An overview of machine learning and deep learning techniques for predicting epileptic seizures.

Epilepsy is a neurological disorder (the third most common, following stroke and migraines). A key aspect of its diagnosis is the presence of seizures that occur without a known cause and the potential for new seizures to occur. Machine learning has shown potential as a cost-effective alternative for rapid diagnosis. In this study, we review the current state of machine learning in the detection and prediction of epileptic seizures. The objective of this study is to portray the existing machine learning methods for seizure prediction. Internet bibliographical searches were conducted to identify relevant literature on the topic. Through cross-referencing from key articles, additional references were obtained to provide a comprehensive overview of the techniques. As the aim of this paper aims is not a pure bibliographical review of the subject, the publications here cited have been selected among many others based on their number of citations. To implement accurate diagnostic and treatment tools, it is necessary to achieve a balance between prediction time, sensitivity, and specificity. This balance can be achieved using deep learning algorithms. The best performance and results are often achieved by combining multiple techniques and features, but this approach can also increase computational requirements.

Immunohistochemical distribution of secretagogin in the mouse brain.

Introduction: Calcium is essential for the correct functioning of the central nervous system, and calcium-binding proteins help to finely regulate its concentration. Whereas some calcium-binding proteins such as calmodulin are ubiquitous and are present in many cell types, others such as calbindin, calretinin, and parvalbumin are expressed in specific neuronal populations. Secretagogin belongs to this latter group and its distribution throughout the brain is only partially known. In the present work, the distribution of secretagogin-immunopositive cells was studied in the entire brain of healthy adult mice. Methods: Adult male C57BL/DBA mice aged between 5 and 7 months were used. Their whole brain was sectioned and used for immunohistochemistry. Specific neural populations were observed in different zones and nuclei identified according to Paxinos mouse brain atlas. Results: Labelled cells were found with a Golgi-like staining, allowing an excellent characterization of their dendritic and axonal arborizations. Many secretagogin-positive cells were observed along different encephalic regions, especially in the olfactory bulb, basal ganglia, and hypothalamus. Immunostained populations were very heterogenous in both size and distribution, as some nuclei presented labelling in their entire extension, but in others, only scattered cells were present. Discussion: Secretagogin can provide a more complete vision of calcium-buffering mechanisms in the brain, and can be a useful neuronal marker in different brain areas for specific populations.

Biohybrid restoration of the hippocampal loop re-establishes the non-seizing state in anin vitromodel of limbic seizures.

Objective. The compromise of the hippocampal loop is a hallmark of mesial temporal lobe epilepsy (MTLE), the most frequent epileptic syndrome in the adult population and the most often refractory to medical therapy. Hippocampal sclerosis is found in >50% of drug-refractory MTLE patients and primarily involves the CA1, consequently disrupting the hippocampal output to the entorhinal cortex (EC). Closed-loop deep brain stimulation is the latest frontier to improve drug-refractory MTLE; however, current approaches do not restore the functional connectivity of the hippocampal loop, they are designed by trial-and-error and heavily rely on seizure detection or prediction algorithms. The objective of this study is to evaluate the anti-ictogenic efficacy and robustness of an artificial bridge restoring the dialog between hippocampus and EC.Approach. In mouse hippocampus-EC slices treated with 4-aminopyridine and in which the Schaffer Collaterals are severed, we established an artificial bridge between hippocampus and EC wherein interictal discharges originating in the CA3 triggered stimulation of the subiculum so to entrain EC networks. Combining quantification of ictal activity with tools from information theory, we addressed the efficacy of the bridge in controlling ictogenesis and in restoring the functional connectivity of the hippocampal loop.Main results. The bridge significantly decreased or even prevented ictal activity and proved robust to failure; when operating at 100% of its efficiency (i.e., delivering a pulse upon each interictal event), it recovered the functional connectivity of the hippocampal loop to a degree similar to what measured in the intact circuitry. The efficacy and robustness of the bridge stem in mirroring the adaptive properties of the CA3, which acts as biological neuromodulator.Significance. This work is the first stepping stone toward a paradigm shift in the conceptual design of stimulation devices for epilepsy treatment, from function control to functional restoration of the salient brain circuits.

An affordable implantable vagus nerve stimulator system for use in animal research.

In this research, a vagus nerve stimulator has been developed and miniaturized for use in epilepsy research. The board contains all the components necessary for its operation during the standard duration of the experiments, being possible to control it once implanted and even being able to reuse it. The VNS system has been designed for rodents since the VNS devices available for human are not only too large for laboratory animals, but also too expensive. With this solution the expenditure on materials made by laboratories is greatly reduced and bioethical considerations were kept in mind. The system was validated in hamsters. This article is part of the theme issue 'Advanced neurotechnologies: translating innovation for health and well-being'.

