Abstracts

Confrontation naming tasks have long been presumed to be sensitive to left temporal dysfunction and are frequently used in the evaluation of surgical epilepsy patients. Despite wide and frequent use, few studies have examined the validity of confrontation naming tasks in individuals with epilepsy. In 2001, Kubu and colleagues examined the diagnostic utility of the Boston Naming Test (BNT) and the Benton Naming Test using likelihood ratios in a sample of 53 individuals with temporal lobe epilepsy (TLE). They found that neither test improved significantly upon pre-test probability of identifying patients who ultimately underwent left temporal lobectomies when using standard normative data to define level of impairment. The current study sought to replicate and extend the Kubu et al. study in a larger sample of TLE patients to further examine the utility of the BNT in determining ultimate side of surgery. The current study examined the presurgical BNT performance of 348 patients with TLE (Left=175; Right=173) who eventually underwent temporal lobectomies. All patients were right-handed, over the age of 17, and had full scale IQs of [ge] to 70. Likelihood ratios were calculated for all possible cutoff scores between 59 and 27, but were not calculated for lower scores as no right TLE patients scored below 27. The likelihood ratios for BNT cut scores from 59 to 27 ranged from 1.03 to 9.89. As in the Kubu et al. study, results indicated that, when using standard normative data to define level of impairment (scores below 10-12^th percentile), BNT scores did not significantly improve upon pre-test probability of determining patients who would eventually undergo left temporal lobectomy. However, it was apparent that lower cut scores did significantly improve upon the pre-test probability of approximately .50, with post-test probabilities ranging from 50.7% to 90.8%. This study supports the clinical utility of the BNT in identifying epilepsy patients who ultimately go on to have left temporal lobectomies when levels of impairment are extended beyond the standard normative cutoff. In general, patients with lower scores on the BNT are more likely to undergo left temporal lobectomies than TLE patients with higher scores, indicating that the BNT is indeed sensitive to left temporal dysfunction. The potential increase in diagnostic accuracy gained by administering the BNT routinely to pre-surgical TLE patients must be weighed against the costs associated with the test in those patients in whom the test does not significantly improve prediction beyond the known pre-test probability. We argue that the low cost, small time investment, and ease of administering this instrument support continued use of the BNT in the pre-surgical neuropsychological evaluation of TLE patients. In addition, the BNT provides a useful baseline measure against which post-operative changes in confrontation naming can be assessed.