Search

KM3NeT/ORCA is a dense array that constitutes the low-energy branch of the KM3NeT project with the main goal of resolving the question of the neutrino mass ordering. At present, the KM3NeT/ORCA Phase 1 has already been deployed, which means that six out of the planned 115 detection lines are operational. Even with this limited configuration, neutrino oscillations can already be measured and studied. In this contribution, the sensitivity to the neutrino Non-Standard Interactions (NSI) parameter $\epsilon_{\mu \tau}$ using the current stage of the KM3NeT/ORCA detector together with the projections for the final configuration are presented.

KM3NeT is a multi-purpose cubic-kilometer neutrino observatory under construction in the Mediterranean Sea. It consists of ORCA and ARCA (for Oscillation and Astroparticle Research with Cosmics in the Abyss, respectively), currently both have a few detection lines in operation. Although having different primary goals, both detectors can be used for neutrino astronomy over a wide energy range, from a few tens of MeVs to a few tens of PeV. In view of the growing field of time-domain astronomy, it is crucial to be able to identify neutrino candidates in real-time. This online neutrino sample will allow triggered neutrino alerts that will be sent to the astronomy community and to look for time/space coincidences around external electromagnetic and multi-messenger triggers. These real-time searches can significantly increase the discovery potential of transient cosmic accelerators and refine the pointing directions in the case of poorly localized triggers, such as gravitational waves. In the field of core-collapse supernovae (CCSN), the detection of the MeV-scale CCSN neutrinos is crucial as an early warning of the electromagnetic follow-up. KM3NeT's digital optical modules act as good detectors for these supernovae neutrinos. This proceeding presents the status of KM3NeT's real-time multi-messenger activities, including online event reconstruction, event classification and selection, alert distribution, and supernova monitoring., 6 pages, 6 figures, VLVNT2021 conference proceedings

[EN] To adjust expectations efficiently, prediction errors need to be associated with the precise features that gave rise to the unexpected outcome, but this credit assignment may be problematic if stimuli differ on multiple dimensions and it is ambiguous which feature dimension caused the outcome. Here, we report a potential solution: neurons in four recorded areas of the anterior fronto-striatal networks encode prediction errors that are specific to feature values of different dimensions of attended multidimensional stimuli. The most ubiquitous prediction error occurred for the reward-relevant dimension. Feature-specific prediction error signals a) emerge on average shortly after non-specific prediction error signals, b) arise earliest in the anterior cingulate cortex and later in dorsolateral prefrontal cortex, caudate and ventral striatum, and c) contribute to feature-based stimulus selection after learning. Thus, a widely-distributed feature-specific eligibility trace may be used to update synaptic weights for improved feature-based attention., This work was supported by grant MOP 102482 from the Canadian Institutes of Health Research (T.W.) and the Natural Sciences and Engineering Research Council of Canada (T.W.), as well as by the Brain in Action CREATE-IRTG program (M.O. and T.W.), and by grant LPDS 2012-08 from the Deutsche Akademie der Naturforscher Leopoldina (S.W.). Imaging data provided by the Duke Center for In Vivo Microscopy, an NIH Biomedical Technology Resource (NIHP41EB015897, 1S10OD010683-01). The funders had no role in study design, data collection and analysis, the decision to publish, or the preparation of this manuscript. The authors would like to thank Hongying Wang for technical support

The KM3NeT research infrastructure is under construction in the Mediterranean Sea. It consists of two water Cherenkov neutrino detectors, ARCA and ORCA, aimed at neutrino astrophysics and oscillation research, respectively. Instrumenting a large volume of sea water with $$\sim {6200}$$ ∼ 6200 optical modules comprising a total of $$\sim {200{,}000}$$ ∼ 200 , 000 photomultiplier tubes, KM3NeT will achieve sensitivity to $$\sim {10} \ \mathrm{MeV}$$ ∼ 10 MeV neutrinos from Galactic and near-Galactic core-collapse supernovae through the observation of coincident hits in photomultipliers above the background. In this paper, the sensitivity of KM3NeT to a supernova explosion is estimated from detailed analyses of background data from the first KM3NeT detection units and simulations of the neutrino signal. The KM3NeT observational horizon (for a $$5\,\sigma $$ 5 σ discovery) covers essentially the Milky-Way and for the most optimistic model, extends to the Small Magellanic Cloud ($$\sim {60} \ \mathrm{kpc}$$ ∼ 60 kpc ). Detailed studies of the time profile of the neutrino signal allow assessment of the KM3NeT capability to determine the arrival time of the neutrino burst with a few milliseconds precision for sources up to 5–8 kpc away, and detecting the peculiar signature of the standing accretion shock instability if the core-collapse supernova explosion happens closer than 3–5 kpc, depending on the progenitor mass. KM3NeT’s capability to measure the neutrino flux spectral parameters is also presented.