Mimicking CA3 Temporal Dynamics Controls Limbic Ictogenesis.

Mesial temporal lobe epilepsy (MTLE) is the most common partial complex epilepsy in adults and the most unresponsive to medications. Electrical deep brain stimulation (DBS) of the hippocampus has proved effective in controlling seizures in epileptic rodents and in drug-refractory MTLE patients. However, current DBS paradigms implement arbitrary fixed-frequency or patterned stimuli, disregarding the temporal profile of brain electrical activity. The latter, herein included hippocampal spontaneous firing, has been shown to follow lognormal temporal dynamics. Here, we present a novel paradigm to devise DBS protocols based on stimulation patterns fashioned as a surrogate brain signal. We focus on the interictal activity originating in the hippocampal subfield CA3, which has been shown to be anti-ictogenic. Using 4-aminopyridine-treated hippocampus-cortex slices coupled to microelectrode array, we pursue three specific aims: (1) address whether lognormal temporal dynamics can describe the CA3-driven interictal pattern, (2) explore the possibility of restoring the non-seizing state by mimicking the temporal dynamics of this anti-ictogenic pattern with electrical stimulation, and (3) compare the performance of the CA3-surrogate against periodic stimulation. We show that the CA3-driven interictal activity follows lognormal temporal dynamics. Further, electrical stimulation fashioned as a surrogate interictal pattern exhibits similar efficacy but uses less pulses than periodic stimulation. Our results support the possibility of mimicking the temporal dynamics of relevant brain signals as a straightforward DBS strategy to ameliorate drug-refractory epilepsy. Further, they herald a paradigm shift in neuromodulation, wherein a compromised brain signal can be recreated by the appropriate stimuli distribution to bypass trial-and-error studies and attain physiologically meaningful DBS operating modes.

Inferior Colliculus Transcriptome After Status Epilepticus in the Genetically Audiogenic Seizure-Prone Hamster GASH/Sal.

The Genetic Audiogenic Seizure Hamster from Salamanca (GASH/Sal), an animal model of reflex epilepsy, exhibits generalized tonic-clonic seizures in response to loud sound with the epileptogenic focus localized in the inferior colliculus (IC). Ictal events in seizure-prone strains cause gene deregulation in the epileptogenic focus, which can provide insights into the epileptogenic mechanisms. Thus, the present study aimed to determine the expression profile of key genes in the IC of the GASH/Sal after the status epilepticus. For such purpose, we used RNA-Seq to perform a comparative study between the IC transcriptome of GASH/Sal and that of control hamsters both subjected to loud sound stimulation. After filtering for normalization and gene selection, a total of 36 genes were declared differentially expressed from the RNA-seq analysis in the IC. A set of differentially expressed genes were validated by RT-qPCR showing significant differentially expression between GASH/Sal hamsters and Syrian control hamsters. The confirmed differentially expressed genes were classified on ontological categories associated with epileptogenic events similar to those produced by generalized tonic seizures in humans. Subsequently, based on the result of metabolomics, we found the interleukin-4 and 13-signaling, and nucleoside transport as presumably altered routes in the GASH/Sal model. This research suggests that seizures in GASH/Sal hamsters are generated by multiple molecular substrates, which activate biological processes, molecular processes, cellular components and metabolic pathways associated with epileptogenic events similar to those produced by tonic seizures in humans. Therefore, our study supports the use of the GASH/Sal as a valuable animal model for epilepsy research, toward establishing correlations with human epilepsy and searching new biomarkers of epileptogenesis.

Smart Buildings IoT Networks Accuracy Evolution Prediction to Improve Their Reliability Using a Lotka-Volterra Ecosystem Model.

Internet of Things (IoT) is the paradigm that has largely contributed to the development of smart buildings in our society. This technology makes it possible to monitor all aspects of the smart building and to improve its operation. One of the main challenges encountered by IoT networks is that the the data they collect may be unreliable since IoT devices can lose accuracy for several reasons (sensor wear, sensor aging, poorly constructed buildings, etc.). The aim of our work is to study the evolution of IoT networks over time in smart buildings. The hypothesis we have tested is that, by amplifying the Lotka-Volterra equations as a community of living organisms (an ecosystem model), the reliability of the system and its components can be predicted. This model comprises a set of differential equations that describe the relationship between an IoT network and multiple IoT devices. Based on the Lotka-Volterra model, in this article, we propose a model in which the predators are the non-precision IoT devices and the prey are the precision IoT devices. Furthermore, a third species is introduced, the maintenance staff, which will impact the interaction between both species, helping the prey to survive within the ecosystem. This is the first Lotka-Volterra model that is applied in the field of IoT. Our work establishes a proof of concept in the field and opens a wide spectrum of applications for biology models to be applied in IoT.