Dynamic searching in the brain

Cognitive functions rely on the extensive use of information stored in the brain, and the searching for the relevant information for solving some problem is a very complex task. Human cognition largely uses biological search engines, and we assume that to study cognitive function we need to understand the way these brain search engines work. The approach we favor is to study multi-modular network models, able to solve particular problems that involve searching for information. The building blocks of these multimodular networks are the context dependent memory models we have been using for almost 20 years. These models work by associating an output to the Kronecker product of an input and a context. Input, context and output are vectors that represent cognitive variables. Our models constitute a natural extension of the traditional linear associator. We show that coding the information in vectors that are processed through association matrices, allows for a direct contact between these memory models and some procedures that are now classical in the Information Retrieval field. One essential feature of context-dependent models is that they are based on the thematic packing of information, whereby each context points to a particular set of related concepts. The thematic packing can be extended to multimodular networks involving input-output contexts, in order to accomplish more complex tasks. Contexts act as passwords that elicit the appropriate memory to deal with a query. We also show toy versions of several ‘neuromimetic’ devices that solve cognitive tasks as diverse as decision making or word sense disambiguation. The functioning of these multimodular networks can be described as dynamical systems at the level of cognitive variables.     

Research Implications of Science-Informed,Value-Based Decision Making

In ‘Hard’ science, scientists correctly operate as the ‘guardians of certainty,’ using hypothesis testing formulations and value judgements about error rates and time discounting that make classical inferential methods appropriate. But these methods can neither generate most of the inputs needed by decision makers in their time frame, nor generate them in a form that allows them to be integrated into the decision in an analytically coherent and transparent way. The need for transparent accountability in public decision making under uncertainty and value conflict means the analytical coherence provided by the stochastic Bayesian decision analytic approach, drawing on the outputs of Bayesian science, is needed. If scientific researchers are to play the rôle they should be playing in informing value-based decision making, they need to see themselves also as ‘guardians of uncertainty,’ ensuring that the best possible current posterior distributions on relevant parameters are made available for decision making, irrespective of the state of the certainty-seeking research. The paper distinguishes the actors employing different technologies in terms of the focus of the technology (knowledge, values, choice); the ‘home base’ mode of their activity on the cognitive continuum of varying analysis-to-intuition ratios; and the underlying value judgements of the activity (especially error loss functions and time discount rates). Those who propose any principle of decision making other than the banal ‘Best Principle,’ including the ‘Precautionary Principle,’ are properly interpreted as advocates seeking to have their own value judgements and preferences regarding mode location apply. The task for accountable decision makers, and their supporting technologists, is to determine the best course of action under the universal conditions of uncertainty and value difference/conflict.     

The frustrated brain: from dynamics on motifs to communities and networks

Cognitive function depends on an adaptive balance between flexible dynamics and integrative processes in distributed cortical networks. Patterns of zero-lag synchrony likely underpin numerous perceptual and cognitive functions. Synchronization fulfils integration by reducing entropy, while adaptive function mandates that a broad variety of stable states be readily accessible. Here, we elucidate two complementary influences on patterns of zero-lag synchrony that derive from basic properties of brain networks. First, mutually coupled pairs of neuronal subsystems—resonance pairs—promote stable zero-lag synchrony among the small motifs in which they are embedded, and whose effects can propagate along connected chains. Second, frustrated closed-loop motifs disrupt synchronous dynamics, enabling metastable configurations of zero-lag synchrony to coexist. We document these two complementary influences in small motifs and illustrate how these effects underpin stable versus metastable phase-synchronization patterns in prototypical modular networks and in large-scale cortical networks of the macaque (CoCoMac). We find that the variability of synchronization patterns depends on the inter-node time delay, increases with the network size and is maximized for intermediate coupling strengths. We hypothesize that the dialectic influences of resonance versus frustration may form a dynamic substrate for flexible neuronal integration, an essential platform across diverse cognitive processes.     

Mapping indicator models: From intuitive problem structuring to quantified decision-making in sustainable forest management

Most commonly, sustainability indicator sets presented as lists do not take into account interactions among indicators in a systematic manner. Vice versa, existing environmental indicator systems do not provide a formalized approach for problem structuring and quantitative decision support. In this paper, techniques for considering indicator relationships are highlighted and a coupled approach between a qualitative and a quantitative method is analysed. Cognitive mapping (CM) is used for structuring indicators and three different causal maps are derived based on established sustainability concepts: (a) criteria and indicators (C&I hierarchy), (b) indicator network, and (c) Driving Force-Pressure-State-Impact-Response (DPSIR) system. These maps are transferred to the Analytic Network Process (ANP) to allow their application in multi-criteria decision analysis (MCDA).In an application example, Pan-European indicators for sustainable forest management (SFM) are utilized in an ANP-based assessment. The effects of the model structure on the overall evaluation result are demonstrated by means of three reporting periods on Austrian forestry.In a comparative analysis of CM and ANP it is tested whether their measures of indicator significance do correspond. Both centrality in CM and single limited priorities in ANP have been reported to identify key indicators that play an important role in networks. We found out that the correspondence between CM and ANP is the stronger the more rigidly cause-effect relationships are interpreted, which is the case for the DPSIR system of SFM indicators.It is demonstrated that using indicator sets without consideration of the indicator interactions will cause shortcomings for evaluation and assessment procedures in SFM. Given strict and consistent definition of causal indicator relationships, a coupled use of CM and ANP is recommendable for both enhancing the process of problem structuring as well as supporting preference-based evaluation of decision alternatives.     