[EN] The study of oscillations, waves, and sound is included in most first-year courses on Physics, however, analyzing audio recordings to understand and test physics experiments in laboratory practices is not a common practice, compared for example with the use of visual techniques. In this paper, we fill in this gap showing the usefulness of the application of sound recording and its analysis in Physics Laboratory practices of first-year Engineering University studies. Sound recording is very simple and implemented in commonly available technology tools, such as smartphones. The analysis can be done with ease in free open-source applications, such as Audacity. This means that this experimental procedure can be easily implemented and extensively used, even in distance learning, which is particularly convenient in a pandemic context. In fact, we illustrate in this work how this approach let us to successfully transform two in-person lab practices into sessions that can be run remotely: the study of free fall and measurement of the coefficient of restitution of a ball bouncing when released from a certain height, and the measurement of the speed of vehicles by analyzing the Doppler effect of the sound that the motor vehicles produce. With this, we conclude that this is a powerful technique that should be considered, alone or in combination with other techniques, for instance video analysis, when planning the lab practices of Physics courses., S.A. was supported by the CIDEGENT Program from the Generalitat Valenciana CIDEGENT/2019/043.

[EN] Cognition involves using attended information, maintained in working memory (WM), to guide action. During a cognitive task, a correct response requires flexible, selective gating so that only the appropriate information flows from WM to downstream effectors that carry out the response. In this work, we used biophysically-detailed modeling to explore the hypothesis that network oscillations in prefrontal cortex (PFC), leveraging local inhibition, can independently gate responses to items in WM. The key role of local inhibition was to control the period between spike bursts in the outputs, and to produce an oscillatory response no matter whether the WM item was maintained in an asynchronous or oscillatory state. We found that the WM item that induced an oscillatory population response in the PFC output layer with the shortest period between spike bursts was most reliably propagated. The network resonant frequency (i.e., the input frequency that produces the largest response) of the output layer can be flexibly tuned by varying the excitability of deep layer principal cells. Our model suggests that experimentally-observed modulation of PFC beta-frequency (15-30 Hz) and gamma -frequency (30-80 Hz) oscillations could leverage network resonance and local inhibition to govern the flexible routing of signals in service to cognitive processes like gating outputs from working memory and the selection of rule-based actions. Importantly, we show for the first time that nonspecific changes in deep layer excitability can tune the output gate's resonant frequency, enabling the specific selection of signals encoded by populations in asynchronous or fast oscillatory states. More generally, this represents a dynamic mechanism by which adjusting network excitability can govern the propagation of asynchronous and oscillatory signals throughout neocortex., This work was supported by the U.S. Army Research Office under award number ARO W911NF-12-R-0012-02 to N. K., the U.S. Office of Naval Research under award number ONR MURI N00014-16-1-2832 to M. H. and E. M., the National Institute of Mental Health under award number NIMH R37MH087027 to E. M., and The MIT Picower Institute Faculty Innovation Fund to E. M. We would like to acknowledge Joachim Hass and Michelle McCarthy for early discussions of our modeling results, as well as Andre Bastos and Mikael Lundqvist for discussions relating our modeling work to their experiments.

Persistent activity, the maintenance of neural activation over short periods of time in cortical networks, is widely thought to underlie the cognitive function of working memory. A large body of modeling studies has reproduced this kind of activity using cell assemblies with strengthened synaptic connections. However, almost all of these studies have considered persistent activity within networks with homogeneous neurons and synapses, making it difficult to judge the validity of such model results for cortical dynamics, which is based on highly heterogeneous neurons. Here, we consider persistent activity in a detailed, strongly data-driven network model of the prefrontal cortex with heterogeneous neuron and synapse parameters. Surprisingly, persistent activity could not be reproduced in this model without incorporating further constraints. We identified three factors that prevent successful persistent activity: heterogeneity in the cell parameters of interneurons, heterogeneity in the parameters of short-term synaptic plasticity and heterogeneity in the synaptic weights. Our model predicts that persistent activity is recovered if the heterogeneity in the activity of individual interneurons is diminished, which could be achieved by a homeostatic plasticity mechanism. Such a plasticity scheme could also compensate the heterogeneities in the synaptic weights and short-term plasticity when applied to the inhibitory synapses. Cell assemblies shaped in this way may be potentially targeted by distinct inputs or become more responsive to specific tuning or spectral properties. Furthermore, the model predicts that a network that exhibits persistent activity is not able to dynamically produce intrinsic in vivo-like irregular activity at the same time, because heterogeneous synaptic connections are required for these dynamics. Thus, the background noise in such a network must either be produced by external input or constitutes an entirely different state of the network, which is brought about, e.g., by neuromodulation.Author summaryTo operate effectively in a constantly changing world, it is crucial to keep relevant information in mind for short periods of time. This ability, called working memory, is commonly assumed to rest on reverberating activity among members of cell assemblies. While effective in reproducing key results of working memory, most cell assembly models rest on major simplifications such as using the same parameters for all neurons and synapses, i.e., assuming homogeneity in these parameters. Here, we show that this homogeneity assumption is necessary for persistent activity to arise, specifically for inhibitory interneurons and synapses. Using a strongly data-driven network model of the prefrontal cortex, we show that the heterogeneities in the above parameters that are implied by in vitro studies prevent persistent activity. When homogeneity is imposed on inhibitory neurons and synapses, persistent activity is recovered. We propose that the homogeneity constraints can be implemented in the brain by means of homeostatic plasticity, a form of learning that keeps the activity of a network in a constant, homeostatic state. The model makes a number of predictions for biological networks, including a structural separation of networks responsible for generating persistent activity and spontaneous, noise-like activity.

