Using Visual Strategies to Remember Verbal Information: An fMRI Study of Working Memory in Children with and without Autism

Friday, May 18, 2012: 10:15 AM
Grand Ballroom West (Sheraton Centre Toronto)
10:15 AM
E. J. Carter1,2, D. L. Williams3, J. F. Lehman4 and N. J. Minshew5, (1)Robotics Institute, Carnegie Mellon University, Pittsburgh, PA, (2)Psychology, Carnegie Mellon University, Pittsburgh, PA, (3)Speech-Language Pathology, Duquesne University, Pittsburgh, PA, (4)Computer Science, Carnegie Mellon University, Pittsburgh, PA, (5)Psychiatry & Neurology, University of Pittsburgh, Pittsburgh, PA
Background: Previous fMRI research suggested that adults with autism performed a verbal working memory task with individual letters using regions related to visual-processing rather than the expected language regions (Koshino et al., 2005, NeuroImage). We expanded this research to children and adolescents with autism using a verbal working memory task with two-letter words that might be more likely to trigger the use of encoding with language; we also included a comparison visual working memory condition.

Objectives: To compare the fMRI brain activity of children with and without autism in response to working memory tasks for two-letter words and simple visual patterns and determine whether visual or verbal strategies are used.

Methods: Twelve 8- to 15-year-old children with autism and twelve age- and IQ-matched typically developing children (TD) who were good readers successfully participated in this IRB-approved fMRI study. Nine task blocks included a 0-back working memory task (i.e., comparing each item to the first item viewed) and a 1-back memory task (i.e., comparing each item to the previous item) and for two types of stimuli: (1) two-letter words and (2) patterns consisting of four angled bars. Each item was displayed for 2s and no feedback was given. Each block lasted 66s with 5s of instructions to indicate the task (and the originating pattern for the zero-back task) and 20 stimulus trials with 1s inter-trial intervals. All of the children had a minimum score of 70% correct with no performance difference between the groups.

Results: For the combined visual tasks versus fixation, both groups showed activity in bilateral occipital cortices (BOcc), superior parietal lobe, and inferior temporal gyri. For the combined verbal tasks versus fixation, the autism group showed BOcc activity and a very limited area of activity in left Broca’s area (LBA). This LBA activity was found only in the verbal 1-back task. TD children showed activity in BOcc and extensive activity in LBA as well as right middle superior temporal gyrus. LBA activity was present for both the 0- and 1-back verbal tasks in this group. When directly comparing the verbal and visual tasks, the autism group only showed significantly increased activity in a limited area of BOcc, whereas the TD group also showed increased activity for the verbal task in LBA and bilateral middle temporal gyrus (BMTG). Group comparisons showed that the TD group had higher LBA and BMTG activity than the children with autism when comparing the verbal tasks to the visual tasks. There were no group differences for the visual > verbal contrast.

Conclusions: The children with autism relied more heavily on visual strategies to perform working memory tasks for two-letter words, particularly for 0-back task, whereas typically developing children used the expected linguistic strategies as indicated by the activation of left Broca’s area. The children with autism demonstrated some use of this region in response to increasing task demands in the verbal 1-back condition, suggesting that cognitive resources from this region could be elicited when needed to accomplish the task.

 

| More