Longitudinal fMRI study of aphasia recovery in frontal stroke Lisa Tabor Connor, a, * Valeria Blasi, d Alexis C. Young, a Frances M. Tucker, a Abraham Z. Snyder, a,b Amy Breier Kwentus, b Christopher Lewis, a and Maurizio Corbetta a,b,c a Department of Radiology, Washington University School of Medicine, St. Louis, MO, USA b Department of Neurology, Washington University School of Medicine, St. Louis, MO, USA c Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, MO, USA d Neuroradiology Institute, San Raffaele Hospital, Milan, Italy Available online 23 July 2004 Introduction Although many who experience aphasia after stroke recover their language abilities, it is unclear what neurobiological mechanisms support language recovery. Regions in the right hemisphere are often recruited in aphasic participants (e.g., Blasi et al., 2002; Buckner, Corbetta, Schatz, Raichle, & Petersen, 1996; Rosen et al., 2000), but it is unknown whether the right hemisphere underlies improvements in language performance. The current study examines language recovery longitudinally in individuals with left frontal stroke with aphasia using event-related fMRI. The goal was to examine regions that may con- tribute to language recovery and determine whether their responses reflect improvements in language performance. Method Participants Eleven individuals (5 females, 6 males) with a single left frontal stroke participated. They ranged in age from 22 to 73 years (M = 44 years), were enrolled within 1 month of their stroke (M = 21 days), and tested again approximately 7 months post-stroke (range 5.75–10 months). All persons had verbal output in the mild- to moderately impaired range based on either the Western Aphasia Battery or the Boston Diagnostic Aphasia Examination administered at both the acute and chronic sessions. Aphasia severity, fluency, and naming were converted to z scores. Improvements were made from acute to chronic testing with the following exceptions: aphasia severity was unchanged in 1 patient, fluency was unchanged in 1 other patient, and naming was unchanged in 4 patients. Procedure Each participant performed two tasks overtly in the scanner: reading ‘‘yes’’ and ‘‘no’’ and generating a word from a three-letter word stem. Stimuli were presented centrally for 3.5 s, followed by a random interstimulus interval of 3.58, 5.94, or 8.30 s. MR scans were acquired with a 1.5 Tesla Siemens Vision scanner. Eight BOLD runs of 128 frames, 4 runs of each task, were acquired with each frame in- cluding 16 contiguous 8 mm axial slices (3.75 · 3.75 mm in-plane resolution). In addition, anatomical images were acquired with a sagittal T1-weighted MP-RAGE sequence (TR = 97 ms, TE = 4 ms, flip angle = 12°, TI = 300 ms). Imaging analysis Functional data were realigned within and across runs to correct for head movement, and coregistered with anatomical data. For each subject, model independent timecourses (8 frames long) were esti- mated at the voxel level using the general linear model (Ollinger, Shulman, & Corbetta, 2001). Individual time-points of each hemo- dynamic response were used to compute voxel-level or regional ANOVAs using a random-effects model. F maps at the voxel level were corrected for multiple comparisons. Regions of interest were selected on the multiple comparison corrected image by including for each region all voxels above the significance threshold (z = 3 at cluster size of 45 voxels). Results and Discussion Regions sensitive to task demands Several regions, primarily in the right hemisphere, produced a larger and more sustained response for word stem completion than for yes/no reading regardless of stage of recovery. Most notably, the right dorsal inferior frontal gyrus (IFG), x = 42, y = 2, z = 28, F(7,70) = 6.68, p < .001, a region that has been implicated in other studies of recovery, showed this pattern. Because the right IFG was involved at the acute stage while participants were not performing very well on word stem completion, this region is unlikely to underlie language recovery. Other regions that produced this same pattern included the right medial frontal gyrus, x = 6, y = 13, z = 37, F(7,70) = 4.88, p < .01, and the right supramarginal gyrus, x = 44, y = )54, z = 36, F(7,70) = 7.29, p < .001. Brain and Language 91 (2004) 82–83 www.elsevier.com/locate/b&l * Corresponding author. Fax: 1-314-362-6911. E-mail address: lconnor@npg.wustl.edu (L.T. Connor). 0093-934X/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bandl.2004.06.044