In an ever-changing visual world, animals' survival depends on their ability to perceive and respond to rapidly changing motion cues. The primary visual cortex (V1) is at the forefront of this sensory processing, orchestrating neural responses to perturbations in visual flow. However, the underlying neural mechanisms that lead to distinct cortical responses to such perturbations remain enigmatic. In this study, our objective was to uncover the neural dynamics that govern V1 neurons' responses to visual flow perturbations using a biologically realistic computational model. By subjecting the model to sudden changes in visual input, we observed opposing cortical responses in excitatory layer 2/3 (L2/3) neurons, namely, depolarizing and hyperpolarizing responses. We found that this segregation was primarily driven by the competition between external visual input and recurrent inhibition, particularly within L2/3 and L4. This division was not observed in excitatory L5/6 neurons, suggesting a more prominent role for inhibitory mechanisms in the visual processing of the upper cortical layers. Our findings share similarities with recent experimental studies focusing on the opposing influence of top-down and bottom-up inputs in the mouse primary visual cortex during visual flow perturbations.
Visual Cortex , Mice , Animals , Visual Cortex/physiology , Photic Stimulation , Neurons/physiology , Sensation , Visual Perception/physiology
[This corrects the article DOI: 10.3389/fnins.2023.1223950.].
The alpha rhythm is often associated with relaxed wakefulness or idling and is altered by various factors. Abnormalities in the alpha rhythm have been linked to several neurological and psychiatric disorders, including Alzheimer's disease. Transcranial alternating current stimulation (tACS) has been proposed as a potential tool to restore a disrupted alpha rhythm in the brain by stimulating at the individual alpha frequency (IAF), although some research has produced contradictory results. In this study, we applied an IAF-tACS protocol over parieto-occipital areas to a sample of healthy subjects and measured its effects over the power spectra. Additionally, we used computational models to get a deeper understanding of the results observed in the experiment. Both experimental and numerical results showed an increase in alpha power of 8.02% with respect to the sham condition in a widespread set of regions in the cortex, excluding some expected parietal regions. This result could be partially explained by taking into account the orientation of the electric field with respect to the columnar structures of the cortex, showing that the gyrification in parietal regions could generate effects in opposite directions (hyper-/depolarization) at the same time in specific brain regions. Additionally, we used a network model of spiking neuronal populations to explore the effects that these opposite polarities could have on neural activity, and we found that the best predictor of alpha power was the average of the normal components of the electric field. To sum up, our study sheds light on the mechanisms underlying tACS brain activity modulation, using both empirical and computational approaches. Non-invasive brain stimulation techniques hold promise for treating brain disorders, but further research is needed to fully understand and control their effects on brain dynamics and cognition. Our findings contribute to this growing body of research and provide a foundation for future studies aimed at optimizing the use of non-invasive brain stimulation in clinical settings.
Nonlinear dynamical systems exhibiting inherent memory can process temporal information by exploiting their responses to input drives. Reservoir computing is a prominent approach to leverage this ability for time-series forecasting. The computational capabilities of analog computing systems often depend on both the dynamical regime of the system and the input drive. Most studies have focused on systems exhibiting a stable fixed-point solution in the absence of input. Here, we go beyond that limitation, investigating the computational capabilities of a paradigmatic delay system in three different dynamical regimes. The system we chose has an Ikeda-type nonlinearity and exhibits fixed point, bistable, and limit-cycle dynamics in the absence of input. When driving the system, new input-driven dynamics emerge from the autonomous ones featuring characteristic properties. Here, we show that it is feasible to attain consistent responses across all three regimes, which is an essential prerequisite for the successful execution of the tasks. Furthermore, we demonstrate that we can exploit all three regimes in two time-series forecasting tasks, showcasing the versatility of this paradigmatic delay system in an analog computing context. In all tasks, the lowest prediction errors were obtained in the regime that exhibits limit-cycle dynamics in the undriven reservoir. To gain further insights, we analyzed the diverse time-distributed node responses generated in the three regimes of the undriven system. An increase in the effective dimensionality of the reservoir response is shown to affect the prediction error, as also fine-tuning of the distribution of nonlinear responses. Finally, we demonstrate that a trade-off between prediction accuracy and computational speed is possible in our continuous delay systems. Our results not only provide valuable insights into the computational capabilities of complex dynamical systems but also open a new perspective on enhancing the potential of analog computing systems implemented on various hardware platforms.
