Information Entropy Measures for Evaluation of Reliability of Deep Neural Network Results.

Deep neural networks (DNN) try to analyze given data, to come up with decisions regarding the inputs. The decision-making process of the DNN model is not entirely transparent. The confidence of the model predictions on new data fed into the network can vary. We address the question of certainty of decision making and adequacy of information capturing by DNN models during this process of decision-making. We introduce a measure called certainty index, which is based on the outputs in the most penultimate layer of DNN. In this approach, we employed iEEG (intracranial electroencephalogram) data to train and test DNN. When arriving at model predictions, the contribution of the entire information content of the input may be important. We explored the relationship between the certainty of DNN predictions and information content of the signal by estimating the sample entropy and using a heatmap of the signal. While it can be assumed that the entire sample must be utilized for arriving at the most appropriate decisions, an evaluation of DNNs from this standpoint has not been reported. We demonstrate that the robustness of the relationship between certainty index with the sample entropy, demonstrated through sample entropy-heatmap correlation, is higher than that with the original signal, indicating that the DNN focuses on information rich regions of the signal to arrive at decisions. Therefore, it can be concluded that the certainty of a decision is related to the DNN's ability to capture the information in the original signal. Our results indicate that, within its limitations, the certainty index can be used as useful tool in estimating the confidence of predictions. The certainty index appears to be related to how effectively DNN heatmaps captured the information content in the signal.

Comparison of care accessibility, costs, and quality with face-to-face and telehealth epilepsy clinic visits.

During the COVID-19 pandemic, restrictions on reimbursement for telehealth visits were lifted and this visit type was suddenly available to patients around the United States of America. Telehealth visits offer potential cost savings for patients and families, which may vary by region of the world studied. Also, aggressiveness of the care patients receive may differ, and patients or families may be more likely to choose one visit type over another based on seizure control. This is a prospective face-to-face clinic versus telehealth clinic visit comparison study involving patients with seizures, their legal guardians, and caretakers who attend clinic. We compared travel distance, work-related factors, childcare, satisfaction of care, changes in seizure medication or diagnostics tests ordered, and willingness to cancel appointments to better understand the behavioral patterns of patients, caretakers, and providers. Our results indicate that many patients and families still prefer in-person interactions with their medical providers. Patient and family satisfaction levels were equal with both visit types. No significant difference was seen in medical management between face-to-face and telehealth visits. Also, prior seizure control did not dictate the type of visit chosen. Telehealth participants were significantly more willing to cancel appointments if asked to switch to face-to-face then face-to-face participants asked to complete telehealth visits. Surprisingly, we found that patients and families choosing telehealth were not statistically more likely to be employed or take less time off work. Also, distance from home to office was not significantly shorter for participants choosing face-to-face visits. Offering a combination of telehealth and face-to-face visits appears to be the optimal strategy in caring for patients with controlled and uncontrolled seizure disorders to ensure adherence with clinic visits and satisfaction with care. Our study suggests that providers are equally willing to adjust medications or order additional diagnostic testing regardless of visit type. Patients and families may be less likely to cancel telehealth visits than face-to-face visits; this finding may translate to improved seizure control and long-term decreased cost of care.

Intracranial EEG and laser interstitial thermal therapy in MRI-negative insular and/or cingulate epilepsy: case series.

OBJECTIVE: The goal of this study was to assess the success rate and complications of stereo-electroencephalogra-phy (sEEG) and laser interstitial thermal therapy (LITT) in the treatment of nonlesional refractory epilepsy in cingulate and insular cortex. METHODS: The authors retrospectively analyzed the treatment response in 9 successive patients who underwent insular or cingulate LITT for nonlesional refractory epilepsy at their center between 2011 and 2019. Localization of seizures was based on inpatient video-EEG monitoring, neuropsychological testing, 3-T MRI, PET scan, magnetoencephalography scan, and/or ictal SPECT scan. Eight patients underwent sEEG, and 1 patient had implantation of both sEEG electrodes and subdural grids for localization of epileptogenic zones. LITT was performed in 5 insular cases (4 left and 1 right) and 3 cingulate cases (all left-sided). One patient also underwent both insular and cingulate LITT on the left side. All of the patients who underwent insular LITT as well as 2 of the 3 who underwent cingulate LITT were right-hand dominant. The patient who underwent insular plus cingulate LITT was also right-hand dominant. RESULTS: Following LITT, 67% of the patients were seizure free (Engel class I) at follow-up (mean 1.35 years, range 0.6-2.8 years). All patients responded favorably to treatment (Engel class I-III). Two patients developed small intracranial hemorrhages during the sEEG implantation that did not require surgical management. One patient developed a large intracranial hemorrhage during an insular LITT procedure that did require surgical management. That patient experienced aphasia, incoordination, and hemiparesis, which resolved with inpatient rehabilitation. No permanent neurological deficits were noted in any of the patients at last follow-up. Neuropsychological status was stable in this cohort before and after LITT. CONCLUSIONS: sEEG can be safely used to localize seizures originating from insular and cingulate cortex. LITT can successfully treat seizures arising from these deep-seated structures. The insula and cingulum should be evaluated more frequently for seizure onset zones.

