Many patients with epilepsy have memory deficits in the setting of otherwise normal intelligence. Unfortunately, the treatment options for memory dysfunction in patients with epilepsy are limited. The investigators are conducting a study to evaluate the effects of memantine for the treatment of verbal memory dysfunction in subjects with localization-related seizures. The study involves randomization to memantine therapy or placebo, with cognitive testing and EEG pre- and post-treatment, as well as after an open-label memantine treatment phase.

The primary aim of this study is to evaluate the efficacy of memantine for the treatment of verbal memory dysfunction in subjects with left temporal lobe epilepsy. The investigators expect that verbal memory task performance will improve in those taking memantine, but not in those taking a placebo.

The investigators propose that the expected benefit of memantine is specific to verbal memory in subjects with left temporal lobe seizures, rather than representing an overall improvement in cognitive function. The investigators expect no improvement on other cognitive tasks in either the memantine or placebo groups.

The investigators will evaluate whether subjects with left temporal lobe epilepsy and memory difficulties have self-reported improvement in memory while taking memantine. The investigators expect improvement of self-rated memory function on the Quality of Life in Epilepsy Patient Inventory (QOLIE-89) in the memantine group, but no change on this scale in the placebo group.

Patients with epilepsy frequently demonstrate subtle cognitive difficulties in the setting of otherwise normal intelligence. Those with left temporal lobe seizures often have particular deficits in verbal memory (Blum 2001). These memory difficulties may be the most distressing aspect of epilepsy for the patients.

Unfortunately, treatment options for memory dysfunction are limited. Cognitive therapy, for example, may help patients to cope, but does not treat the memory loss or address the underlying pathologic process. Two studies examined the pharmacologic management of memory dysfunction in patients with epilepsy using donepezil (Aricept), but with inconsistent results and questionable benefit. A pilot study by Fisher et al. (2001) showed some promise for use of the drug. The study found improved immediate recall and consistent long-term retrieval scores on the Buschke Selective Reminding Test after three months of open-label treatment when compared to a pre-treatment baseline. A more recent randomized, double-blind, placebo-controlled cross-over trial of donepezil, however, showed no effect on memory as measured by delayed recall on the Hopkins Verbal Learning Test (Hamberger et al. 2007).

Use of donepezil, an acetylcholinesterase inhibitor, may pose a risk of seizure exacerbation in this population. Fisher et al. (2001) reported a significant increase in the frequency of generalized tonic-clonic seizures during donepezil treatment. Cholinergic agents have been shown to cause seizures in animal models as well (Turski et al. 1989). Given isolated case reports of seizures associated with donepezil use, the manufacturer issued an advisory note warning of a possible relationship, although data have been insufficient to establish causality. While an increase in seizures was not noted in the Hamberger et al. (2007) study, seizure exacerbation remains a concern regarding the use of this drug in patients with epilepsy.

The mechanism for the postulated effect of donepezil is unclear. Cholinergic transmission has not traditionally been viewed as a contributor to hippocampal pathology. It is believed that excitotoxicity, mediated by glutamate acting on NMDA receptors in the hippocampus, causes hippocampal sclerosis. This process leads to further seizures and memory dysfunction. Alteration of this excitotoxic pathway would be a novel, and potentially safer and more effective, approach to the treatment of memory loss.

The possible effect of intervention at the level of excitotoxicity is supported by animal data. Such studies demonstrate that induced seizures in a rat model of epilepsy will lead to decrements in performance of a spatial memory task, the Morris water maze. This memory dysfunction, however, can be mitigated by NMDA antagonists, such as MK-801, administered prior to seizure induction. The underlying concept is that NMDA receptor antagonists would block the pathway of excitotoxicity that leads to hippocampal injury and memory loss (Kelsey et al. 2000).

An NMDA antagonist, memantine (Namenda), is prescribed in humans for treatment of moderate to severe Alzheimer's disease (Tariot et al. 2004, Reisberg et al. 2003, 2006). Patients with Alzheimer's disease have attained significant cognitive improvements with use of the drug, as measured by the Severe Impairment Battery. The time-course of benefit is less clear, with some studies demonstrating sustained improvement (Tariot et al. 2004) and others showing more transient benefits over the first several weeks of treatment (Reisberg et al. 2003).