We design scalable neural networks adapted to translational symmetries in dynamical systems, capable of inferring untrained high-dimensional dynamics for different system sizes. We train these networks to predict the dynamics of delay-dynamical and spatiotemporal systems for a single size. Then, we drive the networks by their own predictions. We demonstrate that by scaling the size of the trained network, we can predict the complex dynamics for larger or smaller system sizes. Thus, the network learns from a single example and by exploiting symmetry properties infers entire bifurcation diagrams.
Microring resonators (MRRs) are a key photonic component in integrated devices, due to their small size, low insertion losses, and passive operation. While the MRRs have been established for optical filtering in wavelength-multiplexed systems, the nonlinear properties that they can exhibit give rise to new perspectives on their use. For instance, they have been recently considered for introducing optical nonlinearity in photonic reservoir computing systems. In this work, we present a detailed numerical investigation of a silicon MRR operation, in the presence of external optical feedback, in a time delay reservoir computing scheme. We demonstrate the versatility of this compact, passive device, by exploiting different operating regimes and solving computing tasks with diverse memory requirements. We show that when large memory is required, as it occurs in the Narma 10 task, the MRR nonlinearity does not play a significant role when the photodetection nonlinearity is involved, while the contribution of the external feedback is significant. On the contrary, for computing tasks such as the Mackey-Glass and the Santa Fe chaotic timeseries prediction, the MRR and the photodetection nonlinearities contribute both to efficient computation. The presence of optical feedback improves the prediction of the Mackey-Glass timeseries while it plays a minor role in the Santa Fe timeseries case.
Synchronization between neuronal populations is hypothesized to play a crucial role in the communication between brain networks. The binding of features, or the association of computations occurring in spatially segregated areas, is supposed to take place when a stable synchronization between cortical areas occurs. While a direct cortico-cortical connection typically fails to support this mechanism, the participation of a third area, a relay element, mediating in the communication was proposed to overcome this limitation. Among the different structures that could play the role of coordination during the binding process, the thalamus is the best placed region to carry out this task. In this paper we study how information flows in a canonical motif that mimics a cortico-thalamo-cortical circuit composed by three mutually coupled neuronal populations (also called the V-motif). Through extensive numerical simulations, we found that the amount of information transferred between the oscillating neuronal populations is determined by the delay in their connections and the mismatch in their oscillation frequencies (detuning). While the transmission from a cortical population is mostly restricted to positive detuning, transmission from the relay (thalamic) population to the cortical populations is robust for a broad range of detuning values, including negative values, while permitting feedback communication from the cortex at high frequencies, thus supporting robust bottom up and top down interaction. In this case, a strong feedback transmission between the cortex to thalamus supports the possibility of robust bottom-up and top-down interactions in this motif. Interestingly, adding a cortico-cortical bidirectional connection to the V-motif (C-motif) expands the dynamics of the system with distinct operation modes. While overall transmission efficiency is decreased, new communication channels establish cortico-thalamo-cortical association loops. Switching between operation modes depends on the synaptic strength of the cortico-cortical connections. Our results support a role of the transthalamic V-motif in the binding of spatially segregated cortical computations, and suggest an important regulatory role of the direct cortico-cortical connection.
The structure of the brain network exhibits modularity at multiple spatial scales. The effect of the modular structure on the brain dynamics has been the focus of several studies in recent years but many aspects remain to be explored. For example, it is not well-known how the delays in the transmission of signals between the neurons and the brain regions interact with the modular structure to determine the brain dynamics. In this paper, we show an important impact of the delays on the collective dynamics of brain networks with modular structure; that is, the degree of the synchrony between different brain regions depends on the oscillating frequency. In particular, we show that when increasing the frequency of the nodes the network transits from a global synchrony state to an asynchronous state, through a transition region over which the local synchrony inside the modules is stronger than the global synchrony. When the delays depend on the distance between the nodes, the modular structure of different spatial scales appears in the correlation matrix over different specific frequency bands, so that, finer spatial modular structures reveal in higher frequency bands. The results are corroborated by a simple theoretical argument and elaborated by simulations on several simplified modular networks and the connectome with different spatial resolutions.
