Abstracts

Rationale:
The fornix and cingulum bundle are the primary afferent and efferent pathway of the hippocampus. Adult unilateral TLE patients demonstrate differences in white matter organization in both of these areas, with increased mean diffusivity and reduced fractional anisotropy (FA) bilaterally in the fornix and cingulum bundle. However, to our knowledge, these pathways have not been investigated in pediatric focal epilepsy. This is important to examine the neural basis of memory and potentially inform surgical procedures, as these white matter tracts may be affected during surgical resection. We determined the extent to which degree of white matter organization in the memory network is different in focal epilepsy and typically developing (TD) controls for 1) adults and 2) children. We expect reduced white matter organization (decreased FA) in the fornix and cingulum bundle for both epilepsy groups.
Method:
46 participants completed the study; Adults: 21 patients with focal epilepsy (8 L, 9 R, 3 bilateral, 1 undetermined), 14 TD controls; Pediatric: 7 patients (2 L, 4 R, 1 undetermined), 4 TD controls. All patients were referred to the Clinical Epilepsy Section at NIH for presurgical evaluation. For each participant, an MPRAGE T1 and  45 diffusion-weighted images were acquired on a 3T scanner (AP-PA dual phase-encoded, b-values of 0, 200, 500 and 1100) were acquired on a 3T scanner. Diffusion data were preprocessed using TORTOISE, correcting for motion, EPI and eddy current distortions. We used the AFNI toolbox FATCAT to estimate diffusion tensors (DT) and perform probabilistic tractography with minimum thresholds of: 0.2 FA, 60-degree angles, and 20mm tract length. Each MPRAGE was segmented in Freesurfer6.0 with the Desikan-Killiany atlas. We selected 11 regions-of-interest (ROI) associated with the memory network (parahippocampal, entorhinal, fusiform, amygdala, thalamus, nucleus accumbens, isthmus/posterior cingulate (PCC), and frontal regions) to compute average DT metrics of white matter bundles connecting all ROIs from the ipsilateral left or right hippocampus. Mixed Model ANOVAs in SPSS were conducted separately for left and right hippocampal connectivity to these ipsilateral ROIs andcompared between focal epilepsy and TD groups.
Results:
We found a main effect of the memory network ROI separately for each hippocampus in both the adult and pediatric group (p’s< 0.05), with the strongest FA value overall to ipsilateral PCC. For only the left hippocampus for adults, there was also a trend for an ROI x group interaction (p=0.059), with reduced FA to the ipsilateral isthmus of the cingulate, thalamus, and lateral orbital frontal cortex (latter two connections at the trend level). With an additional model including side of epilepsy focus, we found a main effect of focus for the left hippocampus only: the adult left focal group had reduced FA compared to TDs (p=0.03). Limited number of pediatric patients precluded side of epilepsy focus analysis.
Conclusion:
For both age groups, the strongest connection was between each hippocampus and the ipsilateral PCC, an important target for the cingulum. For adults with left focal epilepsy, there was also overall reduced FA to these memory network ROIs ipsilateral to the seizure focus.  Adult patients with focal epilepsy potentially show reduced FA from the left hippocampus to specific ipsilateral targets related to the fornix (e.g., thalamus) and cingulum (e.g., orbital frontal), as hypothesized based on prior studies. This small pediatric group did not demonstrate these same differences, potentially representing a developmental effect. 
Funding:
:This work was supported by the Susan Spencer Clinical Research Fellowship and K23 NS093152 from NINDS to LNS and NINDS Division of Intramural Research.