Quantifying unintelligible speech

Evaluations for children with speech sound disorders (SSDs) are pretty standard across the board. We take a speech sample through imitated words, picture naming, and/or free conversation. We measure which consonants they produce correctly vs. incorrectly, do an oral mech exam, and call it a day. But when children have long stretches of unintelligible speech, things can get complicated. How can we analyze speech if we have no idea what the child is saying? Öster et al. attempts to help us quantify unintelligible speech using a measure called percentage of intelligible correct syllables (PICS).  


SLPs measured intelligible correct syllables by listening to a short sample of connected speech in children ages 4–10 with SSD (approximately 100 words). Since syllables are still evident in unintelligible speech, the SLPs were able to measure the number of accurate syllables divided by the total number of syllables in the sample to obtain the percentage of correct syllables. Furthermore, children’s’ PICS scores were correlated with their percentage of correct consonants (PCC) showing us that PICS can be used as a valid measure for assessing speech sounds. For those tricky children who are unintelligible more often than not, PICS could provide a baseline from which to measure progress.  

These researchers also used a novel method for collecting a connected speech sample. Using concurrent commenting, children watched a silent video, typically a clip from a children’s TV episode, and were instructed to describe the ongoing events of the video to a listener. This procedure allows SLPs to gain a sample of narration, which usually contains more linguistically complex and longer utterances compared to conversational speech. The video also provides contextual clues for the SLP to use when transcribing the sample.

Ready to jump in? Check out the article for even more details on how to assess a child using PICS and concurrent commenting. Both are handy tools to consider using when assessing highly unintelligible children.


Öster, C.A.M., Ode, C., & Strömbergsson, S. (2019). Dealing with the unknown—addressing challenges in evaluating unintelligible speech. Clinical Linguistics and Phonetics. doi:10.1080/02699206.2019.1622787.

Speech homework: The parents’ perspective

May Quotes, website.png

If you are an SLP who works with children with speech sound disorders (SSD), you are aware of: (1) how important home practice is, and (2) how difficult it is to ensure it happens. You send home cute activities and worksheets only to find out they’re never being done. You email parents with suggestions but wonder how much parents actually use them.

Some researchers decided to interview parents about their experiences with home practice in order to better understand their perspective. They wanted to hear from parents how SLPs could better support and encourage their attempts!

The researchers interviewed six parents of children aged 3–6 who had participated in speech–language therapy for an SSD. Several themes emerged throughout the conversations. They boiled them down these: 

Evolution over time

Parents expressed that their experiences with home practice changed over time. At the beginning, parents often felt confused, overwhelmed, and unsure of how to complete the activities. Over time, parents felt increasingly confident in the activities and what was expected of them.

Different roles

Parents saw the SLP as the expert who could provide materials and instruction, but saw themselves as ultimately responsible for supporting their child’s speech and language.


Parents stressed the importance of several things to them: their child’s speech and communication growth, their own role and involvement in therapy, consistent home practice, and rapport with the SLP. They felt that when there was rapport, they and their child were more motivated to do home practice and they saw more progress.

Managing the practicalities of home practice

Parents expressed difficulty with the logistics of home practice. All parents reported that it was challenging to find the time to do the activities and most admitted that they did not complete the full amount of time suggested by the SLP. Parents described receiving activities that were not motivating to their child or did not suit them as a family. They were often also unsure of how to complete the activities or how to do the technical components of therapy.

Taken together, these findings leave us SLPs with some helpful takeaways. First, it’s important for parents and SLPs to have a clear discussion about both of their expectations at the beginning of therapy. These parents spoke about how their expectations did not always match the realities of what therapy looked like and it took a while for them to adjust and figure out the ropes. Second, it may be beneficial to regularly share data about the child’s progress, as parents found that when their child made progress, they were more motivated to continue home practice. Last, SLPs should work with the family to provide family-appropriate materials as well as sufficient training and clear instructions for how to complete them.


Sugden, E., Munro, N., Trivette, C.M., Baker, E., Williams, A.L. (2019) Parents’ experiences of completing home practice for speech sound disorders. Journal of Early Intervention. doi: 10.1177/1053815119828409.

Ultrasound visual biofeedback: a new tool for treating SSD?