Brain/physiology , Connectome/methods , Nerve Net/physiology , Neural Networks, Computer , Humans , Neurons/physiology , Time Factors
Hippocampal firing is organized in theta sequences controlled by internal memory processes and by external sensory cues, but how these computations are coordinated is not fully understood. Although theta activity is commonly studied as a unique coherent oscillation, it is the result of complex interactions between different rhythm generators. Here, by separating hippocampal theta activity in three different current generators, we found epochs with variable theta frequency and phase coupling, suggesting flexible interactions between theta generators. We found that epochs of highly synchronized theta rhythmicity preferentially occurred during behavioral tasks requiring coordination between internal memory representations and incoming sensory information. In addition, we found that gamma oscillations were associated with specific theta generators and the strength of theta-gamma coupling predicted the synchronization between theta generators. We propose a mechanism for segregating or integrating hippocampal computations based on the flexible coordination of different theta frameworks to accommodate the cognitive needs.
In the brain, a vast number of neurons coordinate their activity to support complex cognitive processes. One of the best places to see this in action is the hippocampus, a brain structure with a key role in memory and navigation. The hippocampus shows waves of electrical activity, which represent the synchronized firing of large numbers of neurons. The hippocampus can generate multiple rhythms at once. The two main rhythms are theta and gamma. Theta waves are slow, with a frequency of about 8 Hertz. Gamma waves are faster with a frequency of up to 120 Hertz or even more. Theta waves are always present in the brains of freely moving animals, whereas gamma waves occur in brief bursts. These bursts usually correspond to a particular point on the theta wave. One burst may occur just before each peak of the theta wave, for example, whereas another burst may occur just after the peak. This separation enables individual bursts of gamma to carry different messages without them becoming mixed up. This is similar to how radio stations broadcast their signals at different carrier frequencies to avoid interference. By recording hippocampal activity in rats exploring a maze, Lopez-Madrona et al. now show that the hippocampus has not one, but three generators of theta waves. Having three sources of theta, each of which can be synchronized with gamma, provides a more versatile system for encoding and sending information. It also means that the three theta generators can vary the degree to which they coordinate their firing. This helps the brain combine or separate streams of information as required. By working together to create a single theta rhythm, for example, the three theta generators can help animals combine information stored in memory with incoming sensory input. How the coordination of theta rhythms in the hippocampus influences the activity of other brain regions involved in learning and memory remains unclear. However, uncoupling of theta and gamma waves seems to be an early sign of Alzheimer's disease and can also be seen in the brains of people with schizophrenia and other psychiatric disorders. Understanding how this process occurs could provide clues to the origin of these disorders.
Exploratory Behavior/physiology , Hippocampus/physiology , Memory/physiology , Theta Rhythm/physiology , Animals , Male , Rats , Rats, Long-Evans
Neurons encode and transmit information in spike sequences. However, despite the effort devoted to understand the encoding and transmission of information, the mechanisms underlying the neuronal encoding are not yet fully understood. Here, we use a nonlinear method of time-series analysis (known as ordinal analysis) to compare the statistics of spike sequences generated by applying an input signal to the neuronal model of Morris-Lecar. In particular, we consider two different regimes for the neurons which lead to two classes of excitability: class I, where the frequency-current curve is continuous and class II, where the frequency-current curve is discontinuous. By applying ordinal analysis to sequences of inter-spike-intervals (ISIs) our goals are (1) to investigate if different neuron types can generate spike sequences which have similar symbolic properties; (2) to get deeper understanding on the effects that electrical (diffusive) and excitatory chemical (i.e., excitatory synapse) couplings have; and (3) to compare, when a small-amplitude periodic signal is applied to one of the neurons, how the signal features (amplitude and frequency) are encoded and transmitted in the generated ISI sequences for both class I and class II type neurons and electrical or chemical couplings. We find that depending on the frequency, specific combinations of neuron/class and coupling-type allow a more effective encoding, or a more effective transmission of the signal.
