Abstracts

Rationale: The complex relation between the interictal markers of focal epilepsy (spikes) and the epileptogenic zone (seizures) is incompletely understood. The shape (morphology) of a spike encodes the spatiotemporal pattern of depolarization-hyperpolarization of the responsible neuronal pool; aspects of the clinical behavior of the underlying epilepsy may be implicit in this pattern. For instance, the complexity of spike morphology may relate to the clinical propensity for seizures (Kalamangalam GP, 2007. Epilepsia 48(s6):2.198; Frost JD et al, 1992. Epilepsia 33:531-536). We explored spike morphology as a marker for ictal onset lateralization and network behavior, in unilateral temporal lobe epilepsy (TLE) with bilateral interictal spikes of comparable abundance. Methods: Continuous scalp EEG data from 6 adults with proven TLE undergoing long-term monitoring were filtered in the passband 0.5-70 Hz, sampled at 200 Hz, and reviewed digitally (Nihon Kohden, Inc.). Epochs of artifact-free EEG during light sleep were selected, totaling 2-4 hours per subject. Spikes were detected semi-automatically by high-pass filters and amplitude thresholds using WAVE_CLUS (Quian Quiroga R et al, 2004. Neur Comput 16:1661-87), and clustered into morphological subgroups by the k-means algorithm with L1 norm (MATLAB Statistics Toolbox). The tightness of clustering provided an index of morphology variability (MV) within spike populations. Results were compared to the TLE syndrome features of individual patients. Results: Right or left total spike counts varied from 160-1700, with spike count ratios (SCRs) between hemispheres of < 1:2. In two patients, greater spike counts occurred opposite the side of ictal onset. In two patients with right TLE (right:left SCRs=226:167; 1711:1330) there was greater right-sided spike MV (right-left difference >10%). In two patients with left TLE (right:left SCRs=708:443; 809:1425), but with rapid left to right seizure spread, right-left spike MVs were equivalent (right-left difference < 2%). In the fifth patient with left hippocampal sclerosis but only right-sided seizures (right:left SCR=930:1184), right MV was significantly (>10%) larger than left MV. In the sixth patient (right:left SCR=611:782) with bilateral independent seizures with no contralateral spread but more frequent left-sided seizures, left spike MV was significantly (>10%) larger than the right. Conclusions: Morphological analysis of spikes offers insight into right versus left epileptogenicity in TLE. Such insight is independent of simpler methods such as raw spike counts (Krendl R et al, 2008. Neurology 71:413-418), which would have been misleading in two of our patients. Large (>10%) asymmetries in spike MV indicate either a unilateral, or significantly asymmetric, seizure propensity. Equivalent MVs (<2% right-left difference) indicate a tendency for rapid contralateral involvement during seizures. We hypothesize that equivalent MVs also characterize bilateral TLE patients who seize equally from the two hemispheres. A larger study is under way to confirm these preliminary results.