Abstracts

Rationale: A large proportion (up to 60%) of patients who undergo surgery of the left temporal lobe (TL) experience a decline in naming. It is important to identify the patients who may be at greater risk for such a decline for presurgical counseling and planning purposes. Magnetic Source Imaging (MSI) is used to reliably lateralize/localize language specific regions in the brain. This study investigated the use of MSI language mapping to predict language decline in epilepsy patients who underwent surgical resection of the left TL. We also compared the predictive power of MSI to other previously indentified risk factors including age of seizure onset (ASO), presurgical naming ability, and language performance on the intracarotid amobarbital procedure (IAP). Methods: The sample was N = 15 with 47% male and 87% right handed. The mean age at surgery, education level, and ASO were 34.6, 13.0, and 10.4, respectively. For the MSI, an auditory word recognition task was used to activate TL language-specific cortex. The MEG unit consisted of a 148-channel Magnes 2500 system. Single equivalent current dipole model across each whole hemisphere was used in determining the number of language activations. For the IAP, only language responses occurring prior to motor recovery were counted in the score. A language asymmetry index (AI) was calculated using the formula R-L/R+L for both the MSI and IAP. Scores ranged from -1 (left dominant) to +1 (right dominant). Pre and post surgical neuropsychological exams were obtained. The mean pre-surgical Full Scale IQ was 90.8. The 60-item Boston Naming Test (BNT) was used as a measure of naming ability. The mean presurgical BNT raw score was 45.33. A BNT change score (pre - post) was calculated to characterize the patient's naming outcome following surgery. Results: In this sample, 67% showed some decline on the BNT (mean change score = -7.1), with 40% showing a mean change score of -19.0 following left TL surgery. The mean AIs for the MSI and IAP were .01 and -.82, respectively. The following 2-tailed Pearson correlation coefficients (r) between the BNT change score and variables of interest were obtained in order of magnitude: presurgical BNT = .26(ns), MSI AI = .22(ns), IAP AI = -.09(ns), and ASO = -.07 (ns). Conclusions: The majority of patients showed a decline in naming ability following left TL surgery. However, none of the variables examined were able to significantly predict this decline, likely due to our small sample size. There was a trend for greater MSI language activation in the left hemisphere to be related with larger declines in naming post surgery. Furthermore, this trend was comparable in magnitude to an established risk factor (i.e., presurgical naming) and stronger than the IAP and ASO predictors. We intend to study these preliminary results further by increasing our sample size with the goal of improving our assessment of the risks and benefits of TL surgery.