[EN] Classical accounts of biased competition require an input bias to resolve the competition between neuronal ensembles driving downstream processing. However, flexible and reliable selection of behaviorally relevant ensembles can occur with unbiased stimulation: striatal D1 and D2 spiny projection neurons (SPNs) receive balanced cortical input, yet their activity determines the choice between GO and NO-GO pathways in the basal ganglia. We here present a corticostriatal model identifying three mechanisms that rely on physiological asymmetries to effect rate- and time-coded biased competition in the presence of balanced inputs. First, tonic input strength determines which one of the two SPN phenotypes exhibits a higher mean firing rate. Second, low-strength oscillatory inputs induce higher firing rate in D2 SPNs but higher coherence between D1 SPNs. Third, high-strength inputs oscillating at distinct frequencies can preferentially activate D1 or D2 SPN populations. Of these mechanisms, only the latter accommodates observed rhythmic activity supporting rule-based decision making in prefrontal cortex, We thank T. Womelsdorf for helpful suggestions on an earlier version of the manuscript. We also thank the two reviewers for the constructive comments that enhanced the quality of the manuscript. In particular, their question regarding the resonant properties of SPNs under distinct mean input helped us to uncover how the resonance of D2 SPNs shifts in frequency space (Fig. 3E). Our research was supported by the Army Research Office (ARO) Grant W911NF-12-R-0012-02 (to N.K.). Additionally, S.A. and N.K. were supported by NSF Grant DMS-1042134, and M.M.M. was supported by the Collaborative Research in Computational Neuroscience (CRCNS) NIH Grant CRCNS 1R01N5081716

[EN] Oscillations are ubiquitous features of brain dynamics that undergo task-related changes in synchrony, power, and frequency. The impact of those changes on target networks is poorly understood. In this work, we used a biophysically detailed model of prefrontal cortex (PFC) to explore the effects of varying the spike rate, synchrony, and waveform of strong oscillatory inputs on the behavior of cortical networks driven by them. Interacting populations of excitatory and inhibitory neurons with strong feedback inhibition are inhibition-based network oscillators that exhibit resonance (i.e., larger responses to preferred input frequencies). We quantified network responses in terms of mean firing rates and the population frequency of network oscillation; and characterized their behavior in terms of the natural response to asynchronous input and the resonant response to oscillatory inputs. We show that strong feedback inhibition causes the PFC to generate internal (natural) oscillations in the beta/gamma frequency range (>15 Hz) and to maximize principal cell spiking in response to external oscillations at slightly higher frequencies. Importantly, we found that the fastest oscillation frequency that can be relayed by the network maximizes local inhibition and is equal to a frequency even higher than that which maximizes the firing rate of excitatory cells; we call this phenomenon population frequency resonance. This form of resonance is shown to determine the optimal driving frequency for suppressing responses to asynchronous activity. Lastly, we demonstrate that the natural and resonant frequencies can be tuned by changes in neuronal excitability, the duration of feedback inhibition, and dynamic properties of the input. Our results predict that PFC networks are tuned for generating and selectively responding to beta- and gamma-rhythmic signals due to the natural and resonant properties of inhibition-based oscillators. They also suggest strategies for optimizing transcranial stimulation and using oscillatory networks in neuromorphic engineering., This material is based upon research supported by the U. S. Army Research Office under award number ARO W911NF-12-R-0012-02 to N. K., the U. S. Office of Naval Research under award number ONR MURI N00014-16-1-2832 to M. H., and the National Science Foundation under award number NSF DMS-1042134 (Cognitive Rhythms Collaborative: A Discovery Network) to N. K. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

SummaryPrediction errors signal unexpected outcomes indicating that expectations need to be adjusted. For adjusting expectations efficiently prediction errors need to be associated with the precise features that gave rise to the unexpected outcome. For many visual tasks this credit assignment proceeds in a multidimensional feature space that makes it ambiguous which object defining features are relevant. Here, we report of a potential solution by showing that neurons in all areas of the medial and lateral fronto-striatal networks encode prediction errors that are specific to separate features of attended multidimensional stimuli, with the most ubiquitous prediction error occurring for the reward relevant features. These feature specific prediction error signals (1) are different from a non-specific prediction error signal, (2) arise earliest in the anterior cingulate cortex and later in lateral prefrontal cortex, caudate and ventral striatum, and (3) contribute to feature-based stimulus selection after learning. These findings provide strong evidence for a widely-distributed feature-based eligibility trace that can be used to update synaptic weights for improved feature-based attention.HighlightsNeural reward prediction errors carry information for updating feature-based attention in all areas of the fronto-striatal network.Feature specific neural prediction errors emerge earliest in anterior cingulate cortex and later in lateral prefrontal cortex.Ventral striatum neurons encode feature specific surprise strongest for the goal-relevant feature.Neurons encoding feature-specific prediction errors contribute to attentional selection after learning.

Classical accounts of biased competition (BC) require an input bias to resolve the competition between neuronal ensembles driving downstream processing. However, flexible and reliable selection of behaviorally-relevant ensembles can occur with unbiased stimulation: striatal D1 and D2 medium spiny neurons (MSNs) receive balanced cortical input, yet their activity determines the choice between GO and NO-GO pathways in the basal ganglia. We present a corticostriatal model identifying three candidate mechanisms that rely on physiological asymmetries to effect rate- and time-coded BC in the presence of balanced inputs. First, tonic input strength determines which MSN phenotype exhibit higher mean firing rate (FR). Second, low strength oscillatory inputs induce higher FR in D2 MSNs but higher coherence between D1 MSNs. Third, high strength inputs oscillating at distinct frequencies preferentially activate D1 or D2 MSN populations. Of these candidate mechanisms, only the latter accommodates observed rhythmic activity supporting rule-based decision making in prefrontal cortex.

