Different eigenvalue distributions encode the same temporal tasks in recurrent neural networks.
Ontology highlight
ABSTRACT: Supplementary information
The online version contains supplementary material available at 10.1007/s11571-022-09802-5.
SUBMITTER: Jarne C
PROVIDER: S-EPMC9020562 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Training recurrent neural networks (RNNs) has become a go-to approach for generating and evaluating mechanistic neural hypotheses for cognition. The ease and efficiency of training RNNs with backpropagation through time and the availability of robustly supported deep learning libraries has made RNN modeling more approachable and accessible to neuroscience. Yet, a major technical hindrance remains. Cognitive processes such as working memory and decision making involve neural population dynamics over a long period of time within a behavioral trial and across trials. It is difficult to train RNNs to accomplish tasks where neural representations and dynamics have long temporal dependencies without gating mechanisms such as LSTMs or GRUs which currently lack experimental support and prohibit direct comparison between RNNs and biological neural circuits. We tackled this problem based on the idea of specialized skip-connections through time to support the emergence of task-relevant dynamics, and subsequently reinstitute biological plausibility by reverting to the original architecture. We show that this approach enables RNNs to successfully learn cognitive tasks that prove impractical if not impossible to learn using conventional methods. Over numerous tasks considered here, we achieve less training steps and shorter wall-clock times, particularly in tasks that require learning long-term dependencies via temporal integration over long timescales or maintaining a memory of past events in hidden-states. Our methods expand the range of experimental tasks that biologically plausible RNN models can learn, thereby supporting the development of theory for the emergent neural mechanisms of computations involving long-term dependencies.
| S-EPMC10592728 | biostudies-literature
Project description:A recurrent spiking neural network is proposed that implements planning as probabilistic inference for finite and infinite horizon tasks. The architecture splits this problem into two parts: The stochastic transient firing of the network embodies the dynamics of the planning task. With appropriate injected input this dynamics is shaped to generate high-reward state trajectories. A general class of reward-modulated plasticity rules for these afferent synapses is presented. The updates optimize the likelihood of getting a reward through a variant of an Expectation Maximization algorithm and learning is guaranteed to convergence to a local maximum. We find that the network dynamics are qualitatively similar to transient firing patterns during planning and foraging in the hippocampus of awake behaving rats. The model extends classical attractor models and provides a testable prediction on identifying modulating contextual information. In a real robot arm reaching and obstacle avoidance task the ability to represent multiple task solutions is investigated. The neural planning method with its local update rules provides the basis for future neuromorphic hardware implementations with promising potentials like large data processing abilities and early initiation of strategies to avoid dangerous situations in robot co-worker scenarios.
| S-EPMC4758071 | biostudies-literature
Project description:The ability to simultaneously record from large numbers of neurons in behaving animals has ushered in a new era for the study of the neural circuit mechanisms underlying cognitive functions. One promising approach to uncovering the dynamical and computational principles governing population responses is to analyze model recurrent neural networks (RNNs) that have been optimized to perform the same tasks as behaving animals. Because the optimization of network parameters specifies the desired output but not the manner in which to achieve this output, "trained" networks serve as a source of mechanistic hypotheses and a testing ground for data analyses that link neural computation to behavior. Complete access to the activity and connectivity of the circuit, and the ability to manipulate them arbitrarily, make trained networks a convenient proxy for biological circuits and a valuable platform for theoretical investigation. However, existing RNNs lack basic biological features such as the distinction between excitatory and inhibitory units (Dale's principle), which are essential if RNNs are to provide insights into the operation of biological circuits. Moreover, trained networks can achieve the same behavioral performance but differ substantially in their structure and dynamics, highlighting the need for a simple and flexible framework for the exploratory training of RNNs. Here, we describe a framework for gradient descent-based training of excitatory-inhibitory RNNs that can incorporate a variety of biological knowledge. We provide an implementation based on the machine learning library Theano, whose automatic differentiation capabilities facilitate modifications and extensions. We validate this framework by applying it to well-known experimental paradigms such as perceptual decision-making, context-dependent integration, multisensory integration, parametric working memory, and motor sequence generation. Our results demonstrate the wide range of neural activity patterns and behavior that can be modeled, and suggest a unified setting in which diverse cognitive computations and mechanisms can be studied.
