Empirical Techniques for Determining the Reliability, Magnitude, and Pattern of Neuropsychological Change After Epilepsy Surgery

PURPOSE: We assessed test-retest neuropsychological performance in patients with complex partial seizures to derive reliable change indices (RCIs) and regression-based norms for change, indices that may be helpful in assessing cognitive outcome after anterior temporal lobectomy. METHODS: Forty patients with complex partial seizures (CPS) who did not undergo epilepsy surgery were administered a comprehensive neuropsychological battery on two occasions. Their test-retest data were used to compute both RCIs and regression-based norms for change for each neuropsychological index. RCIs corrected for practice effects provide a confidence interval (CI) indicating the degree of performance change required to exceed the variability attributable to sources of error (e.g., practice, test-retest reliability). Regression-based norms for change also correct for several sources of measurement error and examine observed versus expected test-retest changes on a common metric, thereby facilitating determination of the degree and relative magnitude of change across cognitive domains. RESULTS: Mean changes in test-retest performance were generally modest, but were evident across several test measures. Our data indicate a considerable degree of individual variability in test-retest performance. CONCLUSIONS: RCIs and regression-based norms are complementary indexes and can be particularly useful in examining the test-retest performance of individual patients who undergo epilepsy surgery as well as in the more general investigation of cognitive outcome after epilepsy surgery.