Computer Simulation , Models, Neurological , Neurons/metabolism , Synapses , Synaptic Transmission , Animals , Humans
The specific connectivity of a neuronal network is reflected in the dynamics of the signals recorded on its nodes. The analysis of how the activity in one node predicts the behaviour of another gives the directionality in their relationship. However, each node is composed of many different elements which define the properties of the links. For instance, excitatory and inhibitory neuronal subtypes determine the functionality of the connection. Classic indexes such as the Granger causality (GC) quantifies these interactions, but they do not infer into the mechanism behind them. Here, we introduce an extension of the well-known GC that analyses the correlation associated to the specific influence that a transmitter node has over the receiver. This way, the G-causal link has a positive or negative effect if the predicted activity follows directly or inversely, respectively, the dynamics of the sender. The method is validated in a neuronal population model, testing the paradigm that excitatory and inhibitory neurons have a differential effect in the connectivity. Our approach correctly infers the positive or negative coupling produced by different types of neurons. Our results suggest that the proposed approach provides additional information on the characterization of G-causal connections, which is potentially relevant when it comes to understanding interactions in the brain circuits.
Computational Biology/methods , Hippocampus/physiology , Models, Neurological , Nerve Net/physiology , Computer Simulation , Electrophysiological Phenomena/physiology , Humans , Neurons/physiology
Synchronization is one of the brain mechanisms allowing the coordination of neuronal activity required in many cognitive tasks. Anticipated Synchronization (AS) is a specific type of out-of-phase synchronization that occurs when two systems are unidirectionally coupled and, consequently, the information is transmitted from the sender to the receiver, but the receiver leads the sender in time. It has been shown that the primate cortex could operate in a regime of AS as part of normal neurocognitive function. However it is still unclear what is the mechanism that gives rise to anticipated synchronization in neuronal motifs. Here, we investigate the synchronization properties of cortical motifs on multiple scales and show that the internal dynamics of the receiver, which is related to its free running frequency in the uncoupled situation, is the main ingredient for AS to occur. For biologically plausible parameters, including excitation/inhibition balance, we found that the phase difference between the sender and the receiver decreases when the free running frequency of the receiver increases. As a consequence, the system switches from the usual delayed synchronization (DS) regime to an AS regime. We show that at three different scales, neuronal microcircuits, spiking neuronal populations and neural mass models, both the inhibitory loop and the external current acting on the receiver mediate the DS-AS transition for the sender-receiver configuration by changing the free running frequency of the receiver. Therefore, we propose that a faster internal dynamics of the receiver system is the main mechanism underlying anticipated synchronization in brain circuits.
Neurons can anticipate incoming signals by exploiting a physiological mechanism that is not well understood. This article offers a novel explanation on how a receiver neuron can predict the sender's dynamics in a unidirectionally-coupled configuration, in which both sender and receiver follow the evolution of a multi-scale excitable system. We present a novel theoretical viewpoint based on a mathematical object, called canard, to explain anticipation in excitable systems. We provide a numerical approach, which allows to determine the transient effects of canards. To demonstrate the general validity of canard-mediated anticipation in the context of excitable systems, we illustrate our framework in two examples, a multi-scale radio-wave circuit (the van der Pol model) that inspired a caricature neuronal model (the FitzHugh-Nagumo model) and a biophysical neuronal model (a 2-dimensional reduction of the Hodgkin-Huxley model), where canards act as messengers to the senders' prediction. We also propose an experimental paradigm that would enable experimental neuroscientists to validate our predictions. We conclude with an outlook to possible fascinating research avenues to further unfold the mechanisms underpinning anticipation. We envisage that our approach can be employed by a wider class of excitable systems with appropriate theoretical extensions.
