Abstracts

RATIONALE: Evaluation of memory function in neurosurgical patients focuses on medial temporal lobe (MTL) structures. Intracarotid amytal (Wada) testing attempts to predict potential amnesia from MTL resection, but is an unreliable predictor of post-operative deficits. Recent fMRI studies of normal subjects implicate dorsolateral frontal lobe as well as MTL in memory processes, with lateralized frontal activation during memory encoding depending on the material remembered.
METHODS: Memory function was assessed in two groups of patients. If lateralization of memory function during Wada memory testing is driven by frontal perfusion (rather than posteriorly perfused MTL), a strong lateralizing effect should be seen with memory for faces compared to memory for words. During Wada testing, words and faces were presented to each hemisphere after injection. Recognition memory for both items was tested after recovery from anesthetic.
In a second group of patients with frontal subdural grid electrodes placed for seizure monitoring (10 patients, 5 left sided, all patients left dominant for language), stimulation mapping of memory assessed whether frontal regions were essential (at least acutely) for memory processes. Memory was studied by presenting four slides: a picture which was named, two slides which were reading tasks (distracters) and a fourth slide which cued recall of the initial picture. At each site mapped, stimuli were applied either during naming (encoding), during the distracter, or during recall. For each trial, stimulation was applied only once (during either encoding, distraction, or recall).
RESULTS: In all 12 Wada pts (all left dominant for language) later memory for words, but not faces, was impaired with left-sided injection with the opposite effect for right injection (p[lt].0001). In another case, verbal memory was lateralized opposite to speech and frontal fMRI activation with encoding predicted this lateralization.
For the grid patients, one left sided patient had poor baseline memory preventing mapping. One right-sided patients made no errors with stimulation and was found to have cortical dysplasia in the area that disrupted memory in other patients. The remaining four left-sided and four right-sided patients all showed sites disrupting encoding or recall. These sites closely matched the regions predicted from the imaging studies.
CONCLUSIONS: Convergent results from imaging studies of normals, Wada tests, and stimulation mapping suggest a role for frontal lobe in memory. Frontal lobe regions likely act with medial temporal lobe regions to produce successful memory. These findings may explain the poor correlation between Wada memory testing and memory function following medial temporal resections.
Support: NS32979, NS06883, NS07205, NS36527.