Small temporal pole encephaloceles: A treatable cause of “lesion negative” temporal lobe epilepsy


Address correspondence to John S. Archer, Department of Neurology, Austin Health, Melbourne, Australia. E-mail:


Epilepsy due to encephaloceles of the temporal pole may be an under recognized, treatable cause of refractory temporal lobe epilepsy (TLE). We describe three adult patients initially labeled “lesion negative” TLE. In all, video–electroencephalography (EEG) revealed ictal theta in the left temporal region and positron emission tomography (PET) showed temporal lobe hypometabolism, but neuropsychology revealed preserved verbal memory. Close inspection of structural magnetic resonance imaging (MRI) suggested subtle abnormalities at the tip of the left temporal lobe. High resolution computed tomography (CT) confirmed bony defects in the inner table of the skull. 3T MRI with fine coronal and sagittal slices indicated cerebrospinal fluid (CSF) and brain tissue protruding into the defects. All proceeded to resection of the temporal tip and became seizure free. Patients with “lesion negative” TLE should have careful review of images covering the temporal pole. If encephalocele is suspected, further imaging with high-resolution CT and MRI can be helpful. Temporal polar resection, sparing mesial structures, appears to be curative.

Encephaloceles are defects of the skull and meninges, which are usually congenital. They are often associated with defects of the cerebral parenchyma, and are a recognized, if rare, cause of seizures. We describe three patients, initially thought to have “lesion negative” refractory left temporal lobe epilepsy, in whom a standard temporal lobectomy would have had a high risk of producing verbal memory deficits. In each case, the recognition of a small anterior temporal encephalocele by close inspection of imaging and further high-resolution studies altered the surgical strategy.

Description of Cases

Case 1: 39-year-old woman

Complex partial seizures began at age 24 years, manifesting as impaired consciousness, right hand posturing, and head version. At the time of assessment the patient was having multiple daily attacks, frequently with secondary generalization. Four anticonvulsants in various doses and combinations had failed to control seizures. The patient was born premature at 30 weeks of gestation, but early childhood development and schooling were otherwise unremarkable. Febrile seizures had not occurred. Video–electroencephalography (EEG) monitoring captured eight typical seizures, all originating from the anterior left temporal region (F7, T1, T3). Prior cerebral magnetic resonance imaging (MRI), including coronal cuts through the hippocampi, had been reported as normal, with formal volumetric and T2 relaxometry studies confirming normal hippocampal values and no left–right asymmetry. Cerebral fluorodeoxyglucose positron emission tomography (FDG-PET) demonstrated left temporal lobe hypometabolism. Neuropsychologic assessment of this right-handed subject confirmed normal verbal memory. Therefore, the electroclinical diagnosis was “lesion negative” left temporal lobe epilepsy (TLE). There was reluctance to proceed to left temporal lobectomy due to the likely negative impact on verbal memory.

High-resolution coronal MRI showed a subtle asymmetry in the anterobasal part of the left temporal pole close to the entorhinal cortex. A small rounded area of gray matter protruded inferiorly, with no associated signal abnormality. The floor of the middle cranial fossa appeared slightly scalloped at this site, but a bony defect could not be visualized on MRI. For this reason, a high-resolution CT of the floor of the middle cranial fossa was performed, demonstrating a well-defined defect in the inner table of the greater wing of left sphenoid bone, consistent with an encephalocele (Fig. 1).

Figure 1.

Case 1. (A, B) Coronal T1 and T2 images demonstrating subtle lobulated appearance of left temporal pole (arrowed). (C) Computed tomography (CT) with fine coronal cuts confirming defect in inner skull table of middle cranial fossa (circled).

The patient underwent left temporal pole resection, sparing mesial temporal structures, so as to preserve verbal memory. The anterior 3 cm of the middle and inferior temporal gyri were resected. The medial aspect of the temporal pole was clearly abnormal and protruding through the floor of the anteromedial middle fossa, where there was evidence of scalloping of the bone. Subsequent histopathology of the resected specimen showed prominent white matter neurons and molecular layer gliosis, but no frank dysplasia.

The patient remains seizure free 22 months postoperatively, with no change in verbal memory or other cognitive parameters detected on follow-up formal neuropsychological testing.

Case 2: 26-year-old woman

This right-handed female university graduate began having seizures at age 12 characterized by a rising abdominal sensation, inability to speak, and reduced awareness. Seizures continued weekly with occasional secondary generalization, despite trials of seven anticonvulsants in various doses and combinations. Febrile convulsions had not occurred. Three previous MRI examinations at different institutions had been reported as normal. Video-EEG monitoring confirmed complex partial seizures of left temporal onset and FDG-PET showed left anterior temporal hypometabolism. Neuropsychology assessment demonstrated intact verbal memory, but a mild naming deficit.

The previous imaging was reassessed and increased T2-weighted signal was noted in the greater wing of the left sphenoid bone, immediately anterior to the left temporal pole. T1-weighted images showed low signal in the same region. To clarify the nature of the abnormality, high-resolution MRI using surface coils and high-resolution CT of the middle cranial fossa were obtained, confirming a small left anterior temporal encephalocele with a small well-defined lobulated defect in the inner table of the sphenoid bone (supplementary Fig. 1).

