Modern computation based on the von Neumann architecture is today a mature cutting-edge science. In the Von Neumann architecture, processing and memory units are implemented as separate blocks interchanging data intensively and continuously. This data transfer is responsible for a large part of the power consumption. The next generation computer technology is expected to solve problems at the exascale with 1018 calculations each second. Even though these future computers will be incredibly powerful, if they are based on von Neumann type architectures, they will consume between 20 and 30 megawatts of power and will not have intrinsic physically built-in capabilities to learn or deal with complex data as our brain does. These needs can be addressed by neuromorphic computing systems which are inspired by the biological concepts of the human brain. This new generation of computers has the potential to be used for the storage and processing of large amounts of digital information with much lower power consumption than conventional processors. Among their potential future applications, an important niche is moving the control from data centers to edge devices. The aim of this Roadmap is to present a snapshot of the present state of neuromorphic technology and provide an opinion on the challenges and opportunities that the future holds in the major areas of neuromorphic technology, namely materials, devices, neuromorphic circuits, neuromorphic algorithms, applications, and ethics. The Roadmap is a collection of perspectives where leading researchers in the neuromorphic community provide their own view about the current state and the future challenges for each research area. We hope that this Roadmap will be a useful resource by providing a concise yet comprehensive introduction to readers outside this field, for those who are just entering the field, as well as providing future perspectives for those who are well established in the neuromorphic computing community.

Whether new neurons are added in the postnatal cerebral cortex is still debated. Here, we report that the meninges of perinatal mice contain a population of neurogenic progenitors formed during embryonic development that migrate to the caudal cortex and differentiate into Satb2 neurons in cortical layers II-IV. The resulting neurons are electrically functional and integrated into local microcircuits. Single-cell RNA sequencing identified meningeal cells with distinct transcriptome signatures characteristic of (1) neurogenic radial glia-like cells (resembling neural stem cells in the SVZ), (2) neuronal cells, and (3) a cell type with an intermediate phenotype, possibly representing radial glia-like meningeal cells differentiating to neuronal cells. Thus, we have identified a pool of embryonically derived radial glia-like cells present in the meninges that migrate and differentiate into functional neurons in the neonatal cerebral cortex.

Deep-brain electrical or transcranial magnetic stimulation may represent a therapeutic tool for controlling seizures in patients presenting with epileptic disorders resistant to antiepileptic drugs. In keeping with this clinical evidence, we have reported that repetitive electrical stimuli delivered at approximately 1 Hz in mouse hippocampus-entorhinal cortex (EC) slices depress the EC ability to generate ictal activity induced by the application of 4-aminopyridine (4AP) or Mg 2+ -free medium (Barbarosie, M., Avoli, M., 1997. CA3-driven hippocampalentorhinal loop controls rather than sustains in vitro limbic seizures. J. Neurosci. 17,[9308][9309][9310][9311][9312][9313][9314]. Here, we confirmed a similar control mechanism in rat brain slices analyzed with field potential recordings during 4AP (50 MM) treatment. In addition, we used intrinsic optical signal (IOS) recordings to quantify the intensity and spatial characteristics of this inhibitory influence. IOSs reflect the changes in light transmittance throughout the entire extent of the slice, and are thus reliable markers of limbic network epileptiform synchronization. First, we found that in the presence of 4AP, the IOS increases, induced by a train of electrical stimuli (10 Hz for 1 s) or by recurrent, single-shock stimulation delivered at 0.05 Hz in the deep EC layers, are reduced in intensity and area size by low-frequency (1 Hz), repetitive stimulation of the subiculum; these effects were observed in all limbic areas contained in the slice. Second, by testing the effects induced by repetitive subicular stimulation at 0.2-10 Hz, we identified maximal efficacy when repetitive stimuli are delivered at 1 Hz. Finally, we discovered that similar, but slightly less pronounced, inhibitory effects occur when repetitive stimuli at 1 Hz are delivered in the EC, suggesting that the reduction of IOSs seen during repetitive stimulation is pathway dependent as well as activity dependent. Thus, the activation of limbic networks at low frequency reduces the intensity and spatial extent of the IOS changes that accompany ictal synchronization in an in vitro slice preparation. This conclusion supports the view that repetitive stimulation may represent a potential therapeutic tool for controlling seizures in patients with pharmacoresistant epileptic disorders. D 2004 Elsevier Inc. All rights reserved.