Have a student (...or 50?) on your caseload with articulation goals? Have some of them been there for a long time? While some of our kids with speech sound disorders (SSD) make quick progress with traditional methods, what can we try with those whose errors persist despite treatment? One new—and cool— option is ultrasound visual biofeedback (u-VBF).

“Wait, what? Ultrasound??”

Yes, it’s actually been an active area of research for both assessment and intervention for many years, with positive or mixed outcomes reported. Ultrasound is non-invasive and uses high-frequency sound waves to make an image of the tongue, which clients can see as they speak, in real-time. This technology is most helpful for place errors involving the tongue, since labial sounds can be seen using a mirror. And while it sounds super-expensive, ultrasound units have actually become so affordable that we’re starting to see clinics and even school districts jump on board!

This newest study from Cleland et al. used u-VBF to treat a diverse range of children who presented with a range of SSDs, comorbid conditions like DLD or ASD, and different speech targets. (Note: /r/ was excluded as a possible target in this particular study.) The authors conducted a thorough assessment for each child to determine errors and stimulability, which guided the target selection for each child (check out the article for a cool flow chart illustrating this). After an initial baseline phase, 10–12 treatment sessions were conducted, following procedures you can find in this resource manual from the same authors.


Most children made some improvement toward their speech sound, although those without a co-occurring diagnosis made the most improvements, which the authors took to mean that children with comorbidities might benefit from higher dosage. These results show us that ultrasound can be used to treat lingual articulation errors in a wide range of SSDs. Given the design and sample size, we can’t say much about treatment of specific targets or client subtypes at this point. And since this paper didn’t address the dreaded /r/, see here, here, and here for more on that.

Cleland, J., Scobbie, J. M., Roxburgh, Z., Heyde, C., & Wrench, A. (2019). Enabling new articulatory gestures in children with persistent speech sound disorders using ultrasound visual biofeedback. Journal of Speech, Language, and Hearing Research. doi:10.1044/2018_JSLHR-S-17-0360

On intelligibility: why use it, and options for measurement

It’s been suggested that we use intelligibility as part of comprehensive speech assessment, and measurement of treatment outcomes. Why? Well, because intelligibility is kind of the point of speech therapy in the first place, right? Also, intelligibility can pick up on phonological changes that other measures (like percent consonants correct, PCC) can’t.

So which intelligibility measures are we supposed to be using, exactly? Or, more appropriately—what are our options?

First, there are many ways to measure intelligibility. We can use rating scales, single word measures, or connected speech; and raters may include the clinician, family, peers, or unfamiliar listeners. Each of these have their own pros and cons in terms of reliability, validity, and compatibility with clinical practice. But the gold standard has been to calculate the percent of words understood, by unfamiliar listeners, in a connected speech sample (Gordan-Brannan & Hodson, 2010).

And while speech samples + few unfamiliar listeners as raters may be ideal, that carries a time burden for clinicians. Further, we also really want data on how the child is functioning in his or her everyday life. These considerations are what make the Intelligibility in Context Scale (ICS) particularly enticing—developed and measured over the past several years from (ongoing) research by McLeod and colleagues.

What is the ICS? It’s a brief, 7-item rating scale, completed by the parents of preschool and school-aged children. It can supplement other clinical measures for a nice look at functional speech. The scale can be found here (also on the last page of this article). Additional things to know about it:


Multilingual populations: It’s been translated into 60 languages (free, online!), and being multilingual doesn’t affect the scores (McLeod et al., 2015). It’s recommended that you use a separate sheet for each language the child speaks.

Screening: Use for preschool screening would be appropriate, especially as additional normative data is collected by future research. For now, this article can help you identify appropriate scores for your environment. You’ll extrapolate to your clinical population by looking at the scores they found in their sample of 4- and 5-year-olds (see Table 3 of the study). Do keep in mind the limitations of their study (read Limitations section). But basically, the scores in this study are relatively conservative, so children are generally likely to require further speech evaluation if their scores are lower than that 2015 study.

Psychometric properties: You can find this data throughout several of their articles; in particular, this one, which provides support for the ICS as a valid and reliable measurement of preschool children’s intelligibility.

Though we started by looking at the McLeod et al., 2015 paper, research for this Throwback review ended up sending us toward several papers on intelligibility, linked out above. Enjoy!

