both accuracy in the session and generalization to new words.
It was designed for school-age children with speech sound disorders, and its developers report using it from age 6 through adulthood, including with residual speech errors and childhood apraxia of speech. Suitable targets are typically sounds the child says correctly less than half the time at the word level. A pre-practice phase comes first, to get at least three correct productions of each target sequence before chained practice begins.
A traditional hierarchy often asks for a set accuracy within a session before moving from syllables to words. SMC changes the task after every block of six attempts, so a child who is succeeding can move from syllables to sentences in about two minutes. It also fades feedback on purpose, from detailed feedback about the movement toward simple right or wrong feedback and self-rating.
It varies. In one study summarized by the developers, six children with residual /r/ errors improved by about 33 percent on untrained words after seven one-hour sessions. The developers also note that many but not all children generalize, and that some need more hours. A chain counts as mastered when the child completes all five levels on two separate occasions.
Isolation to conversation, step by step: what each level looks like, common criteria for moving up, a cueing hierarchy for fading support, home practice at each level, and a free printable chart.
Why /r/ is so challenging, evidence-based treatment approaches, and home practice activities to help children master this tricky sound.
A guide to childhood apraxia of speech (CAS) for parents and SLPs. The signs, how CAS is diagnosed, evidence-based treatments, and how to support practice at home.