Frontiers in Neuroscience

Journal: Frontiers in Neuroscience

Abbreviation

Front Neurosci

Publisher

Frontiers Media

ISSN

1662-453X
1662-4548

Show all metadata

Search Results

Publications1 - 10 of 144

Multiple Single-Unit Long-Term Tracking on Organotypic Hippocampal Slices Using High-Density Microelectrode Arrays
Gong, Wei; Senčar, Jure; Bakkum, Douglas J.; et al. (2016)
Frontiers in Neuroscience
A novel system to cultivate and record from organotypic brain slices directly on high-density microelectrode arrays (HD-MEA) was developed. This system allows for continuous recording of electrical activity of specific individual neurons at high spatial resolution while monitoring at the same time, neuronal network activity. For the first time, the electrical activity patterns of single neurons and the corresponding neuronal network in an organotypic hippocampal slice culture were studied during several consecutive weeks at daily intervals. An unsupervised iterative spike-sorting algorithm, based on PCA and k-means clustering, was developed to assign the activities to the single units. Spike-triggered average extracellular waveforms of an action potential recorded across neighboring electrodes, termed “footprints” of single-units were generated and tracked over weeks. The developed system offers the potential to study chronic impacts of drugs or genetic modifications on individual neurons in slice preparations over extended times.
Neuropeptides as Primary Mediators of Brain Circuit Connectivity
Guillaumin, Mathilde C.C.; Burdakov, Denis (2021)
Frontiers in Neuroscience
Across sleep and wakefulness, brain function requires inter-neuronal interactions lasting beyond seconds. Yet, most studies of neural circuit connectivity focus on millisecond-scale interactions mediated by the classic fast transmitters, GABA and glutamate. In contrast, neural circuit roles of the largest transmitter family in the brain–the slow-acting peptide transmitters–remain relatively overlooked, or described as “modulatory.” Neuropeptides may efficiently implement sustained neural circuit connectivity, since they are not rapidly removed from the extracellular space, and their prolonged action does not require continuous presynaptic firing. From this perspective, we review actions of evolutionarily-conserved neuropeptides made by brain-wide-projecting hypothalamic neurons, focusing on lateral hypothalamus (LH) neuropeptides essential for stable consciousness: the orexins/hypocretins. Action potential-dependent orexin release inside and outside the hypothalamus evokes slow postsynaptic excitation. This excitation does not arise from modulation of classic neurotransmission, but involves direct action of orexins on their specific G-protein coupled receptors (GPCRs) coupled to ion channels. While millisecond-scale, GABA/glutamate connectivity within the LH may not be strong, re-assessing LH microcircuits from the peptidergic viewpoint is consistent with slow local microcircuits. The sustained actions of neuropeptides on neuronal membrane potential may enable core brain functions, such as temporal integration and the creation of lasting permissive signals that act as “eligibility traces” for context-dependent information routing and plasticity. The slowness of neuropeptides has unique advantages for efficient neuronal processing and feedback control of consciousness.
Research topic: neuromorphic engineering systems and applications. A snapshot of neuromorphic systems engineering
Delbrück, Tobias; Van Schaik, André; Hasler, Jennifer (2014)
Frontiers in Neuroscience
The 14 papers in this research topic were solicited primarily from attendees to the two most important hands-on workshops in neuromorphic engineering: the Telluride Neuromorphic Cognition Engineering Workshop (www.ine-web.org) and the Capo Caccia Cognitive Neuromorphic Engineering Workshop (capocaccia.ethz.ch). The papers show the results of feasibility studies of new concepts, as well as neuromorphic systems that have been constructed from more established neuromorphic technologies. Five papers exploit neuromorphic dynamic vision sensor (DVS) events that mimic the asynchronous and sparse spikes on biology's optic nerve fiber (Delbruck and Lang, 2013; O'Connor et al., 2013; Rea et al., 2013; Brandli et al., 2014; Camunas-Mesa et al., 2014; Clady et al., 2014). Two papers are on the hot topic (based on largest number of views) of event-driven computation in deep belief networks (DBNs) (O'Connor et al., 2013; Neftci et al., 2014). Two papers use floating gate technology for neuromorphic analog circuits (Gupta and Markan, 2014; Marr and Hasler, 2014). The collection is rounded out by papers on central pattern generators (Ambroise et al., 2013), neural fields for cognitive architectures (Sandamirskaya, 2014), sound perception (Coath et al., 2014), polychronous spiking networks (Wang et al., 2014), and automatic parameter tuning for large network simulations (Carlson et al., 2014).
