Left mesial temporal FLAIR lesion with nocturnal seizure clusters
MRI suggests left mesial temporal sclerosis with focal cortical dysplasia, and scalp EEG agrees on the left temporal lobe. Why explore the insula, orbitofrontal cortex and pulvinar before treating it?
Presentation
A right-handed man in his early 30s with drug-resistant focal epilepsy that began in his late 20s. There were no perinatal complications, febrile seizures, significant head injury or CNS infection. His father probably has untreated temporal lobe seizures.
Other history includes anxiety and depression (possibly bipolar disorder) and past substance use. His neurological examination is normal. He currently takes levetiracetam, gabapentin and cenobamate; he had not tried other antiseizure medications before these.
Semiology
As reported: an aura of impending doom and déjà vu, followed by whole-body paresthesias (he can’t recall which side starts), loss of awareness, and at times bilateral tonic-clonic seizures. Post-ictal confusion follows. Seizures last under 5 minutes and come in clusters of 5 to 10 in a day, every 2 to 3 months; the longest seizure-free interval has been about 3.5 months. Sleep deprivation and binge drinking are triggers.
As recorded on video-EEG: eight clinical seizures over two days, several arising from sleep. In the fullest examples he woke and pressed the event button, then had stiffening of the right arm, chewing automatisms, forceful blinking and heavy breathing or grunting, and could not answer questions or follow commands. Others were limited to chewing, or to staring with heavy breathing. Each lasted 40 seconds to 1 minute 40 seconds, and he could answer questions as soon as each one ended.
Pause and think Which features lateralize, which localize, and what doesn't fit a purely mesial temporal onset?
Right arm stiffening (contralateral tonic posturing) lateralizes to the left hemisphere. The aura of fear and déjà vu, the oroalimentary automatisms and the brief, quickly resolving impairment point to the mesial temporal lobe.
Several features are harder to explain with a purely mesial onset. Whole-body paresthesias suggest the insula or somatosensory cortex. Heavy breathing, grunting and chewing can be generated by insulo-opercular cortex. The phase I team also read the semiology as spreading from the left mesial temporal to the left mesial frontal region. Each of these is a reason to sample beyond the mesial temporal lobe.
Scalp EEG
Interictal: no epileptiform abnormalities.
Ictal: 16 seizures with a left temporal onset were captured, 8 electroclinical and 8 electrographic only. Each began with rhythmic 5–7 Hz activity, maximal at T3, that spread to the left frontal region as 3–4 Hz activity and ended with background attenuation.
Imaging
MRI: non-expansile, non-enhancing FLAIR hyperintensity throughout the left mesial temporal lobe, especially the amygdala, hippocampus and parahippocampal gyrus, with blurring of the gray-white junction. The findings suggest focal cortical dysplasia with probable mesial temporal sclerosis; a low-grade glioma was considered less likely because the lesion is not expansile.
There were also scattered punctate FLAIR foci in the cerebral white matter, the left dorsolateral midbrain, the right medial thalamus and the cerebellum. A vasculitic process was raised, and an atypical demyelinating process was not excluded.
PET, MEG, ictal SPECT and fMRI: not performed.
Neuropsychology
The profile was neither clearly lateralizing nor localizing, with variable verbal comprehension, perceptual reasoning, processing speed and verbal memory. Variable processing speed, reduced confrontation naming and poor recall of a word list and geometric figures may suggest subcortical and dominant temporal involvement. The poor visual memory may relate to the right medial thalamic lesion.
Because story recall was intact, resection of dominant mesial temporal structures carries some risk of post-operative cognitive decline. fMRI or a Wada test was recommended to confirm language dominance and refine that risk.
