Abstracts

Rationale: Parietal lobe comprises large areas of association cortex extensively connected to other lobes. We hypothesize such widespread network projections can produce erroneous localization and misinterpretation of scalp EEG and clinical semiology in parietal lobe epilepsies (PLE). We investigated the reliability of electroclinical features in PLE as compared to frontal (FLE) and temporal lobe epilepsy (TLE). Methods: Fifty consecutive pharmacoresistant focal epilepsy patients seizure free for ? 12 months (median 23 months; IQR 14.75-35.75) following resections limited to frontal, temporal or parietal lobe were identified. Interictal EEG discharges (iiEEG), single representative ictal EEG (iEEG) and seizure video were extracted from archived files of scalp video EEG monitoring (median 5 days; IQR 4-6). Two raters (R1 R2) blinded to other clinical data independently reviewed the EEG. Seizure videos were then presented and raters formulated their clinical impression (CI) based on all findings to either PLE/TLE/FLE/non-specified. Patients with more than one iEEG and previous neurosurgery were excluded. Results: 16 PLE, 17 FLE and 17 TLE patients were studied. Groups did not differ significantly in demographics, absence of iiEEG (5 FLE, 2 TLE, and 3 PLE), epilepsy onset or presence of normal MRI (4 FLE, 1 PLE). Interobserver agreement was substantial across different aspects of the iiEEG (? .79-.96). PLE iiEEG showed greatest scatter outside the lobe of origin: unilateral iiEEG in anterotemporal region (electrodes (ele) Sp/FT, n=7); unilateral iiEEG in parietal region (P ele, n=6); unilateral frontopolar or bilateral frontal iiEEG (Fp ele, n=3); iiEEG central region (Cz ele, n=2) in 2; and contralateral iiEEG in anterotemporal region (n=5). FLE group did not show contralateral iiEEG and were generalized (bifrontal maximal, n=5), frontopolar (n=3), parasagittal frontal (n=2) and central (n=3). TLE iiEEG showed most consistent lobar localization: ipsilateral anterotemporal (n=10), ipsilateral temporal (T7/8 ele, n=6), ipsilateral frontopolar (n=4) and contralateral anterotemporal (n=3). There was a significant difference in number of iiEEG populations between PLE (median 2; IQR 1-2.75) and FLE (median 1; IQR 0-1) (p=.032). Interobserver agreement of iEEG was substantial (? .61; p<0.005). Raters were able to localize and lateralize iEEG to the epileptogenic region in all TLE (R1,R2), (R1-11/17; R2-15/17) FLE and (R1-7/16; R2-10/16) PLE patients (p=.002; p=.011). CI agreement for all epilepsy types was moderate (? .58; p<.005). In patients whereby raters confidently categorized their CI to one epilepsy type, PLE was often misidentified (? .52; p=.002): R1 correct/specified FLE 8/8, TLE 12/14, PLE 4/8; p=.034); R2 (FLE 14/14, TLE 13/13, PLE 6/10; p=.002). TLE (R1 correctly identified/rest 14/17; R2 13/17) are more often accurately localized compared with PLE (R1 4/16; R2 6/16) (p=.032; p=.013).