A 54 year-old right-handed female underwent invasive EEG (iEEG) as part of a pre-surgical evaluation for drug-resistant seizures. A high-resolution brain MRI was normal. PET demonstrated left temporal hypometabolism. Interictal scalp EEG revealed left> right bitemporal epileptiform discharges (75:25 ratio). Independent left (n=2) and right (n=2) temporal onset was noted on scalp EEG (scEEG). The semiology demonstrated a sudden blank stare, impaired consciousness, and manual automatisms for 40 seconds following which she was agitated and "couldn't talk" (left temporal onset on scEEG) in addition to a second semiology, infrequent by report, manifest as a blank stare, impaired consciousness, and left arm posturing. She was told she was a non-surgical candidate. At the urgings of her aging parents a second opinion was sought. After comprehensive risk-benefit counseling, iEEG was opted (see below).
Figure: (A) iEEG with left regional ictal fast activity in the depth at seizure onset (B) periodic complexes in the right depth and fast activity in the strips at termination. L/RST=left/right superior temporal; L/RTA=left/right temporal anterior; L/RTI=left/right temporal inferior; L/RTP=left/right temporal parietal; L/RTD=left/right temporal depths. Sensitivity 50 uv; filter settings of 1-70 Hz.
Invasive EEG is frequently used when discordant localizing information is present on the non-invasive evaluation. Common neocortical low-voltage fast activity such as beta and gamma frequencies may be attenuated by scEEG and the true seizure onset zone may be falsely localized from extratemporal sources. Invasive electrodes may detect seizures earlier than scalp recording (see above), though imply a greater degree of complexity. Our patient underwent iEEG with bitemporal depth electrodes and subdural strips. All seizures (n=8) demonstrated left temporal onset (A) with a “switch” to the right strips/depths (B) in 2 seizures that were associated with left sided dystonic posturing. A “switch” from one hemisphere to another may affect semiology and falsely lateralize the scEEG to suggest bitemporal epilepsy as in our patient. Propagated frequencies are often slower than ictal onset frequencies (1) facilitating scEEG detection of the propagated semiology and “pseudo-ictal” onset. EEG patterns during seizures termination are poorly predictive of the surgical outcome (2). Our patient underwent a left temporal lobectomy and is seizure free at 6 months post-operation.
Adapted from Newsletter, American Clinical Neurophysiology Society, Volume 3, Issue 3 (in press) with permission.
© 2014 Epilepsy.com. All rights reserved.