[EN] DynaSim is an open-source MATLAB/GNU Octave toolbox for rapid prototyping of neural models and batch simulation management. It is designed to speed up and simplify the process of generating, sharing, and exploring network models of neurons with one or more compartments. Models can be specified by equations directly (similar to XPP or the Brian simulator) or by lists of predefined or custom model components. The higher-level specification supports arbitrarily complex population models and networks of interconnected populations. DynaSim also includes a large set of features that simplify exploring model dynamics over parameter spaces, running simulations in parallel using both multicore processors and high-performance computer clusters, and analyzing and plotting large numbers of simulated data sets in parallel. It also includes a graphical user interface (DynaSim GUI) that supports full functionality without requiring user programming. The software has been implemented in MATLAB to enable advanced neural modeling using MATLAB, given its popularity and a growing interest in modeling neural systems. The design of DynaSim incorporates a novel schema for model specification to facilitate future interoperability with other specifications (e.g., NeuroML, SBML), simulators (e.g., NEURON, Brian, NEST), and web-based applications (e.g., Geppetto) outside MATLAB. DynaSim is freely available at http://dynasimtoolbox.org. This tool promises to reduce barriers for investigating dynamics in large neural models, facilitate collaborative modeling, and complement other tools being developed in the neuroinformatics community., This material is based upon research supported by the U.S. Army Research Office under award number ARO W911NF-12-R-0012-02, the U.S. Office of Naval Research under award number ONR MURI N00014-16-1-2832, and the National Science Foundation under award number NSF DMS-1042134 (Cognitive Rhythms Collaborative: A Discovery Network)

[EN] Noradrenaline is believed to support cognitive flexibility through the alpha 2A noradrenergic receptor (a2A-NAR) acting in prefrontal cortex. Enhanced flexibility has been inferred from improved working memory with the a2A-NA agonist Guanfacine. But it has been unclear whether Guanfacine improves specific attention and learning mechanisms beyond working memory, and whether the drug effects can be formalized computationally to allow single subject predictions. We tested and confirmed these suggestions in a case study with a healthy nonhuman primate performing a feature-based reversal learning task evaluating performance using Bayesian and Reinforcement learning models. In an initial dose-testing phase we found a Guanfacine dose that increased performance accuracy, decreased distractibility and improved learning. In a second experimental phase using only that dose we examined the faster feature-based reversal learning with Guanfacine with single-subject computational modeling. Parameter estimation suggested that improved learning is not accounted for by varying a single reinforcement learning mechanism, but by changing the set of parameter values to higher learning rates and stronger suppression of non-chosen over chosen feature information. These findings provide an important starting point for developing nonhuman primate models to discern the synaptic mechanisms of attention and learning functions within the context of a computational neuropsychiatry framework., This research was supported by grants from the Canadian Institutes of Health Research (CIHR), the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Ontario Ministry of Economic Development and Innovation (MEDI). We thank Dr. Hongying Wang for invaluable help with drug administration and animal care

[EN] Background: It is widely held that single cells in anterior cingulate and lateral prefrontal cortex (ACC/PFC) coordinate their activity during attentional processes, although cellular activity that may underlie such coordination across ACC/PFC has not been identified. We thus recorded cells in five ACC/PFC subfields of macaques engaged in a selective attention task, characterized those spiking events that indexed attention, and identified how spiking of distinct cell populations synchronized between brain areas. Results: We found that cells in ACC/PFC increased the firing of brief 200 Hz spike bursts when subjects shifted attention and engaged in selective visual processing. In contrast to non-burst spikes, burst spikes synchronized over large distances to local field potentials at narrow beta (12-20 Hz) and at gamma (50-75 Hz) frequencies. Long-range burst synchronization was anatomically specific, functionally connecting those subfields in area 24 (ACC) and area 46 (PFC) that are key players of attentional control. By splitting cells into putative excitatory (pE) and inhibitory (pI) cells by their broad and narrow spikes, we identified that bursts of pI cells preceded and that bursts of pE cells followed in time periods of maximal beta coherent network activity. In contrast, gamma bursts were transient impulses with equal timing across cell classes. Conclusions: These findings suggest that processes underlying burst firing and burst synchronization are candidate mechanisms to coordinate attention information across brain areas. We speculate that distinct burst-firing motifs realize beta and gamma synchrony to trigger versus maintain functional network states during goal-directed behavior., We thank Daniel Kaping, Johanna Stucke, Iman Janemi, and Michelle Bale for help with the electrophysiological recordings and reconstruction of recording sites. This research was supported by grants from the Canadian Institutes of Health Research (CIHR), the Natural Sciences and Engineering Research Council of Canada (NSERC), and the Ontario Ministry of Economic Development and Innovation (MEDI) (T.W.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

