Sentence production after listener and echoic training by prelingual deaf children with cochlear implants.

The team worked with three preschoolers who were deaf and had cochlear implants. None could say full sentences yet.

The kids got matrix training on short sentence parts. Think of a grid: rows were subjects like 'I' or 'You,' columns were verbs like 'want' or 'see.' They practiced hearing and then saying each diagonal cell.

After learning only a handful of trained sentences, each child suddenly understood and said brand-new combos they had never been taught. The untaught sentences popped out right away.

The kids also started labeling things around them more often. Their tacting, or naming, went up.

Meyer et al. (1987) did this first. They showed that matrix training gives kids with intellectual disability new two-word phrases without extra teaching. Perez et al. (2015) moved the same idea to deaf preschoolers and full sentences.

Jimenez-Gomez et al. (2019) and Marya et al. (2021) found the same magic in children with autism. One used spoken words, the other used speech-generating devices. All saw untaught combos appear.

McQuaid et al. (2024) pushed it online. They coached parents through telehealth and still got recombinative generalization for color-shape words. The matrix keeps working even on Zoom.

You can save hours of drill time. Set up a small grid of sentence parts, teach the diagonal, then probe for the rest. If the child mixes the words into new, correct sentences, you have proof of learning. If not, you know exactly where to add teaching. Try it next session with any preschooler who needs longer utterances.