Telehealth perceptions in patients with epilepsy and providers during the COVID-19 pandemic.

Coronavirus disease 2019 (COVID-19) has required novel solutions for issues that arise with social distancing. Telehealth has become one of those solutions in many clinics around the U.S. As we look beyond, the pandemic telehealth can be utilized as an important tool for clinics in the future. Patient satisfaction will most likely affect reimbursement, while provider perception will affect implementation. We see this as a valuable tool to many epilepsy clinics. The goal of our survey was to determine the perceptions and satisfaction of patients with intractable epilepsy and providers with telehealth during the COVID-19 pandemic; we surveyed patient and providers. We evaluated the first 111 patients who participated in our telehealth visits. We conducted telephone surveys with the first 68 patients who agreed to participate. We also conducted surveys by email with our providers who participated in these telehealth visits. We found that 66% of patients and 67% of providers would use a telehealth visit in the future if given the option. Review of our patients' and providers' comments provides valuable insights for building a long-term successful intractable epilepsy telehealth clinic.

Altered integrity of corpus callosum in generalized epilepsy in relation to seizure lateralization after corpus callosotomy.

OBJECTIVE: Generalized-onset seizures are usually conceptualized as engaging bilaterally distributed networks with no clear focus. However, the authors previously reported a case series demonstrating that in some patients with generalized-onset seizures, focal seizure onset could be discovered after corpus callosotomy. The corpus callosum is considered to be a major pathway for seizure generalization in this group of patients. The authors hypothesized that, in patients with generalized-onset seizures, the structure of the corpus callosum could be different between patients who have lateralized seizures and those who have nonlateralized seizures after corpus callosotomy. The authors aimed to evaluate the structural difference through statistical analysis of diffusion tensor imaging (DTI) scalars between these two groups of patients. METHODS: Thirty-two patients diagnosed with generalized-onset motor seizures and without an MRI lesion were included in this study. Among them, 16 patients developed lateralized epileptic activities after corpus callosotomy, and the remaining 16 patients continued to have nonlateralized seizures after corpus callosotomy. Presurgical DTI studies were acquired to quantify the structural integrity of the corpus callosum. RESULTS: The DTI analysis showed significant reduction of fractional anisotropy (FA) and increase in radial diffusivity (RD) in the body of the corpus callosum in the lateralized group compared with the nonlateralized group. CONCLUSIONS: The authors' findings indicate the existence of different configurations of bilateral epileptic networks in generalized epilepsy. Generalized seizures with focal onset relying on rapid spread through the corpus callosum might cause more structural damage related to demyelination in the corpus callosum, showing reduced FA and increased RD. This study suggests that presurgical DTI analysis of the corpus callosum might predict the seizure lateralization after corpus callosotomy.

Successful Intrathecal Rituximab Administration in Refractory Nonteratoma Anti-N-Methyl-D-Aspartate Receptor Encephalitis: A Case Report.

N-methyl-D-aspartate receptor (NMDA-R) antibody encephalitis is an immune-mediated disorder characterized by the presence of anti-NMDA antibody in serum and cerebrospinal fluid, with a characteristic combination of psychological and neurological signs and symptoms. The scientific knowledge pertaining to the management of anti-NMDA-R encephalitis is growing. It is important that neuroscience nurses be aware of treatments as well as the newest novel treatment options available. Early aggressive intervention is imperative to recovery. The first line of treatment often includes high-dose steroids, intravenous immunoglobulin, and therapeutic plasma exchange. Second-line therapy for refractory NMDA-R encephalitis includes intravenous rituximab and cyclophosphamide. Even with these treatments, up to 25% of patients may be left with severe deficits or have a fatal outcome. It is well known that penetration of monoclonal anti-CD20 antibody therapy (rituximab) into the cerebrospinal fluid is 0.1% of that in the serum. Therefore, efficacy of rituximab in the treatment of NMDA encephalitis may be improved by intrathecal administration in selected cases with a poor response to intravenous rituximab. We present a case of anti-NMDA-R encephalitis that was refractory to first- and second-line therapies, who responded to intrathecal rituximab, to highlight a novel treatment that may be able to prevent long-term disability and improve clinical outcomes.