And more

Hwa-Froelich & Matsuo found that children who were adopted internationally had pragmatic skills within the average range, yet their scores were lower than their non-adopted, typically-developing peers. Understanding the language skill profiles of children adopted internationally is important so that we don’t over-refer or misdiagnose these kids. P.S. We’ve reviewed this team’s research with this same population before here.

In the largest study of its kind to date, Potter, Nievergelt, & VanDam found that children with speech sound disorders have similar tongue strength as their typically-developing peers. This study adds to the evidence base that disputes the use of non-speech oral motor exercises in speech therapy.

Rivera Pérez et al. wondered whether monolingual SLPs could use audio prompting (i.e., pre-recorded stimuli in the home language) to facilitate vocabulary learning in Spanish–English bilingual preschoolers with typical language abilities. Children were taught vocabulary in either English only or in both English and audio prompt-delivered Spanish. All children learned English vocabulary, and only the group receiving audio prompting improved on Spanish vocabulary measures, suggesting audio prompting may help improvement in the home language. We should note that their design didn’t compare the English-only and English-plus-audio-prompting conditions and participants were typically developing children taught by SLPs. Still, more research like this could help identify ways SLPs can better serve their bilingual students. Exciting! 

Roberts et al. found positive effects of teaching preschoolers (including some dual language learners) letter name and letter–sound correspondence. No surprise there—we know how important that skill is! It is interesting that they found no advantage for teaching letter names before letter sounds: the jury’s still out on whether one should be taught before the other.

A study by Sue et al. reminds us to consider generalization not only across contexts but across receptive–expressive language modalities. In a single case design on vocabulary training in children with ASD, where children were taught a set of words either receptively or expressively, they found that some but not all of those words taught were acquired in the untrained modality. More expressive-to-receptive transfer was noted—which makes a lot of sense. There are still open questions about the optimal teaching order (if there is one) and what the implications are for dosage.


Hwa-Froelich, D. A., & Matsuo, H. (2019). Pragmatic language performance of children adopted internationally. American Journal of Speech–Language Pathology. doi:10.1044/2018_AJSLP-18-0075

Potter, N. L., Nievergelt, Y., & VanDam, M. (2019). Tongue strength in children with and without speech sound disorders. American Journal of Speech–Language Pathology. doi:10.1044/2018_AJSLP-18-0023

Rivera Pérez, J. F., Creaghead, N. A., Washington, K., Guo, Y., Raisor-Becker, L., & Combs, S. (2019). Using Audio Prompting to Assist Monolingual Speech–Language Pathologists to Teach English–Spanish Vocabulary to English Learners. Communication Disorders Quarterly. doi:10.1177/2F1525740118819659

Roberts, T. A., Vadasy, P. F., & Sanders, E. A. (2019). Preschoolers’ alphabet learning: Cognitive, teaching sequence, and English proficiency influences. Reading Research Quarterly. doi:10.1002/rrq.242

Su, P. L., Castle, G., & Camarata, S. (2019). Cross-modal generalization of receptive and expressive vocabulary in children with autism spectrum disorder. Autism & Developmental Language Impairments. doi:10.1177/2F2396941518824495

And more...

Baylis and Shriberg found that 14 of 17 children (82.4%) with 22q11.2 deletion syndrome (aka DiGeorge syndrome and velocardiofacial syndrome) had comorbid motor speech disorders. Speech motor delay and childhood dysarthria were more common than CAS. These initial prevalence estimates add to a growing body of evidence that helps us better understand the profile of 22q syndrome.

Glover et al. found that young children (preschool through 3rd grade) had more negative attitudes toward stuttering than their parents. By 5th grade, those attitudes improved and were similar to attitudes of parents.

Hammarström et al. found that an intense treatment (4 sessions per week for 6 weeks) was effective for a 4 year old, Swedish-speaking child with a severe speech sound disorder. Treatment incorporated multiple approaches—integral stimulation, nonlinear phonology, and a core vocabulary approach. After therapy, the child produced more target words, word shapes, and consonants correctly.

Kraft et al. replicated an earlier study to find that effortful control (an aspect of temperament) was the most important factor predicting stuttering severity in children. They recommend addressing self-regulation as part of the holistic treatment of stuttering.

