Brain Regions Underlying Repetition and Auditory-Verbal Short-term Memory Deficits in Aphasia: Evidence from Voxel-based Lesion Symptom Mapping

Abstract

Background: A deficit in the ability to repeat auditory-verbal information is common among individuals with aphasia. The neural basis of this deficit has traditionally been attributed to the disconnection of left posterior and anterior language regions via damage to a white matter pathway, the arcuate fasciculus. However, a number of lesion and imaging studies have called this notion into question.

Aims: The goal of this study was to identify the neural correlates of repetition and a related process, auditory-verbal short-term memory (AVSTM). Both repetition and AVSTM involve common elements such as auditory and phonological analysis and translation to speech output processes. Based on previous studies, we predicted that both repetition and AVSTM would be most dependent on posterior language regions in left temporo-parietal cortex.

Methods & procedures: We tested 84 individuals with left hemisphere lesions due to stroke on an experimental battery of repetition and AVSTM tasks. Participants were tested on word, pseudoword, and number-word repetition, as well as digit and word span tasks. Brain correlates of these processes were identified using a statistical, lesion analysis approach known as voxel-based lesion symptom mapping (VLSM). VLSM allows for a voxel-by-voxel analysis of brain areas most critical to performance on a given task, including both grey and white matter regions.

Outcomes & results: The VLSM analyses showed that left posterior temporo-parietal cortex, not the arcuate fasciculus, was most critical for repetition as well as for AVSTM. The location of maximal foci, defined as the voxels with the highest t values, varied somewhat among measures: Word and pseudoword repetition had maximal foci in the left posterior superior temporal gyrus, on the border with inferior parietal cortex, while word and digit span, as well as number-word repetition, were centered on the border between the middle temporal and superior temporal gyri and the underlying white matter.

Conclusions: Findings from the current study show that 1) repetition is most critically mediated by cortical regions in left posterior temporo-parietal cortex; 2) repetition and AVSTM are mediated by partially overlapping networks; and 3) repetition and AVSTM deficits can be observed in different types of aphasia, depending on the site and extent of the brain injury. These data have implications for the prognosis of chronic repetition and AVSTM deficits in individuals with aphasia when lesions involve critical regions in left temporo-parietal cortex.

Keywords: aphasia; conduction aphasia; parietal cortex; repetition; short-term memory; temporal cortex.

Figure 1

Overlay of participants' reconstructed lesions. This figure depicts only those voxels with a minimum of at least 5 individuals' lesions (shown in purple), which was the minimum number of individuals per voxel necessary to be included in the VLSM analyses to protect against spurious findings.

Figure 2

VLSM map of overall performance on the experimental repetition/span battery. All voxels shown in color exceeded the critical threshold for significance, and the colors reflect increasing t-values from purple to red.

Figure 3

VLSM map showing correlates of single word repetition. Only significant voxels are shown, and the colors reflect increasing t-values from purple to red.

Figure 4

VLSM map showing correlates of pseudoword repetition. Only significant voxels are shown, and the colors reflect increasing t-values from purple to red.

Figure 5

VLSM map showing correlates of number-word repetition. Only significant voxels are shown, and the colors reflect increasing t-values from purple to red.

Figure 6

VLSM map showing correlates of word span performance. Only significant voxels are shown, and the colors reflect increasing t-values from purple to red.

Figure 7

VLSM map showing correlates of digit span performance. Only significant voxels are shown, and the colors reflect increasing t-values from purple to red.

Figure 8

On the left, the maximum t-values are shown for single word repetition (orange) and pseudoword repetition (yellow), located at the border of left posterior STG/IP cortex. On the right, the maximum t-values are shown for word span (blue), digit span (red), and number-word repetition (green), located at the medial extent of the temporal gyri adjacent to underlying white matter. Note that word span is shown on the same slice as digit span and number-word repetition (x=−46) but was actually a few mm medial (x=−42).

Figure 9

Probabilistic diffusion tensor imaging-based map of the arcuate fasciculus (shown in dark yellow), thresholded at 75% of cases, overlaid on the VLSM map for overall repetition/span performance. The colored regions and legend indicate the t-values from the original VLSM analysis of overall repetition/span performance shown in Figure 2.