Authors :
Presenting Author: Alex Bender, MD, PhD – Massachusetts General Hospital
Sydney Cash, MD, PhD – Neurology – Massachusetts General Hospital; Alice Lam, MD, PhD – Neurology – Massachusetts General Hospital
Rationale:
Memory impairment in epilepsy is a common and debilitating comorbidity. Patients complain of problems ranging from short to long intervals and long-term impairment has been demonstrated even in the presence of normal short-term performance, yet currently standard neuropsychological testing does not assess memory on this timescale. In this study, we developed a multi-day memory testing protocol to study overnight memory and the learning and forgetting of information that occurs over days-to-weeks.
Methods:
We studied 18 adult patients admitted to the Epilepsy Monitoring Unit (EMU) at Massachusetts General Hospital who did not have any prior brain surgery, lesion, or learning disability. A global cognitive assessment was completed using the MOCA. Two separate types of repeated memory testing were subsequently administered. The Rey Auditory Verbal Learning Test (RAVLT), a 15-item word list, was administered with a 30-min recall and an additional overnight delayed recall. An alternate version was administered again two to three days later. In the second type of memory test, a modified version of the Boston Remote Assessment for Neurocognitive Health (BRANCH was used. This novel web-based platform is comprised of face-name, grocery-price, and symbol-digit associations that are repeated over 7 consecutive days to measure “learning” performance. Recall-only versions were then repeated 4 additional times over the following 3 weeks to measure “forgetting."
Results:
We enrolled 18 subjects of mean age 38 years (range 21-68). Scores on the MOCA ranged from 17-30 with a mean of 24.7 (std dev 3.7). On the RAVLT, performance at a 30-minute delay was below average, with subjects recalling a mean of eight words (std dev 4.7) in comparison to an age-adjusted normative mean of 11 words (t=-2.9, p=0.01). Overnight, there was an additional forgetting of two words (std dev 1.9). In subjects with repeated RAVLT testing during the latter half of the EMU stay, performance was significantly lower on the overnight delayed recall, with subjects forgetting five words (std dev 4; t=2.8, p=0.01) despite similar performance on the 30-minute delayed recall (mean 10 words, t=-0.8, p=0.4). On the BRANCH, performances improved over the first seven days demonstrating a learning curve that was steepest on face-name association, with a slope of ~7 (% correct/day). Over longer-term intervals of up to thee weeks, there was a slow forgetting of learned items with an average slope of -0.5.
Conclusions:
Memory performance in patients admitted to the EMU was impaired at short delay intervals. However, additional forgetting of learned information occurred overnight and over days, and in particular only the overnight recall worsened in the latter half of the EMU stay with relative preservation of the short delay recall. We speculate that this is related to detrimental effects on sleep in the setting of increased seizures or epileptiform activity while in the EMU. Importantly, this study demonstrates the feasibility of memory testing over a multi-day period in a clinical setting and will provide an important avenue to further understand the factors that influence memory complaints on a multi-day timescale.
Funding:
NINDS R25NS065743