Abstracts

The current gold standard for localizing language and memory function in temporal lobe epilepsy (TLE) pre-surgical evaluation is the intra-carotid sodium amytal (ICA) procedure. While various fMRI paradigms predict language localization, the invasive ICA procedure remains the standard of care due to lack of a reliable method to achieve fMRI hippocampal activation. In this study, we assess the usefulness of an fMRI language localization and hippocampal activation paradigm by comparing it to ICA test a series of TLE patients. Four consecutive TLE patients being evaluated for epilepsy surgery had an ICA test yielding a quantitative representation of functional reserve for both memory and language for each hemisphere. All patients participated in an fMRI study in a 3-T GE magnet, using a protocol which yields reliable bilateral hippocampal activation in normal subjects. The patients were presented with a learning trial followed by 5 identical recall trials in a block design. The learning trial consisted of an individual presentation of each of 10 objects for 3 seconds, for a total of 30 seconds of memory acquisition. For recall trials, the patient was presented with a series of 10 objects for 3 seconds each, 1 or 2 of which were from the original list. The patient pushed one switch to indicate recognition of the object and another switch for absence of recognition. Each epoch was followed by a 30 second rest period; activation maps were generated from a comparison of the 5 recall epochs with the 5 rest epochs. Language localization was 100% congruent between fMRI and ICA test in our series. Specifically, 3 of the 4 patients were left-hemisphere dominant for language while the fourth patient (a right-handed male with seizure onset as a young child) was right hemisphere dominant. Three of the patients exhibited bilateral hippocampal activation. Of these, memory lateralization was not definitively established using ICA test, as both hemispheres appeared critical for memory acquisition; that is, recall scores were reduced by 60-90% from baseline testing for each hemisphere. The fourth patient exhibited unilateral left hippocampal activation during the fMRI memory paradigm, which was congruent with the ICA results, demonstrating memory functions strongly lateralized to the left hemisphere. This fMRI paradigm, previously demonstrated to produce reliable language localization and bilateral hippocampal activation in normal subjects, demonstrates congruence with ICA language and memory findings in our preliminary series of TLE patients. The ICA test did not augment the clinical information obtained from the fMRI for both memory and language localization in this series. This particular fMRI paradigm shows promise as a potential tool for replacing the invasive ICA procedure in select epilepsy surgery candidates.