Neurons/physiology , Action Potentials , Computer Simulation , Electrophysiology , Humans , Models, Neurological , Nervous System Physiological Phenomena
BACKGROUND AND OBJECTIVE: In this work, we develop a fully automatic and real-time ventricular heartbeat classifier based on a single ECG lead. Single ECG lead classifiers can be especially useful for wearable technologies that provide continuous and long-term monitoring of the electrocardiogram. These wearables usually have a few non-standard leads and the quality of the signals depends on the user physical activity. METHODS: The proposed method uses an Echo State Network (ESN) to classify ECG signals following the Association for the Advancement of Medical Instrumentation (AAMI) recommendations with an inter-patient scheme. To achieve real-time classification, the classifier itself and the feature extraction approach are fast and computationally efficient. In addition, our approach allows transferring the knowledge from one database to another without additional training. RESULTS: The classification performance of the proposed model is validated on the MIT-BIH arrhythmia and INCART databases. The sensitivity and precision of the proposed method for MIT-BIH arrhythmia database are 95.3 and 88.8 for the modified lead II and 90.9 and 89.2 for the V1 lead. The results reported are further compared to the existing methodologies in literature. Our methodology is a competitive single lead ventricular heartbeat classifier, that is comparable to state-of-the-art algorithms using multiple leads. CONCLUSIONS: The proposed fully automated, single-lead and real-time heartbeat classifier of ventricular heartbeats reports an improved classification accuracy in different leads of the evaluated databases in comparison with other single lead heartbeat classifiers. These results open the possibility of applying our methodology to wearable long-term monitoring devices with an unconventional placement of the electrodes.
Electrocardiography , Heart Rate , Signal Processing, Computer-Assisted , Wearable Electronic Devices , Heart Ventricles
Humans perform remarkably well in many cognitive tasks including pattern recognition. However, the neuronal mechanisms underlying this process are not well understood. Nevertheless, artificial neural networks, inspired in brain circuits, have been designed and used to tackle spatio-temporal pattern recognition tasks. In this paper we present a multi-neuronal spike pattern detection structure able to autonomously implement online learning and recognition of parallel spike sequences (i.e., sequences of pulses belonging to different neurons/neural ensembles). The operating principle of this structure is based on two spiking/synaptic neurocomputational characteristics: spike latency, which enables neurons to fire spikes with a certain delay and heterosynaptic plasticity, which allows the own regulation of synaptic weights. From the perspective of the information representation, the structure allows mapping a spatio-temporal stimulus into a multi-dimensional, temporal, feature space. In this space, the parameter coordinate and the time at which a neuron fires represent one specific feature. In this sense, each feature can be considered to span a single temporal axis. We applied our proposed scheme to experimental data obtained from a motor-inhibitory cognitive task. The results show that out method exhibits similar performance compared with other classification methods, indicating the effectiveness of our approach. In addition, its simplicity and low computational cost suggest a large scale implementation for real time recognition applications in several areas, such as brain computer interface, personal biometrics authentication, or early detection of diseases.
The emergence of flexible information channels in brain networks is a fundamental question in neuroscience. Understanding the mechanisms of dynamic routing of information would have far-reaching implications in a number of disciplines ranging from biology and medicine to information technologies and engineering. In this work, we show that the presence of a node firing at a higher frequency in a network with local connections, leads to reliable transmission of signals and establishes a preferential direction of information flow. Thus, by raising the firing rate a low degree node can behave as a functional hub, spreading its activity patterns polysynaptically in the network. Therefore, in an otherwise homogeneous and undirected network, firing rate is a tunable parameter that introduces directionality and enhances the reliability of signal transmission. The intrinsic firing rate across neuronal populations may thus determine preferred routes for signal transmission that can be easily controlled by changing the firing rate in specific nodes. We show that the results are generic and the same mechanism works in the networks with more complex topology.
Brain/physiology , Models, Neurological , Nerve Net/physiology , Neurons/physiology , Humans
Anticipated and zero-lag synchronization have been observed in different scientific fields. In the brain, they might play a fundamental role in information processing, temporal coding and spatial attention. Recent numerical work on anticipated and zero-lag synchronization studied the role of delays. However, an analytical understanding of the conditions for these phenomena remains elusive. In this paper, we study both phenomena in systems with small delays. By performing a phase reduction and studying phase locked solutions, we uncover the functional relation between the delay, excitation and inhibition for the onset of anticipated synchronization in a sender-receiver-interneuron motif. In the case of zero-lag synchronization in a chain motif, we determine the stability conditions. These analytical solutions provide an excellent prediction of the phase-locked regimes of Hodgkin-Huxley models and Roessler oscillators.