[EN] Errors indicate the need to adjust attention for improved future performance. Detecting errors is thus a fundamental step to adjust and control attention. These functions have been associated with the dorsal anterior cingulate cortex (dACC), predicting that dACC cells should track the specific processing states giving rise to errors in order to identify which processing aspects need readjustment. Here, we tested this prediction by recording cells in the dACC and lateral prefrontal cortex (latPFC) of macaques performing an attention task that dissociated 3 processing stages. We found that, across prefrontal subareas, the dACC contained the largest cell populations encoding errors indicating (1) failures of inhibitory control of the attentional focus, (2) failures to prevent bottom-up distraction, and (3) lapses when implementing a choice. Error-locked firing in the dACC showed the earliest latencies across the PFC, emerged earlier than reward omission signals, and involved a significant proportion of putative inhibitory interneurons. Moreover, early onset error-locked response enhancement in the dACC was followed by transient prefrontal-cingulate inhibition, possibly reflecting active disengagement from task processing. These results suggest a functional specialization of the dACC to track and identify the actual processes that give rise to erroneous task outcomes, emphasizing its role to control attentional performance., This research was supported by grants from the Canadian Institutes of Health Research (CIHR), the Natural Sciences and Engineering Research Council of Canada (NSERC), and the Ontario Ministry of Economic Development and Innovation (MEDI) (T.W.). S.W. was funded by the "Deutsche Akademie der Naturforscher Leopoldina" (LPDS 2012-08). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Candidate mechanisms of brain function can potentially be identified using biologically detailed computational models. A critical question that arises from the construction and analysis of such models is whether a particular set of parameters is unique or whether multiple different solutions exist, each capable of reproducing some relevant phenomenology. Addressing this issue is difficult, and systematic procedures have been proposed only recently, targeting small systems such as single neurons or small neural circuits [16] (Marder and Taylor, Nat Neurosci 14:133–138, 2011), [1] (Achard and De Schutter, PLoS Comput Biol 2:e94, 2006). However, how to develop a methodology to address the problem of non-uniqueness of parameters in large-scale biological networks is yet to be developed. Here, we describe a computational strategy to explicitly approach this issue on large-scale neural network models, which has been successfully applied to computational models of working memory (WM) and selective attention [2] (Ardid, J Neurosci Off J Soc Neurosci 30:2856–2870, 2010), [3] (Cano-Colino et al., Cereb Cortex 24:2449–2463, 2014). To illustrate the approach, we show in this chapter how our strategy applies to the problem of identifying different mechanisms underlying visuospatial WM. We use a well-established biological neural circuit model in the literature [6] (Compte et al., Cereb. Cortex 10:910–923, 2000) as a reference point, which we then perturb by using the Swarm Optimization Algorithm. This algorithm explores the space of biologically unconstrained parameters in the model under the constraint of preserving a solution defined here as a network in which the activity of model neurons mimics the properties of neurons in the dorsolateral prefrontal cortex (dlPFC) of monkeys performing a visuospatial WM task [7] (Funahashi et al., J Neurophysiol 61:331–349, 1989). The results are: (1) identification of a set of model solutions, composed of alternative and, in principle, feasible and sufficient mechanisms generating WM function in a cortical network. In particular, we found that the dynamics of interneurons play a main role in distinguishing among potential circuit candidates. Secondly we uncovered compensatory mechanisms in a subset of the parameters in the model. In essence, the compensatory mechanisms we observe in the different solutions are based on correlations between sets of parameters that shift the local Excitatory/Inhibitory balance in opposite directions. In summary, our approach is able to identify distinct mechanisms underlying a same function, as well as to propose a dynamic solution to the problem of fine-tuning. Our results from the proposed workflow would be strengthened by additional biological experiments aimed to refine the validity of the results.

[EN] Attention includes processes that evaluate stimuli relevance, select the most relevant stimulus against less relevant stimuli, and bias choice behavior toward the selected information. It is not clear how these processes interact. Here, we captured these processes in a reinforcement learning framework applied to a feature-based attention task that required macaques to learn and update the value of stimulus features while ignoring nonrelevant sensory features, locations, and action plans. We found that value-based reinforcement learning mechanisms could account for feature-based attentional selection and choice behavior but required a value-independent stickiness selection process to explain selection errors while at asymptotic behavior. By comparing different reinforcement learning schemes, we found that trial-by-trial selections were best predicted by a model that only represents expected values for the task-relevant feature dimension, with nonrelevant stimulus features and action plans having only a marginal influence on covert selections. These findings show that attentional control subprocesses can be described by (1) the reinforcement learning of feature values within a restricted feature space that excludes irrelevant feature dimensions, (2) a stochastic selection process on feature-specific value representations, and (3) value-independent stickiness toward previous feature selections akin to perseveration in the motor domain. We speculate that these three mechanisms are implemented by distinct but interacting brain circuits and that the proposed formal account of feature-based stimulus selection will be important to understand how attentional subprocesses are implemented in primate brain networks., This work was supported by grants from the Canadian Institutes of Health Research, the Natural Sciences and Engineering Research Council of Canada, and the Ontario Ministry of Economic Development and Innovation. We thank Johanna Stucke for her help with assisting with animal training and care. We thank anonymous reviewers for helpful comments on earlier versions of this manuscript.

