Abstracts

Rationale: Classic experiments have established a central role for medial temporal lobe (MTL) structures for long-term memory formation. More recent work has identified a role for MTL in certain types of short-term memory, especially memory for complex or detailed information. However, memory is not an all-or-nothing phenomenon. Even if information is successfully recalled, it can be recalled with varying degrees of detail or precision. While at least one study has suggested that MTL plays a role in maintaining precisely bound information, others have identified parietal lobe or visual cortex as the primary substrate visual mnemonic precision. We therefore queried whether precisely recalling visual motion over a short time interval depends on MTL function.  Methods: Healthy controls and patients with temporal lobe epilepsy (TLE) performed a short term memory task. On each memory trial, participants were shown one or three circular apertures filled with coherently moving dots. Each moving dot array was a distinct color with a distinct direction of movement. Participants were instructed to remember the colors and directions. After a brief delay, a colored cue indicated to participants which dot array to recall. Participants then responded by using a computer mouse to orient a radial line to match the remembered direction. This free response data was analyzed using a formal mixture model to estimate various types of errors, including guesses, binding errors (in which subjects mistakenly recalled the direction of an uncued color), and precision (which is determined by the variation of responses around the correct value). Results: Although data collection is ongoing, preliminary results demonstrate that participants with TLE (n=6) are able to perform the task as expected. Across all participants (n=7), precision decreased (2-tailed t(6)=2.47, p=0.048) and reaction time increased (2-tailed t(6)=5.98, p=.001) when comparing one-item and three-item trials. Although there are not enough subjects at present to compute between-group statistics, we would predict that the number of binding errors will be increased and the precision of responses will be decreased in participants with MTL epilepsy relative to normal controls. Conclusions: Early results illustrate typical effects of memory load, in that higher memory load slowed reaction time and decreased mnemonic precision. More data is needed to determine whether patients with TLE will exhibit more binding errors and less precise responses, as predicted. Such a pattern of results would suggest that temporal lobe dysfunction impacts short term memory not only by disrupting the binding of information, but also by impairing the accurate recall of visual information. This result would be at odds with previous work suggesting that parietal lobe activity determines the precision of short-term visual recall.  Funding: NYU Finding a Cure for Epilepsy and Seizures (FACES)