Iatrogenic Exserohilum infection of the central nervous system: mycological identification and histopathological findings.

An outbreak of fungal infections has been identified in patients who received epidural injections of methylprednisolone acetate that was contaminated with environmental molds. In this report, we present the mycological and histopathological findings in an index case of Exserohilum meningitis and vasculitis in an immunocompetent patient, who received a cervical spine epidural steroid injection for chronic neck pain 1 week before the onset of fulminant meningitis with subsequent multiple brain and spinal cord infarcts. The fungus was recovered from two separate cerebrospinal fluid specimens collected before initiation of antifungal therapy and at autopsy on standard bacterial and fungal culture media. The mold was identified phenotypically as Exserohilum species. DNA sequencing targeting the internal transcribed spacer region and D1/D2 region of 28S ribosomal DNA enabled further speciation as E. rostratum. Gross examination at autopsy revealed moderate brain edema with bilateral uncal herniation and a ventriculostomy tract to the third ventricle. The brainstem, cerebellum, and right orbitofrontal cortex were soft and friable, along with hemorrhages in the cerebellar vermis and thalamus. Microscopic examination demonstrated numerous fungi with septate hyphae invading blood vessel walls and inducing acute necrotizing inflammation. The leptomeninges were diffusely infiltrated by mixed inflammatory cells along with scattered foci of fungal elements. This is the first report of iatrogenic E. rostratum meningitis in humans. This report describes the microbiological procedures and histopathological features for the identification of E. rostratum (a pigmented vascularly invasive fungi), the cause of a current nationwide outbreak of fatal fungal meningitis.

Hierarchical interaction structure of neural activities in cortical slice cultures.

Recent advances in the analysis of neuronal activities suggest that the instantaneous activity patterns can be mostly explained by considering only first-order and pairwise interactions between recorded elements, i.e., action potentials or local field potentials (LFP), and do not require higher-than-pairwise-order interactions. If generally applicable, this pairwise approach greatly simplifies the description of network interactions. However, an important question remains: are the recorded elements the units of interaction that best describe neuronal activity patterns? To explore this, we recorded spontaneous LFP peak activities in cortical organotypic cultures using planar, integrated 60-microelectrode arrays. We compared predictions obtained using a pairwise approach with those using a hierarchical approach that uses two different spatial units for describing the activity interactions: single electrodes and electrode clusters. In this hierarchical model, short-range interactions within each cluster were modeled by pairwise interactions of electrode activities and long-range interactions were modeled by pairwise interactions of cluster activities. Despite the relatively low number of parameters used, the hierarchical model provided a more accurate description of the activity patterns than the pairwise model when applied to ensembles of 10 electrodes. Furthermore, the hierarchical model was successfully applied to a larger-scale data of approximately 60 electrodes. Electrode activities within clusters were highly correlated and spatially contiguous. In contrast, long-range interactions were diffuse, suggesting the presence of higher-than-pairwise-order interactions involved in the LFP peak activities. Thus, the identification of appropriate units of interaction may allow for the successful characterization of neuronal activities in large-scale networks.

Neuronal avalanches organize as nested theta- and beta/gamma-oscillations during development of cortical layer 2/3.

Maturation of the cerebral cortex involves the spontaneous emergence of distinct patterns of neuronal synchronization, which regulate neuronal differentiation, synapse formation, and serve as a substrate for information processing. The intrinsic activity patterns that characterize the maturation of cortical layer 2/3 are poorly understood. By using microelectrode array recordings in vivo and in vitro, we show that this development is marked by the emergence of nested - and beta/gamma-oscillations that require NMDA- and GABA(A)-mediated synaptic transmission. The oscillations organized as neuronal avalanches, i.e., they were synchronized across cortical sites forming diverse and millisecond-precise spatiotemporal patterns that distributed in sizes according to a power law with a slope of -1.5. The correspondence between nested oscillations and neuronal avalanches required activation of the dopamine D(1) receptor. We suggest that the repetitive formation of neuronal avalanches provides an intrinsic template for the selective linking of external inputs to developing superficial layers.