[EN] Microcircuits are composed of multiple cell classes that likely serve unique circuit operations. But how cell classes map onto circuit functions is largely unknown, particularly for primate prefrontal cortex during actual goal-directed behavior. One difficulty in this quest is to reliably distinguish cell classes in extracellular recordings of action potentials. Here we surmount this issue and report that spike shape and neural firing variability provide reliable markers to segregate seven functional classes of prefrontal cells in macaques engaged in an attention task. We delineate an unbiased clustering protocol that identifies four broad spiking (BS) putative pyramidal cell classes and three narrow spiking (NS) putative inhibitory cell classes dissociated by how sparse, bursty, or regular they fire. We speculate that these functional classes map onto canonical circuit functions. First, two BS classes show sparse, bursty firing, and phase synchronize their spiking to 3-7 Hz (theta) and 12-20 Hz (beta) frequency bands of the local field potential (LFP). These properties make cells flexibly responsive to network activation at varying frequencies. Second, one NS and two BS cell classes show regular firing and higher rate with only marginal synchronization preference. These properties are akin to setting tonically the excitation and inhibition balance. Finally, two NS classes fired irregularly and synchronized to either theta or beta LFP fluctuations, tuning them potentially to frequency-specific subnetworks. These results suggest that a limited set of functional cell classes emerges in macaque prefrontal cortex (PFC) during attentional engagement to not only represent information, but to subserve basic circuit operations., This research was supported by grants from the Canadian Institutes of Health Research, the Natural Sciences and Engineering Research Council of Canada, and the Ontario Ministry of Economic Development and Innovation. We thank Iman Janemi and Michelle Bale for their help with recording and anatomical reconstruction of the neurophysiological data. We also thank the reviewers for their thoughtful comments and suggestions.

Top-down inputs from prefrontal cortex impact on sensory neurons [1,2], enhancing their selectivity to attended stimuli, while sensory processing of distractors is suppressed [1,3,4]. However, what are the neuro-computational mechanisms that identify the behaviorally relevant information that is worth to bias? Previous studies linked selective attention to learned value [5], suggesting that attentional selection relies on internal mechanisms that track the relevance of sensory information. Consistent with this view, we recently introduced a reinforcement learning (RL) approach for the deployment of selective attention [6]. We tested model-free and model-based versions of RL to identify the mechanisms that most accurately predict behavior: whereas model-based prioritizes attentional selection to task features that are systematically associated with reward, model-free considers all available features. Our results proved that the optimal task-set representation significantly improved predictive power, suggesting that monkeys benefited from model-based mechanisms [6]. Yet, model-based presents two important limitations: i) it is unable to adapt to changes in the association of reward with sensory features that are not included in the learned model, and ii) it excludes the mechanism of learning by which subjects derive the proper task-set representation. The question remains then of how a prioritized task set can be learned to exploit the benefits of employing a model. With the aim to workaround model-based limitations and shed light on the underlying mechanisms that make model-based benefits possible, we propose here the “adaptive learning” mechanism for flexible task-set representation. The mechanism is dynamically tuned according to the statistics of association between sensory features and reward outcome, flexibly adjusting its regime of operation between model-free and modelbased systems. To test the adaptive learning mechanism, we developed it in a RL model and extended our previous analysis of monkey behavior to both, cued [5] and uncued [6], versions of the same attention task. This RL model was able to transition from a naive starting point to an optimal task-set representation, and to flexibly adapt among optimal task-set representations upon changes in reward contingencies. The model achieved so by tracking a separate learning rate a for each stimulus feature in the environment. If the selection of a stimulus feature was systematically correlated with a particular outcome (reward or no-reward), a increased. In contrast, when a feature was unable to consistently predict reward, a decayed. Thus, a changed dynamically, and over a large number of trials any a associated with a non-predictive feature went to zero, effectively eliminating it from the task set, which tuned performance towards that of the model-based system. The adaptive learning mechanism introduced here represents a step further in our understanding of the origins of selective attention. Notably, our results prove that the adaptive learning is an optimal mechanistic candidate to support arbitrary prioritized model-based formation under generic conditions of covert attentional selection, and regardless of whether attentional selection operates on cued or uncued tasks.