Certain differences between brain networks of healthy and epilectic subjects have been reported even during the interictal activity, in which no epileptic seizures occur. Here, magnetoencephalography (MEG) data recorded in the resting state is used to discriminate between healthy subjects and patients with either idiopathic generalized epilepsy or frontal focal epilepsy. Signal features extracted from interictal periods without any epileptiform activity are used to train a machine learning algorithm to draw a diagnosis. This is potentially relevant to patients without frequent or easily detectable spikes. To analyze the data, we use an up-to-date machine learning algorithm and explore the benefits of including different features obtained from the MEG data as inputs to the algorithm. We find that the relative power spectral density of the MEG time-series is sufficient to distinguish between healthy and epileptic subjects with a high prediction accuracy. We also find that a combination of features such as the phase-locked value and the relative power spectral density allow to discriminate generalized and focal epilepsy, when these features are calculated over a filtered version of the signals in certain frequency bands. Machine learning algorithms are currently being applied to the analysis and classification of brain signals. It is, however, less evident to identify the proper features of these signals that are prone to be used in such machine learning algorithms. Here, we evaluate the influence of the input feature selection on a clinical scenario to distinguish between healthy and epileptic subjects. Our results indicate that such distinction is possible with a high accuracy (86%), allowing the discrimination between idiopathic generalized and frontal focal epilepsy types.
Anticipated synchronization (AS) is a counterintuitive behavior that has been observed in several systems. When AS occurs in a sender-receiver configuration, the latter can predict the future dynamics of the former for certain parameter values. In particular, in neuroscience AS was proposed to explain the apparent discrepancy between information flow and time lag in the cortical activity recorded in monkeys. Despite its success, a clear understanding of the mechanisms yielding AS in neuronal circuits is still missing. Here we use the well-known phase-response-curve (PRC) approach to study the prototypical sender-receiver-interneuron neuronal motif. Our aim is to better understand how the transitions between delayed to anticipated synchronization and anticipated synchronization to phase-drift regimes occur. We construct a map based on the PRC method to predict the phase-locking regimes and their stability. We find that a PRC function of two variables, accounting simultaneously for the inputs from sender and interneuron into the receiver, is essential to reproduce the numerical results obtained using a Hodgkin-Huxley model for the neurons. On the contrary, the typical approximation that considers a sum of two independent single-variable PRCs fails for intermediate to high values of the inhibitory coupling strength of the interneuron. In particular, it loses the delayed-synchronization to anticipated-synchronization transition.
Cortical Synchronization/physiology , Models, Neurological , Neurons/physiology , Action Potentials , Animals , Computer Simulation , Neural Inhibition/physiology , Neural Pathways/physiology , Periodicity
Inferring effective connectivity from neurophysiological data is a challenging task. In particular, only a finite (and usually small) number of sites are simultaneously recorded, while the response of one of these sites can be influenced by other sites that are not being recorded. In the hippocampal formation, for instance, the connections between areas CA1-CA3, the dentate gyrus (DG), and the entorhinal cortex (EC) are well established. However, little is known about the relations within the EC layers, which might strongly affect the resulting effective connectivity estimations. In this work, we build excitatory/inhibitory neuronal populations representing the four areas CA1, CA3, the DG, and the EC and fix their connectivities. We model the EC by three layers (LII, LIII, and LV) and assume any possible connection between them. Our results, based on Granger Causality (GC) and Partial Transfer Entropy (PTE) measurements, reveal that the estimation of effective connectivity in the hippocampus strongly depends on the connectivities between EC layers. Moreover, we find, for certain EC configurations, very different results when comparing GC and PTE measurements. We further demonstrate that causal links can be robustly inferred regardless of the excitatory or inhibitory nature of the connection, adding complexity to their interpretation. Overall, our work highlights the importance of a careful analysis of the connectivity methods to prevent unrealistic conclusions when only partial information about the experimental system is available, as usually happens in brain networks. Our results suggest that the combination of causality measures with neuronal modeling based on precise neuroanatomical tracing may provide a powerful framework to disambiguate causal interactions in the brain.
Entorhinal Cortex/physiology , Hippocampus/physiology , Nerve Net/physiology , Animals , Interneurons/physiology , Models, Neurological , Rats , Time Factors