Hypothesis
| Lateralization | Localization | |
|---|---|---|
| Semiology | Left | Left mesial temporal, spreading to left mesial frontal |
| Interictal EEG | — | — |
| Ictal EEG | Left | Left temporal |
| MRI | Left | Left amygdala, hippocampus and parahippocampal gyrus |
| Neuropsychology | Non-lateralizing | Non-localizing |
| PET, MEG, fMRI, Wada | Not done | Not done |
The conference hypothesis was a left mesial temporal epileptogenic zone due to mesial temporal sclerosis with focal cortical dysplasia. The plan was fMRI or Wada testing for surgical risk, and an SEEG admission to explore the left mesial temporal lobe ahead of resection.
Pause and think The MRI lesion, the ictal EEG and the lateralizing signs all agree. Why not go straight to surgery?
The lesion makes the mesial temporal lobe the obvious candidate, but a FLAIR abnormality marks a lesion, not necessarily the whole epileptogenic zone. Scalp EEG can’t separate amygdala, hippocampus, temporal pole and insula, and the semiology contains insulo-opercular and orbitofrontal features. In the language-dominant hemisphere, with intact story recall to protect, the team needs to know whether a limited mesial treatment is enough, or whether the temporal pole, insula or orbitofrontal cortex is part of the onset.
Implantation plan
Thirteen orthogonal depth electrodes were placed in the left hemisphere: six in the temporal lobe (A’, B’, C’, E’, F’, I’), four sampling the insula from different entry points (T’, Q’, R’, S’), two in the orbitofrontal cortex (OR’, J’) and one in the pulvinar (U’).
| Lead | Target | Grid | Why it was placed |
|---|---|---|---|
| A' | Amygdala | E9 | Amygdala, the leading candidate given the aura of impending doom and déjà vu and the FLAIR signal centred on the amygdala. |
| B' | Hippocampal head | E10 | Hippocampal head, within the MRI abnormality and the core target for a mesial temporal sclerosis hypothesis. |
| C' | Entorhinal cortex and hippocampal body | F8 | Entorhinal cortex and hippocampal body. Defines how far back the abnormality and seizure onset extend, which sets the posterior limit of treatment. |
| E' | Entorhinal and perirhinal cortex | E11 | Anterior entorhinal and perirhinal cortex, to separate rhinal from hippocampal or amygdalar onset. |
| F' | Lingual and fusiform gyri | F10 | Lingual and fusiform gyri. Samples basal and posterior mesial spread in the dominant hemisphere and allows stimulation mapping of basal temporal language cortex. |
| I' | Temporal pole (medial to lateral) | D12 | Temporal pole, medial to lateral. Distinguishes temporopolar from amygdala-hippocampal onset, which changes the operation. |
| J' | Medial posterior orbitofrontal cortex (via inferior insula) | C10 | Medial posterior orbitofrontal cortex, crossing the inferior insula and superior temporal pole on the way in. Tests the orbitofrontal side of a temporal plus network. |
| T' | Inferior insula | D9 | Inferior insula. Chewing, heavy breathing and grunting can arise from insulo-opercular cortex, so insular onset or early spread has to be excluded. |
| U' | Pulvinar (via Heschl's gyrus) | E7 | Pulvinar, through Heschl's gyrus and the planum temporale. Shows whether the seizure network reaches a thalamic neuromodulation target. |
| OR' | Medial and mid orbitofrontal cortex | B9 | Medial and mid orbitofrontal cortex, a second orbitofrontal sample for the temporal plus question. |
| Q' | Superior insula (via pars triangularis) | C8 | Anterior superior insula through the pars triangularis. |
| R' | Superior insula (via frontal operculum) | D7 | Superior insula through the frontal operculum; its lateral contacts sample primary motor cortex. |
| S' | Superior/posterior insula (via S1) | E7 | Superior/posterior insula through primary sensory cortex. Reported whole-body paresthesias raise posterior insular involvement. |
References
- Mallela AN, Hect J, Reedy E, et al. A semiology-driven framework for explorations of temporal lobe epilepsy. Epilepsia. 2025. Link
For education only. These cases show how one epilepsy team reasoned about individual patients; they are not clinical guidelines. Cases are de-identified, and some details may be changed to protect patient privacy.