Even the simplest tasks in our everyday life depend on the activity in multiple brain areas that are coordinated in large-scale brain networks (Sporns, 2011). These networks restructure the information flow in the brain on fast time scales whenever we re-focus our attention on novel tasks or initiate novel movements to interact with our environment. This restructuring of information flow is implemented in cortical circuits by functionally changing the identity and composition of cells in the output layers that connect to other long-distant network nodes. Recent anatomical evidence has begun to show that these output cells form highly specific, segregated subnetworks (Krook-Magnuson et al., 2012). Cells within a subnetwork more likely interconnect with each other and share distant projection targets, avoiding interactions with other cells that project elsewhere. It remains unknown, however, how cells of the same deep layer subnetwork are selected when task demands change (Douglas and Martin, 2004). Results from a recent study by Canolty et al. (2012) suggest an interesting possibility to resolve this context-dependent output selection by showing that the composition of cells that fire together and in phase in the beta cycle can be inferred from the strength of local beta rhythmic modulation. Canolty and colleagues recorded the firing of cells in deep layer motor cortex of primates engaged in either of two tasks that required moving a cursor between visual stimuli manually (with their hands), or through brain activity (brain control). During these tasks, the majority of cells in motor cortex fire systematically at particular phases of a beta oscillation that reduces its amplitudes when actual or brain controlled movements are planned and executed. Canolty and colleagues show that beta synchronous firing rates of individual cells increase or decrease in close correspondence to increases or decreases in the amplitude of beta oscillations. This cell-specific mapping between firing rates and beta amplitudes was highly stable for single cells across multiple recording sessions, but it varied for a large subset of cells between different tasks in a reversible and reliable manner. These findings have potentially wide implications for our understanding of the mechanisms of rapid sub-network selection. Figure ​Figure1A1A illustrates the cell-specific mapping between firing rates and beta amplitudes and how a particular beta rhythmic state in a local circuit could signify which cells fire and are therefore selected into the currently active subnetwork. When the beta amplitude of the local field potentials surrounding the cell changes over time, for example when it is reduced during movement initiation, the subnetwork of cells that fired during high beta states dissipates and the circuit switches on cells that prefer firing at low beta amplitudes (Figure ​(Figure1A).1A). Furthermore, Canolty et al. showed that for one third of beta-modulated cells, the rank ordering of firing rates varied systematically during different tasks. As shown in Figure ​Figure1B,1B, this task specificity of reliable beta-to-rate mapping could underlie the flexible and fast selection of task-specific subnetworks. Figure 1 Proposed scheme of subnetwork participation of deep layer cells by amplitude variations of beta oscillations. (A) Canolty et al. (2012) have shown that in deep cortical layers the firing rate of single cells show a highly robust sigmoidal relation to ... The task- and cell- specific mapping between beta amplitude and firing rate reveals a statistically robust relationship of two, in principle, independent signals. As suggested by Canolty and colleagues, this relationship would allow activating brain circuits by changing the beta rhythmic temporal structure—independently from targeting the firing rate of neurons explicitly. Such a mechanism assumes that the beta-rhythm is causal of the firing rate changes. Specifically, as indicated in Figure ​Figure1C1C (left panel), external beta rhythmic input may act as the trigger to entrain synaptic activity of cells in target areas that are resonant to beta rhythmic fluctuations of input. Stronger rhythmicity of beta rhythmic input will thereby select cells according to their beta-specific sensitivity and possibly independent of their overall level of excitation (Akam and Kullmann, 2010). According to this scenario, a change of beta rhythmic activity would serve as a true switch of a local network by causally modulating firing rates. Future studies, using stimulation techniques, will be necessary to test the prediction of beta rhythmicity being causally involved in task-specific subnetwork selection. An alternative possibility is that varying beta amplitudes within a local circuit may not cause the switch of subnetworks, but may rather reflect the consequence of the switch itself (Figure ​(Figure1C,1C, right panel). According to this assumption, beta amplitudes may derive from the intrinsic properties of the cells that are selected, including e.g., beta rhythmic, intrinsic burst firing. Thus, the actual switch of active cells would follow from mechanisms that only indirectly relate to the rhythm generating mechanisms. For example, in computational firing rate models of randomly connected networks, such switches of subnetworks can be achieved by locally biasing the balance of excitation and inhibition, such that selected cells will be released from inhibition and will maintain their selectively activated state via the di-synaptic inhibition of non-selected cells (Vogels and Abbott, 2009). Whatever the actual mechanism that causes a local circuit to switch the subnetworks of cells in deep cortical layers, the finding of a reliable mapping between beta amplitude and firing rate in the majority of deep layer cells in motor cortex critically extends our perspective of the working principles of brain activity. The results by Canolty et al. show that brain signals (1) at the local scale of cell (firing), (2) at the meso-scale of circuits (beta entrainment), and (3) at macro-scales comprising long-distant networks (inter-areal beta-coherence, not discussed here), combine together in systematic ways to subserve the larger goal to establish functional networks that can flexibly switch according to rapidly varying task demands. This “cross-level” relation of activity is only recently moving into the focus of scientific scrutiny. The discussed study and its broad analysis of available neuronal signals from all three levels of neuronal dynamics points the right way into this direction, and promises to critically advance our understanding of how external factors like specific reach movements, or the shifting of attention are implemented by the dynamics of local circuits and their dynamic interplay with larger functional brain networks.

The brain possesses limited resources and utilizes selective attention as the mechanism to manage the massive influx of sensory information into the cortex. Selective attention strengthens the impact of behaviorally relevant information and diminishes distractions from irrelevant inputs. For instance, in visual discrimination or detection tasks, proper allocation of attention improves performance and shortens response times. At the neural level, there are many different effects of attention on the response of sensory neurons: receptive field shrinkages, modulation of neural synchronization and mean activity, variability reduction, interneuronal decorrelations, and more (Reynolds and Chelazzi, 2004). Yet few experiments have attempted to determine how attentional correlates subserve behavioral benefits (Womelsdorf et al., 2006).

