fMRI Identification of Language Dominance: Comparison of Visual Reading, Region of Interest Analysis, and the Intra-Carotid Amytal Test.
Abstract number :
1.243
Submission category :
Year :
2001
Submission ID :
1684
Source :
www.aesnet.org
Presentation date :
12/1/2001 12:00:00 AM
Published date :
Dec 1, 2001, 06:00 AM
Authors :
W.D. Gaillard, MD, Neurology, Children[ssquote]s National Medical Center, Washington, DC; L. Balsamo, MA, Epilepsy Research Branch, NIH/NINDS, Bethesda, MD; B. Xu, PhD, Epilepsy Research Branch, NIH/NINDS, Bethesda, MD; P. Papero, PhD, Neurology, Children
RATIONALE: fMRI is increasingly used to determine language dominance in preparation for epilepsy surgery, but the optimal methods for fMRI interpretation are uncertain. We compared region of interest (ROI) fMRI analysis, visual reading of fMRI, and the Intra-Carotid Amytal Test (IAT) from a reading response naming paradigm.
METHODS: We studied 29 patients (3 left handed, 2 ambidextrous) aged 8-55 years (mean 25 yr.), with a ictal video EEG confirmed temporal lobe focus using whole brain 1.5T fMRI (EPI BOLD). During the task patients read a 6-7 word object description ([dsquote]what is a long yellow fruit[dsquote]), compared to rest, using block design of six control/task cycles (each 64 sec). Data were analyzed with a region of interest analysis from t-maps (t=5,4,3). The number of activated voxels was determined in inferior frontal gyrus (IFG), middle frontal gyrus (MFG), and Wernicke[ssquote]s area using a semi-automated program. An asymmetry index (AI) was calculated [(L-R)/(L+R)] for each region. 19 patients had an IAT, and one surgical confirmation. 3 raters viewed t maps (t=3,4,5) to determine laterality of activation while blinded to subject identity. Studies were deemed left or right dominant, bilateral, or indeterminate. Results were compared using Cramer V and Chi-square tests.
RESULTS: There was strong correlation between IAT, reader rating and ROI analysis (Chi-square p[lt]0.001). Cramer V ranged from 0.59 and 0.70 between individual raters and IAT; ranged between 0.72 and 0.74 for ROI and Wada; and varied between 0.63 and 0.73 between raters for all patients. Only 2 patients had no agreement between raters (both had poor activation); There were no overt right-left disagreements, rather discrepancies arose from deciding bilateral vs. unilateral activation in a minority of patients. Two had poor activation, but one fMRI study identified strong hemispheric dominance with a non-diagnostic IAT. ROI analysis at lower thresholds identified laterality in 3 additional patients.
CONCLUSIONS: For most studies there is excellent agreement between visual and ROI analysis and IAT. Lower thresholds may be reliable and may extend information. In a few instances there is partial disagreement on the extent of language dominance. No studies gave contrary results.
Support: Supported by NINDS KO8-NS1663 & ERB, NINDS
Disclosure: Grant - Supported by NINDS KO8-NS1663 & ERB, NINDS