| S-EPMC4771709 | biostudies-literature
Project description:Working memory and response inhibition are functions that mature relatively late in life, after adolescence, paralleling the maturation of the prefrontal cortex. The link between behavioral and neural maturation is not obvious, however, making it challenging to understand how neural activity underlies the maturation of cognitive function. To gain insights into the nature of observed changes in prefrontal activity between adolescence and adulthood, we investigated the progressive changes in unit activity of recurrent neural networks as they were trained to perform working memory and response inhibition tasks. These included increased delay period activity during working memory tasks and increased activation in antisaccade tasks. These findings reveal universal properties underlying the neuronal computations behind cognitive tasks and explicate the nature of changes that occur as the result of developmental maturation.
| S-EPMC8506971 | biostudies-literature
Project description:Electroencephalogram (EEG) microstates that represent quasi-stable, global neuronal activity are considered as the building blocks of brain dynamics. Therefore, the analysis of microstate sequences is a promising approach to understand fast brain dynamics that underlie various mental processes. Recent studies suggest that EEG microstate sequences are non-Markovian and nonstationary, highlighting the importance of the sequential flow of information between different brain states. These findings inspired us to model these sequences using Recurrent Neural Networks (RNNs) consisting of long-short-term-memory (LSTM) units to capture the complex temporal dependencies. Using an LSTM-based auto encoder framework and different encoding schemes, we modeled the microstate sequences at multiple time scales (200-2,000 ms) aiming to capture stably recurring microstate patterns within and across subjects. We show that RNNs can learn underlying microstate patterns with high accuracy and that the microstate trajectories are subject invariant at shorter time scales (≤400 ms) and reproducible across sessions. Significant drop in the reconstruction accuracy was observed for longer sequence lengths of 2,000 ms. These findings indirectly corroborate earlier studies which indicated that EEG microstate sequences exhibit long-range dependencies with finite memory content. Furthermore, we find that the latent representations learned by the RNNs are sensitive to external stimulation such as stress while the conventional univariate microstate measures (e.g., occurrence, mean duration, etc.) fail to capture such changes in brain dynamics. While RNNs cannot be configured to identify the specific discriminating patterns, they have the potential for learning the underlying temporal dynamics and are sensitive to sequence aberrations characterized by changes in metal processes. Empowered with the macroscopic understanding of the temporal dynamics that extends beyond short-term interactions, RNNs offer a reliable alternative for exploring system level brain dynamics using EEG microstate sequences.
| S-EPMC7267981 | biostudies-literature
Project description:BackgroundThe transport through a quantum-scale device may be uniquely characterized by its transmission eigenvalues ?(n). Recently, highly conductive single-molecule junctions (SMJ) with multiple transport channels (i.e., several ?(n) > 0) have been formed from benzene molecules between Pt electrodes. Transport through these multichannel SMJs is a probe of both the bonding properties at the lead-molecule interface and of the molecular symmetry.ResultsWe use a many-body theory that properly describes the complementary wave-particle nature of the electron to investigate transport in an ensemble of Pt-benzene-Pt junctions. We utilize an effective-field theory of interacting ?-electrons to accurately model the electrostatic influence of the leads, and we develop an ab initio tunneling model to describe the details of the lead-molecule bonding over an ensemble of junction geometries. We also develop a simple decomposition of transmission eigenchannels into molecular resonances based on the isolated resonance approximation, which helps to illustrate the workings of our many-body theory, and facilitates unambiguous interpretation of transmission spectra.ConclusionWe confirm that Pt-benzene-Pt junctions have two dominant transmission channels, with only a small contribution from a third channel with ?(n) << 1. In addition, we demonstrate that the isolated resonance approximation is extremely accurate and determine that transport occurs predominantly via the HOMO orbital in Pt-benzene-Pt junctions. Finally, we show that the transport occurs in a lead-molecule coupling regime where the charge carriers are both particle-like and wave-like simultaneously, requiring a many-body description.
