Minimally Verbal Children with ASD: Neural Mechanisms and a Novel Speech Intervention

Background: It has been estimated that up to 30% of children with ASD are minimally verbal or nonverbal, even after receiving years of intensive therapies. However, almost all clinical and neuroimaging studies have focused on the high-functioning, verbal end of the spectrum. As a result, we know almost nothing about why a large portion of children with ASD are nonverbal. This limited knowledge prevents clinicians from testing new therapies and making appropriate treatment recommendations. Over the past few years, our laboratory has developed a novel speech intervention – known as Auditory-Motor Mapping Training (AMMT) – for nonverbal verbal children with ASD. This intervention works by mapping sounds to oral articulatory actions through intonation and bimanual motor activities. Furthermore, the association of sounds with actions engages an auditory-motor network of brain regions, which has been reported to be dysfunctional in ASD, but is of critical importance for developing speech.

Objectives: Here, we report a series of treatment studies that our laboratory has conducted to test the efficacy of AMMT, including a proof-of-concept study, and a randomized controlled trial comparing AMMT to a control treatment (CT). In addition, we report our ongoing neuroimaging research on nonverbal children with ASD, which seeks to understand the neural correlates of nonverbal ASD, as well as to examine whether treatment can change the relevant brain structures.

Methods: All children who participated in our studies were between 4-10 years of age, and had no intelligible words prior to enrollment. They underwent intensive one-on-one AMMT sessions 5 times per week for 5 weeks, and were assessed on their consonant-vowel productions multiple times before, during, and after therapy. In our proof-of-concept study, 6 children participated. In our AMMT vs CT study, 16 children were randomly assigned to receive either AMMT or CT. In CT,the key components of AMMT (intonation and hand-motor actions) were omitted, but the intervention was also designed to promote speech production. In our brain imaging studies, 12 nonverbal children were scanned using DTI, sMRI, and fMRI without sedation. Their brains were compared with those of age-matched typically-developing children.

Results: Results from our treatment studies showed that AMMT resulted in significant improvements in speech production (i.e., increased repertoire of speech sounds and intelligible words) after only a few weeks. Furthermore, their speech improvements transferred to words that were not trained during the therapy sessions. AMMT also yielded superior outcomes compared to CT, suggesting that the critical components of AMMT (intonation and bimanual activities) were likely to be responsible for the therapeutic effects. Results from our ongoing imaging study showed that the language-related regions and inter-regional connectivity of the brains of the nonverbal children are abnormal.

Conclusions: AMMT appears to have significant clinical potential in facilitating the development of speech children with ASD who are completely nonverbal. Its effectiveness may lie in its ability to engage and facilitate connections between language-related areas that may be abnormal in these children.