Laser-evoked potentials (LEPs) are brain responses to laser radiant heat pulses and reflect the activation of Aδ nociceptors. LEPs are to date the reference standard technique for studying nociceptive pathway function in patients with neuropathic pain. To find out whether LEPs also provide a useful neurophysiological tool for assessing antinociceptive drug efficacy, in this doubleblind placebo-controlled study we measured changes induced by the analgesic tramadol on LEPs in 12 healthy subjects. We found that tramadol decreased the amplitude of LEPs, whereas placebo left LEPs unchanged. The opioid antagonist naloxone partially reversed the tramadol-induced LEP amplitude decrease. We conclude that LEPs may be reliably used in clinical practice and research for assessing the efficacy of antinociceptive drugs.

Purpose of review-Neurosteroids are a family of compounds synthesized directly in the brain by transforming cholesterol into pregnenolone, which is then converted to compounds such as allopregnanolone and allotetrahydrodeoxycorticosterone. In view of their ability to modulate neurotransmission, neurosteroids may influence the clinical course of epileptic disorders. In this review, we highlight two emerging properties of neurosteroids, that is, their anticonvulsant and antiepileptogenic activities.Recent findings-It has been shown that fluctuations in neurosteroid synthesis, such as those seen in response to stress or during the ovarian cycle, determine an increase in seizure threshold. Moreover, increased neurosteroid synthesis, presumably occurring in glial cells during epileptogenesis, delays the appearance of recurrent spontaneous seizures in an animal model of temporal lobe epilepsy; such an effect may be due to augmented tonic γ-aminobutyric acid type A receptor-mediated inhibition. Finally, clinical trials with ganaxolone, an allopregnanolone analogue, have demonstrated beneficial effects in pharmacoresistant epileptic patients, whereas finasteride -which interferes with neurosteroid synthesis -facilitates seizures in catamenial epilepsy.Summary-The overall evidence suggests that neurosteroids may represent a novel therapeutic strategy in epileptic disorders and a future perspective to control epileptogenicity.

This paper presents a new methodology for automatically learning an optimal neurostimulation strategy for the treatment of epilepsy. The technical challenge is to automatically modulate neurostimulation parameters, as a function of the observed EEG signal, so as to minimize the frequency and duration of seizures. The methodology leverages recent techniques from the machine learning literature, in particular the reinforcement learning paradigm, to formalize this optimization problem. We present an algorithm which is able to automatically learn an adaptive neurostimulation strategy directly from labeled training data acquired from animal brain tissues. Our results suggest that this methodology can be used to automatically find a stimulation strategy which effectively reduces the incidence of seizures, while also minimizing the amount of stimulation applied. This work highlights the crucial role that modern machine learning techniques can play in the optimization of treatment strategies for patients with chronic disorders such as epilepsy.

Temporal lobe epilepsy (TLE) is a chronic epileptic disorder involving the hippocampal formation. Details on the interactions between the hippocampus proper and parahippocampal networks during ictogenesis remain, however, unclear. In addition, recent findings have shown that epileptic limbic networks maintained in vitro are paradoxically less responsive than non-epileptic control (NEC) tissue to application of the convulsant drug 4-aminopyridine (4AP). Field potential recordings allowed us to establish here the effects of 4AP in brain slices obtained from NEC and pilocarpinetreated epileptic rats; these slices included the hippocampus and parahippocampal areas such as entorhinal and perirhinal cortices and the amygdala. First, we found that both types of tissue generate epileptiform discharges with similar electrographic characteristics. Further investigation showed that generation of robust ictal-like discharges in the epileptic rat tissue is (i) favored by decreased hippocampal output (ii) reinforced by EC-subiculum interactions and (iii) predominantly driven by amygdala networks. We propose that a functional switch to alternative synaptic routes may promote network hyperexcitability in the epileptic limbic system.

Deep brain stimulation (DBS) is a promising tool for treating drug-resistant epileptic patients. Currently, the most common approach is fixed-frequency stimulation (periodic pacing) by means of stimulating devices that operate under open-loop control. However, a drawback of this DBS strategy is the impossibility of tailoring a personalized treatment, which also limits the optimization of the stimulating apparatus. Here, we propose a novel DBS methodology based on a closed-loop control strategy, developed by exploiting statistical machine learning techniques, in which stimulation parameters are adapted to the current neural activity thus allowing for seizure suppression that is fine-tuned on the individual scale (adaptive stimulation). By means of field potential recording from adult rat hippocampus–entorhinal cortex (EC) slices treated with the convulsant drug 4-aminopyridine we determined the effectiveness of this approach compared to low-frequency periodic pacing, and found that the closed-loop stimulation strategy: (i) has similar efficacy as low-frequency periodic pacing in suppressing ictal-like events but (ii) is more efficient than periodic pacing in that it requires less electrical pulses. We also provide evidence that the closed-loop stimulation strategy can alternatively be employed to tune the frequency of a periodic pacing strategy. Our findings indicate that the adaptive stimulation strategy may represent a novel, promising approach to DBS for individually-tailored epilepsy treatment.