Classification of glioblastoma versus primary central nervous system lymphoma using convolutional neural networks.

A subset of primary central nervous system lymphomas (PCNSL) are difficult to distinguish from glioblastoma multiforme (GBM) on magnetic resonance imaging (MRI). We developed a convolutional neural network (CNN) to distinguish these tumors on contrast-enhanced T1-weighted images. Preoperative brain tumor MRIs were retrospectively collected among 320 patients with either GBM (n = 160) and PCNSL (n = 160) from two academic institutions. The individual images from these MRIs consisted of a training set (n = 1894 GBM and 1245 PCNSL), a validation set (n = 339 GBM; 202 PCNSL), and a testing set (99 GBM and 108 PCNSL). Three CNNs using the EfficientNetB4 architecture were evaluated. To increase the size of the training set and minimize overfitting, random flips and changes to color were performed on the training set. Our transfer learning approach (with image augmentation and 292 epochs) yielded an AUC of 0.94 (95% CI: 0.91-0.97) for GBM and an AUC of 0.95 (95% CI: 0.92-0.98) for PCNL. In the second case (not augmented and 137 epochs), the images were augmented prior to training. The area under the curve for GBM was 0.92 (95% CI: 0.88-0.96) for GBM and an AUC of 0.94 (95% CI: 0.91-0.97) for PCNSL. For the last case (augmented, Gaussian noise and 238 epochs) the AUC for GBM was 0.93 (95% CI: 0.89-0.96) and an AUC 0.93 (95% CI = 0.89-0.96) for PCNSL. Even with a relatively small dataset, our transfer learning approach demonstrated CNNs may provide accurate diagnostic information to assist radiologists in distinguishing PCNSL and GBM.

Insult to injury: transient encephalopathy in a brain-injured adolescent.

We report on an adolescent with prior brain injury who responded to abrupt drug changes and an electrolyte imbalance with a significant alteration in the level of consciousness. In children with prior brain injury, paediatricians should consider the concept of a 'fragile brain' with limited 'cognitive reserve' and increased susceptibility to relatively minor toxic-metabolic insults. Psychoactive and central nervous system drugs should be managed with prudence in these patients.

Electrocorticography for seizure foci mapping in epilepsy surgery.

Patients with refractory focal epilepsy are thoroughly evaluated to identify an area of cortex that, if removed or disconnected, will lead to seizure freedom. Clinical semiology, neuroimaging, and scalp electroencephalogram provide an approximation of this area, whereas intracranial recording may permit a more precise localization and investigation of a selected cortical area. Intraoperative electrocorticography delineates the irritative zone, and subdural electrode implantation also permits cortical stimulation of eloquent areas. Intraoperative electrocorticography rarely captures spontaneous seizures and may be influenced by the effect of anesthetic drugs, and the correlation between complete resection of the irritative zone and postsurgical seizure outcome is unclear. Extraoperative monitoring is often superior to intraoperative electrocorticography but may also be associated with more risk of adverse events. Further development of ultrahigh-density electrode arrays is providing novel insights into the role of microseizures and high-frequency oscillations on ictogenesis and epileptogenesis.

Brain functional networks in syndromic and non-syndromic autism: a graph theoretical study of EEG connectivity.

BACKGROUND: Graph theory has been recently introduced to characterize complex brain networks, making it highly suitable to investigate altered connectivity in neurologic disorders. A current model proposes autism spectrum disorder (ASD) as a developmental disconnection syndrome, supported by converging evidence in both non-syndromic and syndromic ASD. However, the effects of abnormal connectivity on network properties have not been well studied, particularly in syndromic ASD. To close this gap, brain functional networks of electroencephalographic (EEG) connectivity were studied through graph measures in patients with Tuberous Sclerosis Complex (TSC), a disorder with a high prevalence of ASD, as well as in patients with non-syndromic ASD. METHODS: EEG data were collected from TSC patients with ASD (n = 14) and without ASD (n = 29), from patients with non-syndromic ASD (n = 16), and from controls (n = 46). First, EEG connectivity was characterized by the mean coherence, the ratio of inter- over intra-hemispheric coherence and the ratio of long- over short-range coherence. Next, graph measures of the functional networks were computed and a resilience analysis was conducted. To distinguish effects related to ASD from those related to TSC, a two-way analysis of covariance (ANCOVA) was applied, using age as a covariate. RESULTS: Analysis of network properties revealed differences specific to TSC and ASD, and these differences were very consistent across subgroups. In TSC, both with and without a concurrent diagnosis of ASD, mean coherence, global efficiency, and clustering coefficient were decreased and the average path length was increased. These findings indicate an altered network topology. In ASD, both with and without a concurrent diagnosis of TSC, decreased long- over short-range coherence and markedly increased network resilience were found. CONCLUSIONS: The altered network topology in TSC represents a functional correlate of structural abnormalities and may play a role in the pathogenesis of neurological deficits. The increased resilience in ASD may reflect an excessively degenerate network with local overconnection and decreased functional specialization. This joint study of TSC and ASD networks provides a unique window to common neurobiological mechanisms in autism.

Dacrystic seizures: demographic, semiologic, and etiologic insights from a multicenter study in long-term video-EEG monitoring units.

PURPOSE: To provide an estimate of the frequency of dacrystic seizures in video-electroencephalography (EEG) long-term monitoring units of tertiary referral epilepsy centers and to describe the clinical presentation of dacrystic seizures in relationship to the underlying etiology. METHODS: We screened clinical records and video-EEG reports for the diagnosis of dacrystic seizures of all patients admitted for video-EEG long-term monitoring at five epilepsy referral centers in the United States and Germany. Patients with a potential diagnosis of dacrystic seizures were identified, and their clinical charts and video-EEG recordings were reviewed. We included only patients with: (1) stereotyped lacrimation, sobbing, grimacing, yelling, or sad facial expression; (2) long-term video-EEG recordings (at least 12 h); and (3) at least one brain magnetic resonance imaging (MRI) study. KEY FINDINGS: Nine patients (four female) with dacrystic seizures were identified. Dacrystic seizures were identified in 0.06-0.53% of the patients admitted for long-term video-EEG monitoring depending on the specific center. Considering our study population as a whole, the frequency was 0.13%. The presence of dacrystic seizures without other accompanying clinical features was found in only one patient. Gelastic seizures accompanied dacrystic seizures in five cases, and a hypothalamic hamartoma was found in all of these five patients. The underlying etiology in the four patients with dacrystic seizures without gelastic seizures was left mesial temporal sclerosis (three patients) and a frontal glioblastoma (one patient). All patients had a difficult-to-control epilepsy as demonstrated by the following: (1) at least three different antiepileptic drugs were tried in each patient, (2) epilepsy was well controlled with antiepileptic drugs in only two patients, (3) six patients were considered for epilepsy surgery and three of them underwent a surgical/radiosurgical or radioablative procedure. Regarding outcome, antiepileptic drugs alone achieved seizure freedom in two patients and did not change seizure frequency in another patient. Radiosurgery led to moderately good seizure control in one patient and did not improve seizure control in another patient. Three patients were or are being considered for epilepsy surgery on last follow-up. One patient remains seizure free 3 years after epilepsy surgery. SIGNIFICANCE: Dacrystic seizures are a rare but clinically relevant finding during video-EEG monitoring. Our data show that when the patient has dacrystic and gelastic seizures, the cause is a hypothalamic hamartoma. In contrast, when dacrystic seizures are not accompanied by gelastic seizures the underlying lesion is most commonly located in the temporal cortex.