It is unknown, however, if an NMDA antagonist such as memantine would be of benefit in humans with memory dysfunction and seizures. The proposed study tests the hypothesis that treatment with memantine would improve verbal memory test performance in patients with localization-related epilepsy. If beneficial, this would provide a much-needed treatment option.

The study will examine the primary specific aim:

Aim 1: Improvement in memory test performance. The primary aim of this study is to evaluate the efficacy of memantine for the treatment of verbal memory dysfunction in subjects with left temporal lobe epilepsy. We expect that verbal memory task performance, as measured by the Buschke Selective Reminding Test (SRT), will improve in those taking memantine, but not while taking a placebo. Such a finding would support the use of memantine for treatment of memory loss in this population, as well as more generally support the hypothesis that NMDA receptor hyperactivity is an appropriate target for intervention.

The study will examine two secondary specific aims:

Aim 2: Selectivity of response. We propose that the postulated benefit of memantine is specific to verbal memory in subjects with left temporal lobe seizures, and visuospatial memory in subjects with right temporal lobe seizures, rather than representing an overall improvement in cognitive function. We expect no improvement on other cognitive tasks in either the memantine or placebo groups, with measures including the Digit Span (for sustained attention, immediate span), Spatial Span (for visuospatial working memory and span), Block Design (for visuospatial construction), Verbal Fluency, Design Fluency, and Stroop Color Word Interference (for executive function) tests. This would lend support to the hypothesis that blockade of NMDA receptor hyperactivity in the hippocampus would lead to improved performance on cognitive tasks that depend specifically on the integrity of that hippocampus, as opposed to a general benefit in overall cognition.

Aim 3: Improvement in self-reported memory function. We will evaluate whether subjects with localization-related epilepsy and memory difficulties have subjective improvement in memory with the administration of memantine. We expect improvement of self-rated memory function on the Quality of Life in Epilepsy Patient Inventory (QOLIE-89) with memantine, but no change on this scale with placebo. This measure serves to evaluate the hypothesis that memantine treatment leads to clinically meaningful improvement.

Inclusion Criteria:

• 18-65 years of age
• Normal IQ as estimated by the Wechsler Test of Adult Reading (WTAR)
• Able to give consent
• Able to live independently and complete activities of daily living
• Stable frequency of seizures. There is no minimum/maximum criteria for the frequency of partial seizures. Those with infrequent secondary generalized seizures may participate, with infrequent seizures defined as two or fewer per year.
• The subject's treating physician does not believe a change in anticonvulsant regimen to be warranted. The anticonvulsant drugs must remain unchanged during the 26 week trial.
• Partial-onset seizures. Seizure type will be determined by clinical history, MRI, SPECT and/or PET imaging, and interictal and/or ictal EEG.
• Either symptomatic or idiopathic seizures.

Exclusion Criteria:

• Non-epileptic seizures
• Prior surgical resection for treatment of seizures
• Progressive neurologic illness (i.e. tumor evident on MRI)
• Current alcohol or drug abuse, as this may affect memory by other mechanisms. This information may be obtained by self-report, from the referring physician or by medical record.
• Diagnosis of Alzheimer's disease, nutritional deficiency, infection or metabolic/electrolyte disorder causing memory loss.
• Non-native English speaking and/or multilingual.
• Seizure(s) must not have occurred within 3 days of testing.
• Subjects who are pregnant will not be eligible to take part in the study, as memantine is classified as a Pregnancy Category B drug and may pose risk to the fetus.
• Women who are breastfeeding may not participate in this study.
• Those with renal tubular acidosis or infections of the urinary tract will not be eligible for participation, as memantine is renally cleared and conditions that alkalinize the urine may reduce clearance of the drug.
• Subjects with severe renal impairment, defined as a creatinine clearance of ≤29 mL/min, will be excluded as such patients may not tolerate the proposed dosing schedule.