Abstracts

Rationale: EEG spikes-sharp waves (SSW) are focal interictal epileptiform transients (FIET) predictive of partial epilepsy, and their differentiation from FIET not predictive of epilepsy (atypical wicket spikes: aWS) has emphasized expert polygraphic waveform analysis. Conventional criteria of differentiation incorporate relative amplitude against background. Lack of reference indifferent to background has limited reliability of EEG amplitude measurement. It was shown (Matsuo, 2012: www.AESnet.org) that polygraphic channel overlay (PGCO), when applied to 23 common reference (CR) derivations, could segment SSW phases, base-peak-trough-wave (BPTW), by true phase reversal. Voltage difference between maximal deflections of opposite signs is magnitude of equivalent current dipole (VDd: Matsuo, 2016: www.AESnet.org), and indifferent to choice of CR. It was shown in randomly chosen 121 FIET that P- and W-amplitude (VDd-P and VDd-W) correlated with probability of epilepsy. Objective of this study was to develop polygraphic EEG voltage display without reference bias. Methods: 23 head-surface electrode locations included 10-20 System placements, and zygomatic (ZY) and mastoid (MS) electrode pairs. 23 PGCO consisted of replacing CR with each of 23 head-surface electrodes. Operation of comprehensive CR replacement (CCRR) utilized Persyst 13 edit functions, involving 529 (23 X 23) derivations and generating polygraphic display. Method was applied to preliminary examination of effect of display time constant (TC) to VDd-P and VDd-W. Results: Exemplar SSW of right basal frontotemporal origin is illustrated in EEG display for conventional clinical diagnostic review, common average reference (CAR) derivations at TC 0.16 s (FIG-a), and parallel PGCO-CCRR display at TC 3 s with polarity reversed (CR-PGCO: FIG-b). At P, CAR derivations revealed maximal negativity at ZY2 and suggested maximal positivity at vertex region. Waveform of SSW exactly corresponds between basal temporal derivations between FIG-a and –b. When measured in CR-PGCO at cursors P and W, VDd-P and VDd-W were invariant. Invariance of VDd-P also helps confirm that downward PGCO deflections were maximal at vertex region. Higher coherence of transient of interest and relative attenuation of EEG background enhance confidence in detailed waveform examination. Low amplitude notches (cursors c1 and c2) are subtle, but suggest secondary left temporal lobe involvement. TC change from 0.16 to 3 s enhanced VDd-P and VDd-W by 2 % and 11 %, respectively, and delayed peak of W-phase. Random samples of 22 FIET (11 SSW and 11 aWS) were first screened in PGCO-CCRR. Lack of differentiation of W-segment from normal EEG background and/or unstable baseline excluded 10 FIET from VDd-W measurement. Long TC enhanced VDd with relatively wide variation (TABLE: standard deviation in parentheses). Conclusions: PGCO-CCRR was designed as clinical neurophysiological tool for examination of EEG geometry. It neutralizes contribution of reference to EEG display, and can add reliability to assessment of SSW waveform against background, either physiological or non-physiological. Solid angle theorem applies to interpretation of localizing features.