[EN] In this computational work, we investigated gamma-band synchronization across cortical circuits associated with selective attention. The model explicitly instantiates a reciprocally connected loop of spiking neurons between a sensory-type (area MT) and an executive-type (prefrontal/parietal) cortical circuit (the source area for top-down attentional signaling). Moreover, unlike models in which neurons behave as clock-like oscillators, in our model single-cell firing is highly irregular (close to Poisson), while local field potential exhibits a population rhythm. In this "sparsely synchronized oscillation" regime, the model reproduces and clarifies multiple observations from behaving animals. Top-down attentional inputs have a profound effect on network oscillatory dynamics while only modestly affecting single-neuron spiking statistics. In addition, attentional synchrony modulations are highly selective: interareal neuronal coherence occurs only when there is a close match between the preferred feature of neurons, the attended feature, and the presented stimulus, a prediction that is experimentally testable. When interareal coherence was abolished, attention-induced gain modulations of sensory neurons were slightly reduced. Therefore, our model reconciles the rate and synchronization effects, and suggests that interareal coherence contributes to large-scale neuronal computation in the brain through modest enhancement of rate modulations as well as a pronounced attention-specific enhancement of neural synchrony., This work was funded by the Volkswagen Foundation, the Spanish Ministry of Science and Innovation, and the European Regional Development Fund. A.C. is supported by the Researcher Stabilization Program of the Health Department of the Generalitat de Catalunya. X.-J.W. is supported by the National Institutes of Health Grant 2R01MH062349 and the Kavli Foundation. We are thankful to Stefan Treue for fruitful discussions and to Jorge Ejarque for technical support in efficiently implementing the search optimization procedure in a grid cluster computing system. Also, we thankfully acknowledge the computer resources and assistance from the Barcelona Supercomputing Center-Centro Nacional de Supercomputación, Spain.

Selective attention is a fundamental cognitive function that uses top-down signals to orient and prioritize information processing in the brain. Single-cell recordings from behaving monkeys have revealed a number of attention-induced effects on sensory neurons, and have given rise to contrasting viewpoints about the neural underpinning of attentive processing. Moreover, there is evidence that attentional signals originate from the prefrontoparietal working memory network, but precisely how a source area of attention interacts with a sensory system remains unclear. To address these questions, we investigated a biophysically based network model of spiking neurons composed of a reciprocally connected loop of two (sensory and working memory) networks. We found that a wide variety of physiological phenomena induced by selective attention arise naturally in such a system. In particular, our work demonstrates a neural circuit that instantiates the “feature-similarity gain modulation principle,” according to which the attentional gain effect on sensory neuronal responses is a graded function of the difference between the attended feature and the preferred feature of the neuron, independent of the stimulus. Furthermore, our model identifies key circuit mechanisms that underlie feature-similarity gain modulation, multiplicative scaling of tuning curve, and biased competition, and provide specific testable predictions. These results offer a synthetic account of the diverse attentional effects, suggesting a canonical neural circuit for feature-based attentional processing in the cortex., This work was supported by the Volkswagen Foundation, the Spanish Ministry of Education and Science (S.A., A.C.), the European Regional Development Fund (A.C.), and the Swartz Foundation (X.-J.W.). A.C. was supported by a Ramón y Cajal Research Fellowship of the Spanish Ministry of Education and Science and by the Researcher Stabilization Program of the Health Department of the Generalitat de Catalunya.

A hallmark of executive control is the brain’s agility to shift between different tasks depending on the behavioral rule currently in play [1]. Humans and other animals exhibit a remarkable ability to flexibly select an appropriate response to a sensory input, and rapidly switch to another sensory-response mapping when task rule or goal changes. An increasing number of monkey experiments have been performed using task-switching paradigms, combined with single-neuron recording from sensory, parietal, and prefrontal cortical areas. Physiological evidence from these studies suggests that modulation of neural activity by task rule is typically weak [2]. By contrast, most previous models commonly assume that a rule signal is similarly strong as sensory stimulation in affecting activity of cortical neurons [3]. How can small rule modulation explain large (binary) behavioral changes in task switching? In this work, we propose a solution to this puzzle, which we refer to as “the tweaking hypothesis” [4]. The core idea is that network reconfiguration underlying task switching can be realized by very weak top-down signals from rule neurons in prefrontal cortex. This is because a weak input can be greatly amplified through reverberating “attractor” dynamics in categorization and decision circuits, ultimately leading to circuit selection in favor of one sensory-motor mapping over another. We tested the tweaking hypothesis by developing a neural circuit model for task switching that consists of several basic and interacting circuit modules for sensory coding, rule representation, categorization of stimulus features, and action selection, respectively [4]. The model was validated by reproducing salient single-neuron physiological observations [2] and behavioral effects associated with task switching [1,5,6]. Notably, the model identifies specific circuit mechanisms, in terms of neural dynamics and reward-dependent synaptic plasticity, that explain salient and widely observed behavioral effects associated with task switching [4]: (i) Switch cost: response time and error rate increase in trials following a task switch. Switch cost splits into a component that decreases with a longer time for preparation and a residual component that remains [5]. (ii) Task-response interaction: on task repeat trials, the response time is shorter if the same motor response is repeated; by contrast on switch trials, response time is shorter if an alternative motor response is selected [5,6]. (iii) Congruency effect: response times and the error rate are larger when the stimulus is incongruent compared to when it is congruent, which depends on whether the mapped behavioral response is different or the same, according to alternative rules [5,6]. This work represents a neural circuit model for task switching and sheds insights in the brain mechanism of a fundamental cognitive capability; in particular, that category-selective neurons play an essential role in resolving the sensory-motor conflicts that typically appear in task-switching paradigms [4].