Abstracts

Rationale: The CogScreen is a computer-administered and scored instrument which rapidly assesses deficits or changes in cognitive domains including attention, immediate- and short-term memory, and reaction time. It is widely used in clinical trials and in military and industrial research. It has not been used previously to evaluate cognition in epilepsy patients. We compared the baseline CogScreen results in adult temporal lobe epilepsy patients enrolled in a clinical trial with healthy controls in order to assess its utility in this population. Methods: The initial CogScreen data of well-controlled adult epilepsy patients with an estimated intelligence quotient of >70 using the Wechsler Test of Adult Reading were compared with healthy control subjects. The two groups were matched for age, gender and level of education, however, the control group took no centrally-acting medications. We used six CogScreen subtests based on the use of the CogScreen in previous research. Mean number of items completed, reaction time and accuracy for the six selected subtests were compared between patients and controls with Mann-Whitney U. Results: Six subjects and eight controls enrolled. The six subjects had rare seizures and none in the two days prior to the CogScreen test. There were no baseline differences in age (subjects 49 years, controls 43 years) gender (80% women in both groups) and years of education (subjects 16 years, controls 18 years). Subjects were taking between 1 and 3 antiseizure medications, with a median of 3. Controls were not taking any centrally-acting medications; 6 of 8 were not taking any medications. The CogScreen subtest showing a difference between groups was symbol digit coding. There were significant differences in the number of items completed and reaction time between the groups; p=0.014 and p=0.020 respectively. For the control group, the mean number of items completed was 48 and the mean reaction time was 1674 milliseconds. No subject achieved these numbers of items completed or this reaction time; further the mean the mean number of items completed was 33 and the mean reaction time was 2440 milliseconds. There was no difference in the accuracy however. Conclusions: The symbol-digit coding was significantly different between temporal lobe epilepsy patients and controls. This finding is not surprising given that Symbol-digit coding among the most sensitive of all neuropsychological tests to any cognitive dysfunction. Its inclusion in the CogScreen battery may be useful for determining cognitive dysfunction in an epilepsy patient sample. The fact that it is computer self-administered may prove to be cost-effective in applications such as clinical drug trials. A larger sample size may reveal further differences in the comparison groups.