Lancaster and Camarata set out to explain the heterogeneity of language skills in kids with DLD. At this time, it’s looking like a spectrum model (think autism!) fits best, versus labeling kids by subtypes or chalking up the differences to unique, individual profiles; but lots more data is needed. For now, the evidence suggests we should assess and treat kids with DLD based on level of severity *and* individual needs—which is probably what you’re doing already. 

Lane et al. profiled the communication skills of children with Sotos Syndrome using a parent-report measure. They found that most of the children had a language impairment (with issues in both structure and pragmatics), with a relative strength in verbal vs. nonverbal communication and a weakness in using context. These children are likely to need support in peer relationships, too. 

Sutherland et al. found that a standardized language test (the CELF-4) can be reliably administered via telehealth to children with autism. The specific children they tested were between 9 and 12 years old and mostly mainstreamed.


Baylis, A. L., & Shriberg, L. D. (2018). Estimates of the prevalence of speech and motor speech disorders in youth with 22q11.2 deletion syndrome. American Journal of Speech-Language Pathology. doi:10.1044/2018_AJSLP-18-0037

Glover, H. L., St Louis, K. O., & Weidner, M. E. (2018). Comparing stuttering attitudes of preschool through 5th grade children and their parents in a predominately rural Appalachian sample. Journal of Fluency Disorders. Advance online publication. doi:10.1016/j.jfludis.2018.11.001

Hammarström, I. L., Svensson, R., & Myrberg, K. (2018). A shift of treatment approach in speech language pathology services for children with speech sound disorders – a single case study of an intense intervention based on non-linear phonology and motor-learning principles. Clinical Linguistics & Phonetics. Advance online publication. doi:10.1080/02699206.2018.1552990 

Kraft, S. J., Lowther, E., & Beilby, J. (2018). The Role of Effortful Control in Stuttering Severity in Children: Replication Study. American Journal of Speech-Language Pathology. Advance online publication. doi: 10.1044/2018_AJSLP-17-0097

Lancaster, H. S., & Camarata, S. (2018). Reconceptualizing developmental language disorder as a spectrum disorder: Issues and evidence. International Journal of Language and Communication Disorders. Advance online publication. doi:10.1111/1460-6984.12433

Lane, C., Van Herwegen, J., & Freeth, M. (2018). Parent-Reported Communication Abilities of Children with Sotos Syndrome: Evidence from the Children’s Communication Checklist-2. Journal of Autism and Developmental Disorders. Advance online publication. doi:10.1007/s10803-018-3842-0

Sutherland, R., Trembath, D., Hodge, M. A., Rose, V., & Roberts, J. (2018). Telehealth and autism: Are telehealth language assessments reliable and feasible for children with autism? International Journal of Language & Communication Disorders. Advance online publication. doi:10.1111/1460-6984.12440

Random, reading, or rendof: What type of target words maximize treatment gains?

Ah, the age-old question—what words should we use in therapy for kids with speech sound disorders? There are a number of choices, each with some good arguments in its favor:


This study checked out differences in treatment success for each of these three word types for 24 children ages 3 to 7. Therapy for each kid focused on one, word-initial complex phoneme (/r/, /l/, “th” or “ch”), with five target words (either high-frequency, academic vocab, or nonwords, depending on which group the child was assigned to). The article describes the activities within each 50-minute intervention session, and supplemental treatment materials are available on a lab website (woo!). Each child’s progress was compared with his own performance in a baseline condition.

So, the winner? It’s the best possible news for clinicians, really. All the kids improved their phonological skills, with no significant differences among the words types. The authors point out that in reality, you’ll probably want to incorporate multiple types of words into your sessions. Like starting with nonwords for that “clean slate” effect, then moving to academic vocabulary words after a while to help boost those skills. But either way, the initial familiarity of the words likely won’t make or break your therapy. 

A second research question looked at treatment intensity. By splitting their subject pool in half (which, keep in mind, meant the number of kids per condition was pretty small), they found a large effect of treatment after 19 sessions (accuracy of target sounds in new words), and a medium effect after 11. I doubt anyone here is shocked—shocked!—that more therapy leads to better outcomes, but the size of the difference was actually pretty surprising. As we know, many kids don’t just make slow and steady progress, but need to get to that point where things start to “click.” It’s good to realize that the “clicking” place might be a little further away than we think.


