Abstracts

RATIONALE: For neurosurgical planning in medial temporal lobe epilepsy it is important to be able to understand the role of the hippocampus in mnemonic processing and to be able to map hippocampal activity. The hippocampal formation is required for encoding new memories but it has been difficult to show activation of the hippocampus proper in fMRI studies of encoding. Clinically however understanding encoding better is of great concern as anterograde amnesia represents a general failure of memory encoding.
METHODS: In the current study fMRI was performed on 12 subjects (age range 21-30, 8 male, 4 female) performing a verbal association encoding task. 30 pairs of words (high - frequency, concrete nouns) were presented, and the subjects were required to remember what second word was paired with each first word. Ten pairs were presented 12 times in the course of the experiment (easy set), and 20 pairs were presented only 2 times each (hard set). The actual word-pairs comprising each set were arbitrarily divided into these two groups for each subject. Imaging was performed in the coronal-oblique plane (perpendicular to the long axis of the hippocampus) at 1.5 Tesla using gradient-echo EPI.
RESULTS: The post-imaging test outside the magnet yielded excellent performance on a recall test for the easy set (8.4/10 [plusminus]2) and poorer performance on the hard set (4.4/10 [plusminus]3.7). Subtractive analysis of the fMRI images obtained during encoding (easy - hard) failed to show significant activation of the hippocampus, as did subtraction of encoding minus a fixation baseline. Correlation analysis across subjects between fMRI activation and subject performance showed strong left hippocampal activation (p[lt]0.05).
CONCLUSIONS: Previous studies have had difficulty generating unequivocal hippocampal activation in straight subtraction analysis of many encoding tasks, even when composite maps across many subjects are used. However, correlation of hippocampal activity with performance data, whether in an event-related study as shown previously, or in a block design paradigm as shown here, does produce strong hippocampal activation in association with encoding success. This work emphasizes the need to consider performance data in functional studies of the hippocampus and the strong dependence of hippocampal activity on performance could explain the lack of activity observed in this region in many previous studies.
Support: NIH NS-40497.