Optimizing the Energy Consumption of Spiking Neural Networks for Neuromorphic Applications
Sorbaro, Martino; Liu, Qian; Bortone, Massimo; et al. (2020)
Frontiers in Neuroscience
In the last few years, spiking neural networks (SNNs) have been demonstrated to perform on par with regular convolutional neural networks. Several works have proposed methods to convert a pre-trained CNN to a Spiking CNN without a significant sacrifice of performance. We demonstrate first that quantization-aware training of CNNs leads to better accuracy in SNNs. One of the benefits of converting CNNs to spiking CNNs is to leverage the sparse computation of SNNs and consequently perform equivalent computation at a lower energy consumption. Here we propose an optimization strategy to train efficient spiking networks with lower energy consumption, while maintaining similar accuracy levels. We demonstrate results on the MNIST-DVS and CIFAR-10 datasets.
The role of the midcingulate cortex in monitoring others' decisions
Apps, Matthew A.J.; Lockwood, Patricia L.; Balsters, Joshua H. (2013)
Frontiers in Neuroscience
A plethora of research has implicated the cingulate cortex in the processing of social information (i.e., processing elicited by, about, and directed toward others) and reward-related information that guides decision-making. However, it is often overlooked that there is variability in the cytoarchitectonic properties and anatomical connections across the cingulate cortex, which is indicative of functional variability. Here we review evidence from lesion, single-unit recording and functional imaging studies. Taken together, these support the claim that the processing of information that has the greatest influence on social behavior can be localized to the gyral surface of the midcingulate cortex (MCCg). We propose that the MCCg is engaged when predicting and monitoring the outcomes of decisions during social interactions. In particular, the MCCg processes statistical information that tracks the extent to which the outcomes of decisions meet goals when interacting with others. We provide a novel framework for the computational mechanisms that underpin such social information processing in the MCCg. This framework provides testable hypotheses for the social deficits displayed in autism spectrum disorders and psychopathy. Primates live in social environments that require individuals to understand the complex behavior of conspecifics. A plethora of research implicates the dorsal Anterior Cingulate Cortex (ACC) as playing a vital role in processing “social” information (i.e., processing elicited by, about, or directed toward others) (Amodio and Frith, 2006; Somerville et al., 2006; Rudebeck et al., 2008; Behrens et al., 2009; Apps et al., 2012; Hillman and Bilkey, 2012). Indeed, individuals with lesions to the ACC display social deficits so severe that they are said to have “acquired sociopathy” (Anderson et al., 1999). However, the ACC is also engaged by rewards (Doya, 2008), attention and salience (Davis et al., 2005), conflict, and during decision-making (Botvinick et al., 1999; Botvinick, 2007) which are inherently non-social processes. How can the same region be engaged by such a distinct set of processes? It is often overlooked that the area labeled as “ACC” by functional imaging research comprises multiple sub-regions, each with distinct cytoarchitecture and anatomical connections (Vogt et al., 1995; Palomero-Gallagher et al., 2008; Beckmann et al., 2009). Thus, some of the processes that have been reported to elicit an ACC response may in fact be localized to distinct sub-regions. Here, we draw attention to anatomical tracer, neurophysiology, lesion and neuroimaging studies investigating the anatomical and functional properties of the dorsal ACC. Taken together this research highlights one sub-region which processes information about the outcomes of others' decisions and about the decisions made by others during social interactions. This region in fact lies on the gyral surface of the midcingulate cortex (MCCg) and not in the anatomically defined ACC. We contend that whilst the sulcal (MCCs) and gyral (MCCg) regions of the MCC can be differentiated in terms of processing first-person and social information respectively, the two areas process similar information about rewards that guide decision-making. By drawing parallels between the role of the MCCs in processing first-person rewards, and that of the MCCg in processing rewards in social contexts, we provide a new framework for investigating the contribution of the MCC to social decision-making.
