Matrix training to teach tacts using speech generating devices: Replication and extension

Marya et al. (2021) worked with children with autism who use speech-generating devices. They wanted to see if matrix training could teach new noun-verb tacts without teaching every single pair.

The team set up word grids. They taught only the diagonal cells, then checked if the kids could say the untaught combos on their device.

Most kids learned the taught pairs quickly. More important, they also used the device to say untaught noun-verb combos that had never been trained.

The results showed recombinative generalization worked for SGD users too, not just for speaking children.

Jimenez-Gomez et al. (2019) got the same outcome with talking toddlers. Marya’s team copied the method and proved it still works when kids press icons instead of speaking.

Bailey et al. (2010) used matrix training for action-picture pairs in preschoolers. The new study swaps pictures for SGD screens and noun-verb tacts, showing the trick travels across materials.

McQuaid et al. (2024) pushed matrix training into telehealth for color-shape tacts. Marya pushed it into SGDs for noun-verb tacts. Both papers extend the same tool to new tech, so practitioners now have two fresh delivery options.

You can save hours of teaching time. Train the diagonal, then probe the rest. If recombinations pop out, move on. This keeps therapy efficient for kids who rely on devices for speech. Try it next session: pick four nouns and four verbs, build a 4×4 grid on the SGD, teach the diagonal, and test what else they can say.