| S-EPMC3304317 | biostudies-literature
Project description:Combinatorial generalization-the ability to understand and produce novel combinations of already familiar elements-is considered to be a core capacity of the human mind and a major challenge to neural network models. A significant body of research suggests that conventional neural networks cannot solve this problem unless they are endowed with mechanisms specifically engineered for the purpose of representing symbols. In this paper, we introduce a novel way of representing symbolic structures in connectionist terms-the vectors approach to representing symbols (VARS), which allows training standard neural architectures to encode symbolic knowledge explicitly at their output layers. In two simulations, we show that neural networks not only can learn to produce VARS representations, but in doing so they achieve combinatorial generalization in their symbolic and non-symbolic output. This adds to other recent work that has shown improved combinatorial generalization under some training conditions, and raises the question of whether specific mechanisms or training routines are needed to support symbolic processing. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.
| S-EPMC6939346 | biostudies-literature
Project description:The ability to understand similarities and analogies is a fundamental aspect of human advanced cognition. Although subject of considerable research in comparative cognition, the extent to which nonhuman species are capable of analogical reasoning is still debated. This study examined the conditions under which tufted capuchin monkeys (Cebus apella) acquire a same/different concept in a matching-to-sample task on the basis of relational similarity among multi-item stimuli. We evaluated (i) the ability of five capuchin monkeys to learn the same/different concept on the basis of the number of items composing the stimuli and (ii) the ability to match novel stimuli after training with both several small stimulus sets and a large stimulus set. We found the first evidence of same/different relational matching-to-sample abilities in a New World monkey and demonstrated that the ability to match novel stimuli is within the capacity of this species. Therefore, analogical reasoning can emerge in monkeys under specific training conditions.
| S-EPMC3156813 | biostudies-literature
Project description:Entropy is an important index for describing the structure, function, and evolution of network. The existing research on entropy is primarily applied to undirected networks. Compared with an undirected network, a directed network involves a special asymmetric transfer. The research on the entropy of directed networks is very significant to effectively quantify the structural information of the whole network. Typical complex network models include nearest-neighbour coupling network, small-world network, scale-free network, and random network. These network models are abstracted as undirected graphs without considering the direction of node connection. For complex networks, modeling through the direction of network nodes is extremely challenging. In this paper, based on these typical models of complex network, a directed network model considering node connection in-direction is proposed, and the eigenvalue entropies of three matrices in the directed network is defined and studied, where the three matrices are adjacency matrix, in-degree Laplacian matrix and in-degree signless Laplacian matrix. The eigenvalue-based entropies of three matrices are calculated in directed nearest-neighbor coupling, directed small world, directed scale-free and directed random networks. Through the simulation experiment on the real directed network, the result shows that the eigenvalue entropy of the real directed network is between the eigenvalue entropy of directed scale-free network and directed small-world network.
| S-EPMC8216510 | biostudies-literature
Project description:Network analysis has become an increasingly prevalent research tool across a vast range of scientific fields. Here, we focus on the particular issue of comparing network statistics, i.e. graph-level measures of network structural features, across multiple networks that differ in size. Although "normalized" versions of some network statistics exist, we demonstrate via simulation why direct comparison is often inappropriate. We consider normalizing network statistics relative to a simple fully parameterized reference distribution and demonstrate via simulation how this is an improvement over direct comparison, but still sometimes problematic. We propose a new adjustment method based on a reference distribution constructed as a mixture model of random graphs which reflect the dependence structure exhibited in the observed networks. We show that using simple Bernoulli models as mixture components in this reference distribution can provide adjusted network statistics that are relatively comparable across different network sizes but still describe interesting features of networks, and that this can be accomplished at relatively low computational expense. Finally, we apply this methodology to a collection of ecological networks derived from the Los Angeles Family and Neighborhood Survey activity location data.
| S-EPMC5052017 | biostudies-literature