The patient underwent resection of the anterior 2.5 cm of the left temporal pole, including the encephalocele. The dura mater was attenuated around the encephalocele and deficient at the base of the bony defect. The resected encephalocele specimen included subarachnoid and superficial neocortical tissue, the latter showing gliosis and some architectural distortion but no dysplasia. The more proximal temporal neocortex showed some molecular layer gliosis. The patient remains seizure free 20 months postoperatively.

Case 3: 26-year-old man

This right-handed university student began having complex partial seizures at age 20, characterized by déjà vu and confusion, followed by feelings of depression. At the time of evaluation these were occurring twice a week, despite previous trials of four anticonvulsants. Two generalized tonic–clonic seizures had occurred. There was a history of perinatal fetal bradycardia, requiring forceps-assisted delivery, but developmental milestones and school performance had been normal. There was a history of possible febrile convulsion at age 2, and a family history of self-limited childhood seizures in a brother and cousin. Previous cerebral MRI had not revealed definitive abnormalities.

Video-EEG monitoring captured five complex partial seizures, all arising from the left temporal region (supplementary Fig. 2). FDG-PET showed left temporal hypometabolism, and ictal single-photon emission computed tomography (SPECT) showed anterior left temporal hyperperfusion. Neuropsychology assessment confirmed normal cognition including verbal memory. The electroclinical diagnosis was of “lesion negative” left temporal lobe epilepsy (TLE). Once again, concerns about the impact of a standard left anterior temporal lobectomy on verbal memory prompted closer examination of the imaging and further studies.

Figure 2.

Case 3. (A) Coronal T2 through temporal pole with area of encephalocele circled. Unusual morphology just evident. (B) Axial T1 image acquired along “hippocampal plane,” with lesion area circled. (C) In this patient helpful additional imaging included fine sagittal slices (3T, fat saturation sequence shown), revealing pedunculated lesion from tip of left temporal lobe (arrowed). (D) Computed tomography (CT) with fine coronal slices confirmed a defect in the floor of the inner table of the skull, at the tip of the middle cranial fossa.

Structural MRI at 1.5T using surface coils, showed an apparent abnormality at the tip of the left temporal lobe, with tissue of signal characteristics similar to gray matter projecting anteroinferiorly through the floor of the middle cranial fossa. A 3T MRI study clearly demonstrated a well-defined lesion at the tip of his temporal pole. CT with fine coronal cuts showed bony thinning in the area of this lesion, consistent with a left temporal pole encephalocele (Fig. 2).

At operation, the dura at the tip of the middle cranial fossa was pitted, with the brain protruding into the pits and adherent in places. The anterior 2.5 cm of the left temporal pole were resected and the dural pits covered with “dura-seal.” Histopathologic examination confirmed protuberant gliotic cerebral tissue with no underlying dysplastic features. The patient remains seizure free 12 months postoperatively. Postoperative neuropsychologic reassessment showed no change, and he has returned to university studies.


Temporal lobe encephaloceles are a rare cause of TLE. Of the dozen or so previously reported cases associated with TLE, the majority have been noted incidentally at the time of temporal lobe resection, or upon surgical exploration of a “soft tissue mass” in the mastoid region (Hyson et al., 1984; Rosenbaum et al., 1985; Leblanc et al., 1991; Yang et al., 2004). In the cases we report, concerns about the impact of a standard left anterior temporal lobectomy on verbal memory, led to a further careful reexamination of structural imaging of the temporal region, and discovery of quite small lesions.

Although there can at times be clear temporal pole parenchymal abnormalities (Vargas et al., 2008; Whiting et al., 1990), the three patients we report had subtle radiologic features and had previously been characterized as having “lesion negative” TLE. This reflects the difficulty in detecting small structural abnormalities that are iso-intense with gray matter at the tip of the temporal lobe, as well as the limited ability of MRI to display bony abnormalities. Detection was possible in retrospect on some of the standard “epilepsy protocol” MRI scans, but was facilitated by additional sequences, and high-resolution CT. Indeed, the great success of MRI in detecting epileptogenic lesions, compared to x-ray CT, may have led to under recognition of this rare lesion.

Imaging approaches we found helpful were those that increased the MRI signal-to-noise ratio (surface coils, higher field strength), increased resolution (512 × 512 in plane resolution), or both. High-resolution acquisitions in the coronal and sagittal planes were also beneficial. CT with fine coronal reformatted images, showing scalloping or defects of the inner table of the skull provided important confirmatory evidence. Therefore, we would suggest that all cases of “lesion negative” TLE should have the standard MR images reviewed specifically for subtle abnormalities at the temporal tip. If suspicion of encephalocele remains, additional imaging as suggested earlier may help.

These three individuals were seizure free following restricted resections of the temporal pole with follow-up of 12–22 months at the time of writing. This suggests that the encephalocele or immediately adjacent cortex was the epileptogenic region, and that more extensive resections, including mesial temporal structures, were not necessary. Furthermore, sparing of mesial temporal structures has allowed these individuals to continue to have normal verbal memory postoperatively.

The treatment of patients with “lesion negative” TLE is a challenge. Their postoperative surgical outcome is inferior to those with hippocampal sclerosis or foreign tissue lesions (McIntosh et al., 2004). When seizures arise from the dominant temporal lobe, postoperative verbal memory impairment may occur following standard anterior temporal lobectomy, making the risk–benefit ratio less favorable. The discovery of a cryptic lesion such as an encephalocele markedly alters the surgical approach and risk–benefit assessment.


We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.


None of the authors has any conflict of interest to disclose.