Editorial: Bio-inspired Audio Processing, Models and Systems
Liu, Shih-Chii; Harris, John G.; Elhilali, Mounya; et al. (2019)
Frontiers in Neuroscience
Three-Dimensional Iron Oxide Nanoparticle-Based Contrast-Enhanced Magnetic Resonance Imaging for Characterization of Cerebral Arteriogenesis in the Mouse Neocortex
de Bortoli, Till; Boehm-Sturm, Philipp; Koch, Stefan P.; et al. (2021)
Frontiers in Neuroscience
Purpose: Subsurface blood vessels in the cerebral cortex have been identified as a bottleneck in cerebral perfusion with the potential for collateral remodeling. However, valid techniques for non-invasive, longitudinal characterization of neocortical microvessels are still lacking. In this study, we validated contrast-enhanced magnetic resonance imaging (CE-MRI) for in vivo characterization of vascular changes in a model of spontaneous collateral outgrowth following chronic cerebral hypoperfusion. Methods: C57BL/6J mice were randomly assigned to unilateral internal carotid artery occlusion or sham surgery and after 21 days, CE-MRI based on T2*-weighted imaging was performed using ultra-small superparamagnetic iron oxide nanoparticles to obtain subtraction angiographies and steady-state cerebral blood volume (ss-CBV) maps. First pass dynamic susceptibility contrast MRI (DSC-MRI) was performed for internal validation of ss-CBV. Further validation at the histological level was provided by ex vivo serial two-photon tomography (STP). Results: Qualitatively, an increase in vessel density was observed on CE-MRI subtraction angiographies following occlusion; however, a quantitative vessel tracing analysis was prone to errors in our model. Measurements of ss-CBV reliably identified an increase in cortical vasculature, validated by DSC-MRI and STP. Conclusion: Iron oxide nanoparticle-based ss-CBV serves as a robust, non-invasive imaging surrogate marker for neocortical vessels, with the potential to reduce and refine preclinical models targeting the development and outgrowth of cerebral collateralization.
Automatically Detected Microsleep Episodes in the Fitness-to-Drive Assessment
Skorucak, Jelena; Hertig-Godeschalk, Anneke; Achermann, Peter; et al. (2020)
Frontiers in Neuroscience
Study Objectives: Microsleep episodes (MSEs) are short fragments of sleep (1–15 s) that can cause dangerous situations with potentially fatal outcomes. In the diagnostic sleep-wake and fitness-to-drive assessment, accurate and early identification of sleepiness is essential. However, in the absence of a standardised definition and a time-efficient scoring method of MSEs, these short fragments are not assessed in clinical routine. Based on data of moderately sleepy patients, we recently developed the Bern continuous and high-resolution wake-sleep (BERN) criteria for visual scoring of MSEs and corresponding machine learning algorithms for automatic MSE detection, both mainly based on the electroencephalogram (EEG). The present study aimed to investigate the relationship between automatically detected MSEs and driving performance in a driving simulator, recorded in parallel with EEG, and to assess algorithm performance for MSE detection in severely sleepy participants. Methods: Maintenance of wakefulness test (MWT) and driving simulator recordings of 18 healthy participants, before and after a full night of sleep deprivation, were retrospectively analysed. Performance of automatic detection was compared with visual MSE scoring, following the BERN criteria, in MWT recordings of 10 participants. Driving performance was measured by the standard deviation of lateral position and the occurrence of off-road events. Results: In comparison to visual scoring, automatic detection of MSEs in participants with severe sleepiness showed good performance (Cohen’s kappa = 0.66). The MSE rate in the MWT correlated with the latency to the first MSE in the driving simulator (rs = −0.54, p < 0.05) and with the cumulative MSE duration in the driving simulator (rs = 0.62, p < 0.01). No correlations between MSE measures in the MWT and driving performance measures were found. In the driving simulator, multiple correlations between MSEs and driving performance variables were observed. Conclusion: Automatic MSE detection worked well, independent of the degree of sleepiness. The rate and the cumulative duration of MSEs could be promising sleepiness measures in both the MWT and the driving simulator. The correlations between MSEs in the driving simulator and driving performance might reflect a close and time-critical relationship between sleepiness and performance, potentially valuable for the fitness-to-drive assessment.
