Abstracts

Rationale: Reliable pre-surgical guides are necessary to predict and minimize the adverse impact of temporal lobe surgery on memory functioning. Clinical fMRI provides reliable, noninvasive language mapping. Localization of memory function using fMRI is not established in pediatric epilepsy. We developed a paired association learning fMRI task to elicit hippocampal activation (HA) and compared activation in children with focal epilepsy to typically developing (TD) controls. We investigated hippocampal lateralization at the individual level, necessary to aid in presurgical evaluation, and the functional significance of this lateralization. Methods: Twenty-two children with focal epilepsy (mean age=13.6 2.7; 9 left, 8 right, 1 bilateral, 4 undetermined) and 17 TD controls (mean age=12.5 2.2) participated in two EPI BOLD 3T block design paired association learning task (word pair stimuli) with learning and recall blocks. Imaging processing and analyses were conducted in SPM12. Image segmentation and normalization were done using subjects' high-resolution T1 images with the CAT12 Toolbox in SPM12, then applying those parameters to EPI images. A field map was used to correct distortion. Hippocampal region of interest (ROI) was based on the Wake Forest PickAtlas. We calculated a bootstrapped hippocampal ROI lateralization index (LI) using the LI Toolbox (using a common cut-off of 0.2). We examined the relationship between hippocampal LI and memory ability via the Children's Memory Scale (CMS) Verbal and Visual Indices with Pearson's correlation. Results: Patients and controls demonstrated no significant differences in HA LI for learning or recall. Controls had primarily left-lateralized HA, although the mean LI values were not strongly lateralizing (learning: LI=0.20; recall: LI=0.27). Categorically, the variability was evident, particularly for recall where almost half had bilateral HA (learning: 50% left, 25% bilateral, 19% right; recall: 50% left, 44% bilateral, 6% right). Patients had left-lateralized activation for the learning trial and bilateral activation for recall (learning: LI=0.29; recall: LI=0.12). Categorically, the majority were left-lateralized, albeit distribution was also variable, particularly for recall (learning: 59% left, 32% bilateral, 9% right; recall: 55% left, 14% bilateral, 32% right). Left and right focal groups demonstrated no significant difference in HA LI for learning or recall. HA LI during verbal memory fMRI was significantly correlated with CMS Verbal Immediate Index (r=0.45, p=0.01)\, but no other Verbal or Visual Indices.  Conclusions: Controls and patients with focal epilepsy exhibited weakly left-lateralized or bilateral HA on a memory fMRI task, evident both via continuous and categorical analyses of lateralization. Stronger left-lateralized HA on task was related to  better verbal memory and is pattern expected for adults. Thus, children may engage both hippocampi, with hippocampal specialization for specific types of material (left-verbal, right-visual) occurring later in development. Funding: This work was supported by NINDS Division of Intramural Research, Susan S. Spencer Clinical Research Training Fellowship [to LNS] and Avery Translational Research Career Development Program Award [through the Clinical and Translational Science Institute at Children's National to LNS].