Verbal and spatial working memory in autism.

The team asked adults with high-functioning autism to hold short strings of letters in mind while doing math. They also asked them to remember where red squares popped up on a screen.

Both tasks got harder in two steps: two, then four, then six items. The study wanted to see if verbal memory or spatial memory broke first.

Letter recall stayed solid even at the hardest level. Square recall fell apart as soon as the set grew. The gap got bigger when the load increased.

Verbal working memory looked typical; spatial working memory did not. Task complexity was the key lever.

Bled et al. (2024) extends this picture. They showed that autistic adults can keep a mental picture longer than peers, but they scan it differently. Static storage is fine; dynamic updating is shaky.

Emerson et al. (2025) adds physiology. Strong grammar skills predicted less efficient mental-rotation effort in autism, the opposite of typical adults. Together these papers trace a clear line: verbal routes stay open, spatial routes clog.

Stancliffe et al. (2007) seems to disagree at first glance. They found free-recall word lists stayed poor even after strategy training. The clash fades when you see the task: L et al. used working-memory span, J et al. used long-term recall. Verbal working memory can be intact while verbal long-term organization stays weak.

When you give instructions, keep verbal chunks short and steady, but do not count on spatial memory to hold multiple steps. Pair spoken directions with written or visual cues that stay in view. If you must use spatial arrays, start with two items and test before you add more. This small tweak can prevent overload and reduce prompt dependency.