Mixed-Precision Deep Learning Based on Computational Memory
Nandakumar, S.R.; Le Gallo, Manuel; Piveteau, Christophe; et al. (2020)
Frontiers in Neuroscience
Deep neural networks (DNNs) have revolutionized the field of artificial intelligence and have achieved unprecedented success in cognitive tasks such as image and speech recognition. Training of large DNNs, however, is computationally intensive and this has motivated the search for novel computing architectures targeting this application. A computational memory unit with nanoscale resistive memory devices organized in crossbar arrays could store the synaptic weights in their conductance states and perform the expensive weighted summations in place in a non-von Neumann manner. However, updating the conductance states in a reliable manner during the weight update process is a fundamental challenge that limits the training accuracy of such an implementation. Here, we propose a mixed-precision architecture that combines a computational memory unit performing the weighted summations and imprecise conductance updates with a digital processing unit that accumulates the weight updates in high precision. A combined hardware/software training experiment of a multilayer perceptron based on the proposed architecture using a phase-change memory (PCM) array achieves 97.73% test accuracy on the task of classifying handwritten digits (based on the MNIST dataset), within 0.6% of the software baseline. The architecture is further evaluated using accurate behavioral models of PCM on a wide class of networks, namely convolutional neural networks, long-short-term-memory networks, and generative-adversarial networks. Accuracies comparable to those of floating-point implementations are achieved without being constrained by the non-idealities associated with the PCM devices. A system-level study demonstrates 172 × improvement in energy efficiency of the architecture when used for training a multilayer perceptron compared with a dedicated fully digital 32-bit implementation.
Frontiers in neuromorphic engineering
Indiveri, Giacomo; Horiuchi, Timothy K. (2011)
Frontiers in Neuroscience
Neurobiological processing systems are remarkable computational devices. They use slow, stochastic, and inhomogeneous computing elements and yet they outperform today’s most powerful computers at tasks such as vision, audition, and motor control, tasks that we perform nearly every moment that we are awake without much conscious thought or concern. Despite the vast amount of resources dedicated to the research and development of computing, information, and communication technologies, today’s fastest and largest computers are still not able to match biological systems at robustly accomplishing real-world tasks. While the specific algorithms and representations that biological brains use are still largely unknown, it is clear that instead of Boolean logic, precise digital representations, and synchronous operations, nervous systems use hybrid analog/digital components, distributed representations, massively parallel mechanisms, combine communications with memory and computation, and make extensive use of adaptation, self-organization, and learning. On the other hand, as with many successful man-made systems, it is clear that biological brains have been co-designed with the body to operate under a specific range of conditions and assumptions about the world. Understanding the computational principles used by the brain and how they are physically embodied is crucial for developing novel computing paradigms and guiding a new generation of technologies that can combine the strengths of industrial-scale electronics with the computational performance of brains.

Publications1 - 10 of 144

Journal: Frontiers in Neuroscience

Abbreviation

Publisher

Journal Volumes

ISSN

Description

Filters

Search Results