Cummings, A., Hallgrimson, J., & Robinson, S. (2018). Speech Intervention Outcomes Associated With Word Lexicality and Intervention Intensity. Language, Speech, and Hearing Services in Schools. Advance online publication. doi: 10.1044/2018_LSHSS-18-0026.

The speech disorder we aren’t paying attention to

Q: What congenital syndrome…

  • May be more than twice as prevalent as autism?

  • Is often misdiagnosed or not spotted at all?

  • Results in a speech impairment for 90% of affected children?


A: It’s fetal alcohol spectrum disorder (FASD). And while all of us probably learned the hallmark physical features and cognitive/behavioral consequences (Do you know how often “philtrum” comes up on Jeopardy? It’s a lot!), the particulars of the speech impairment haven’t been well studied. Traditionally it’s talked about as a speech delay, but clinical SLPs have found the situation to be more… complicated than that.

Speech delay, disorder, or both? It matters, since you’ll approach treatment differently. But some research is missing to connect the dots and guide our intervention. This study begins that work by analyzing the speech of a group of boys* with FASD and comparing it with another group of typically-developing children. The boys with FASD had:

  • Slightly lower overall intelligibility

  • More consonant errors and some differences in order of mastery (in Dutch, FYI)

  • Extra difficulties with nonword repetition

  • Difficulties with auditory discrimination

  • Difficulties with oral motor skills (especially tongue control), causing issues with specific consonants

  • Speech errors related to craniofacial anomalies (e.g., high palate → denasalization errors)

The takeaways for SLPs? Speech in this population seems to be both delayed and disordered. It may be that motor planning and processing deficits are causing many of the speech issues we see. Beyond that, specific characteristics, such as hearing loss, tongue control issues, high arched palates, and phonological impairments (all of which some, but not all children with FASD will have) have additional effects on speech. Clinicians need to evaluate these underlying differences and difficulties and use that to guide treatment. And remember: These kids need a lot of repetition and practice to learn and generalize skills.

Unfortunately, there are no easy solutions for this population. They need “long-term dedicated treatment that is tailored to the individual profile under the guidance of SLPs who are trained in working with these children.” 

*Why just boys? Evidently no families of girls with FASD were willing to participate. Interesting.  So... sex-related differences in the FASD population are still an open question.


Terband, H., Spruit, M., & Maassen, B. (2018). Speech Impairment in Boys With Fetal Alcohol Spectrum Disorders. American Journal of Speech-Language Pathology / American Speech-Language-Hearing Association, 27(4), 1405–1425.

How do you interpret “educational performance”?

We don’t have to remind you of all the challenges facing children with speech sound disorders (SSD), especially since roughly 90% of school-based SLPs serve students with SSDs. Although we have that in common, we’re pretty different in how we (and our districts/states) interpret “educational performance,” a key phrase from IDEA. These differences have a huge impact on which students ultimately get services—and which students don’t.

By surveying SLPs nationwide, the authors of this article found a lot of variability. The guidelines we use come from different agencies (states, districts, state speech–language–hearing associations, etc.), but at least some of the differences are due to our individual decision making, because the survey found that “SLPs are familiar with their state guidelines but do not consistently use them as evidenced by considerable variability within and between states.”


Essentially, we are taking different factors into account when looking for the impact (or lack thereof) of SSDs on kids’ school success. Are you looking at only at grades? Do you weigh access to the curriculum, oral participation in class, or spelling? Do you take social-emotional adjustment into the mix? Consider how you determine educational impact now, and how either a narrower or broader view of the concept would change your practice. Would you have more artic/phono students? Fewer? Would they get services earlier, or keep them longer? Would you do your evals differently? Having the most possible students in therapy isn’t really the goal (must think least restrictive environment), but under-serving these students is definitely a problem.

Big takeaway here: other SLPs out there are likely making decisions very differently from how you are—and it’s time we talked more about it. As you reflect on the questions above, talk with your SLP coworkers and friends—even consider the conversations you might have with administrators, policy makers, and your local and state agencies. Small changes in policy (or how you and your coworkers apply the policy) could help ensure kids with SSDs get the services they need in the schools.

Farquharson, K., & Boldini, L. (2018). Variability in interpreting “educational performance” for children with speech sound disorders. Language, Speech, and Hearing Services in the Schools. Advance online publication. doi: 10.1044/2018_LSHSS-17-0159.