Abstracts

RATIONALE: To assess the relationships between fMRI activation asymmetry in the anterior hippocampus during a memory task, hippocampal volume asymmetry, and Wada memory asymmetry.
Background: Wada memory asymmetry (WMA) has been shown to predict side of seizure focus and to correlate with hippocampal volume asymmetry (HVA). In a recent study we showed that fMRI activation in the anterior hippocampus during a memory task predicted side of seizure focus. In the present study, we examined the correlation between fMRI hippocampal activation asymmetry (HAA), HVA, and WMA in a further effort to validate the use of fMRI for testing memory in epilepsy surgery candidates.
METHODS: 38 consecutive non-retarded epilepsy surgery candidates underwent fMRI at 1.5 Tesla while performing an episodic memory encoding task. For the memory task, patients were asked to decide if color photographic pictures were indoor or outdoor scenes. In the control task, they were to decide if two halves of randomly retiled scenes matched. This combination of tasks produces bi-hippocampal activation in normals. The fMRI HAA was computed (L-R/L+R) for each epilepsy patient by counting the number of significantly activated voxels within an anterior hippocampal region of interest derived from a probabilistic stereotaxic map of the normal hippocampus. WMA was calculated for each patient by subtracting the number of real objects correctly recognized during Wada testing in the inject left/test right) condition from the number recognized in the inject right/test left condition. HVA (L-R/L+R) was calculated for each patient by tracing the entire hippocampus and subiculum, excluding the parahippocampal region. In each case a positive score indicates greater hippocampal activation (fMRI), memory representation (Wada), or hippocampal volume on the left side. The HAAs and WMAs were correlated with each other (n= 38) and with the HVAs (n=16).
RESULTS: Anterior hippocampal activation asymmetries on fMRI (HAA) and memory asymmetries from Wada (WMA) were both significantly correlated with hippocampal volume asymmetries (r = .64, p = .004 and r = .58, p= .02, respectively). In addition, a modest correlation was found between WMA and HAA (r = .39, p = .007).
CONCLUSIONS: The correlation between asymmetries derived from anterior hippocampal activation on fMRI, hippocampal volumes, and Wada memory testing provide further evidence of the validity of fMRI for assessing hippocampal status in epilepsy surgery candidates.
Support: NINDS RO1 53929 and NINDS RO1 NS33576