Klinische Neurophysiologie 2004; 35 - 314
DOI: 10.1055/s-2004-832226

Under which Condition can fMRI Replace the Intracarotid Amobarbital Procedure (IAP) for Presurgical Language Lateralization?

J Wellmer 1, CE Elger 2, S Weis 3, P Klaver 4, DB Linke 5, H Urbach 6, J Reul 7, J Ruhlmann 8, M Kurthen 9, G Fernandez 10
  • 1Bonn
  • 2Bonn
  • 3Bonn
  • 4Bonn
  • 5Bonn
  • 6Bonn
  • 7Bonn
  • 8Bonn
  • 9Bonn
  • 10Bonn

Introduction: There is a large body of evidence showing that functional magnetic resonance imaging (fMRI) is able to lateralize cortical language representations. Across activation protocols, many studies show good congruence of fMRI-based language lateralization with the gold standard, IAP. However, in up to 25% of patients, fMRI and IAP are discordant. This can be due to a variety of factors. Hence, we developed a simple algorithm that identifies individuals with one strongly dominant hemisphere in whom fMRI can actually replace IAP for language lateralization. Patients and Methods: Forty-five patients with pharmacoresistant epilepsy and suspected atypical language dominance received presurgical language lateralization by fMRI and IAP. For fMRI, we used a semantic-perceptual contrast designed to activate frontal as well as parieto-temporal language areas. It provides an active control condition as well as a continuous performance control. IAP was performed unilaterally on the side of intended surgery. Results: The algorithm: 1) exclusion of patients with possibly BOLD-interfering lesions (n=8); 2) exclusion of patients with non-compliance or protocol atypical activation patterns (n=1); 3) identification of unilateral language dominance across both frontal and parieto-temporal activations according to protocol adjusted volumes of interest and lateralization indices resulted in complete congruence and therefore redundance of fMRI and IAP in 16 of 45 patients. Extrapolating lateralization index thresholds for unilateral language dominance defined in this highly selected group of patients to all epilepsy surgery patients who received language fMRI in Bonn between 2000 and 2002 (n=203) indicates that about two-thirds of IAPs are redundant. Discussion: Without exclusion of interfering factors, fMRI-based language lateralization is error-prone and replacement of IAP by fMRI not recommendable. With the proposed algorithm, however, fMRI is a valuable tool in routine presurgical language lateralization.