Abstracts

Rationale: The piriform region consists of the frontal lobe piriform cortex (FPC), temporal lobe piriform cortex (TPC), and endopiriform nucleus (EPN). These sites are particularly amenable to kindling and other experimental epileptogenesis (as the "area tempestas"), and appear also to have seizure modulatory properties. The EPN is evident in human cadaver studies, but not in current clinical MRI. The FPC and TPC have rarely been studied in temporal lobe epilepsy (TLE) patients. Methods: We identified 8 TLE subjects who had unilateral temporal lobe ictal EEG onsets ipsilateral to hippocampal atrophy on clinical MRI. With IRB-approved consenting, we acquired 7 Tesla brain MRI with T2-weighted turbo spin echo images (0.25 x 0.25 x 1.2 mm3) in contiguous coronal slices of 11 healthy subjects and the TLE patients. We generated a manual segmentation protocol for FPC, TPC, and EPN, based on landmarks derived from coronal histological sections in a human brain atlas. Results: In each healthy and TLE subject the 7T images provided the requisite landmarks for segmentation of the FPC, TPC, and EPN bilaterally (including the critical demarcation of the EPN from the adjacent amygdala, which we could not achieve with clinical 3T MRI). FPC: Healthy group mean volumes: right FPC 118 mm3 (standard deviation 33 mm3) and left FPC 104 mm3 (SD 22 mm3). TLE group mean volumes: right FPC 78 mm3 (SD 19 mm3) and left FPC 74 mm3 (SD 20 mm3). TPC: Healthy group mean volumes: right TPC 134 mm3 (SD 33 mm3) and left TPC 124 mm3 (SD 40 mm3). TLE group mean volumes: right TPC 93 mm3 (SD 23 mm3) and left TPC 91 mm3 (SD 24 mm3). EPN: Healthy group mean volumes: right EPN 76 mm3 (SD 22 mm3) and left EPN 68 mm3 (SD 21 mm3). TLE group mean volumes: right EPN 50 mm3 (SD 8.9 mm3) and left EPN 43 mm3 (SD 13 mm3). In individual TLE patients, the FPC, TPC, and EPN ipsilateral to the sclerotic hippocampus were not relatively smaller than the contralateral homologues, compared to the mean asymmetry in healthy subjects. Conclusions: Increased spatial and contrast resolution of 7T MRI supported manual segmentation of piriform subregions. The FPC, TPC, and EPN appeared bilaterally atrophic in a relatively symmetric fashion in individual TLE patients, despite hippocampal volume asymmetry (with hippocampal atrophy ipsilateral to ictal EEG onsets), on comparisons with piriform subregional volumes in healthy subjects. Bilateral piriform regional atrophy may be due to seizure-related injury; etiological roles of piriform sites in human epileptogenesis or ictogenesis are possible. Further study will be required to determine whether bilateral piriform regional atrophy is present in a larger group of TLE-hippocampal sclerosis patients, and whether this finding is present in non-refractory TLE and in other partial epilepsies. 7T MRI can provide improved piriform subregional anatomical correlation for functional imaging studies. This study was supported by NIH P41 RR008079 & P30 NS057091, and the Keck Foundation.