KEGG Mapper for inferring cellular functions from protein sequences

KEGG is a reference knowledge base for biological interpretation of large‐scale molecular datasets, such as genome and metagenome sequences. It accumulates experimental knowledge about high‐level functions of the cell and the organism represented in terms of KEGG molecular networks, including KEGG pathway maps, BRITE hierarchies, and KEGG modules. By the process called KEGG mapping, a set of protein coding genes in the genome, for example, can be converted to KEGG molecular networks enabling interpretation of cellular functions and other high‐level features. Here we report a new version of KEGG Mapper, a suite of KEGG mapping tools available at the KEGG website ( https://www.kegg.jp/ or https://www.genome.jp/kegg/ ), together with the KOALA family tools for automatic assignment of KO (KEGG Orthology) identifiers used in the mapping.     

Causal inference in biomolecular pathways using a Bayesian network approach and an Implicit method

We introduce here the concept of Implicit networks which provide, like Bayesian networks, a graphical modelling framework that encodes the joint probability distribution for a set of random variables within a directed acyclic graph. We show that Implicit networks, when used in conjunction with appropriate statistical techniques, are very attractive for their ability to understand and analyze biological data. Particularly, we consider here the use of Implicit networks for causal inference in biomolecular pathways. In such pathways, an Implicit network encodes dependencies among variables (proteins, genes), can be trained to learn causal relationships (regulation, interaction) between them and then used to predict the biological response given the status of some key proteins or genes in the network. We show that Implicit networks offer efficient methodologies for learning from observations without prior knowledge and thus provide a good alternative to classical inference in Bayesian networks when priors are missing. We illustrate our approach by an application to simulated data for a simplified signal transduction pathway of the epidermal growth factor receptor (EGFR) protein.     

Extracting representations of cognition across neuroimaging studies improves brain decoding

Cognitive brain imaging is accumulating datasets about the neural substrate of many different mental processes. Yet, most studies are based on few subjects and have low statistical power. Analyzing data across studies could bring more statistical power; yet the current brain-imaging analytic framework cannot be used at scale as it requires casting all cognitive tasks in a unified theoretical framework. We introduce a new methodology to analyze brain responses across tasks without a joint model of the psychological processes. The method boosts statistical power in small studies with specific cognitive focus by analyzing them jointly with large studies that probe less focal mental processes. Our approach improves decoding performance for 80% of 35 widely-different functional-imaging studies. It finds commonalities across tasks in a data-driven way, via common brain representations that predict mental processes. These are brain networks tuned to psychological manipulations. They outline interpretable and plausible brain structures. The extracted networks have been made available; they can be readily reused in new neuro-imaging studies. We provide a multi-study decoding tool to adapt to new data.     

The Decision to Engage Cognitive Control Is Driven by Expected Reward-Value: Neural and Behavioral Evidence

Cognitive control is a fundamental skill reflecting the active use of task-rules to guide behavior and suppress inappropriate automatic responses. Prior work has traditionally used paradigms in which subjects are told when to engage cognitive control. Thus, surprisingly little is known about the factors that influence individuals'' initial decision of whether or not to act in a reflective, rule-based manner. To examine this, we took three classic cognitive control tasks (Stroop, Wisconsin Card Sorting Task, Go/No-Go task) and created novel ‘free-choice’ versions in which human subjects were free to select an automatic, pre-potent action, or an action requiring rule-based cognitive control, and earned varying amounts of money based on their choices. Our findings demonstrated that subjects'' decision to engage cognitive control was driven by an explicit representation of monetary rewards expected to be obtained from rule-use. Subjects rarely engaged cognitive control when the expected outcome was of equal or lesser value as compared to the value of the automatic response, but frequently engaged cognitive control when it was expected to yield a larger monetary outcome. Additionally, we exploited fMRI-adaptation to show that the lateral prefrontal cortex (LPFC) represents associations between rules and expected reward outcomes. Together, these findings suggest that individuals are more likely to act in a reflective, rule-based manner when they expect that it will result in a desired outcome. Thus, choosing to exert cognitive control is not simply a matter of reason and willpower, but rather, conforms to standard mechanisms of value-based decision making. Finally, in contrast to current models of LPFC function, our results suggest that the LPFC plays a direct role in representing motivational incentives.     

miniTUBA: medical inference by network integration of temporal data using Bayesian analysis

MOTIVATION: Many biomedical and clinical research problems involve discovering causal relationships between observations gathered from temporal events. Dynamic Bayesian networks are a powerful modeling approach to describe causal or apparently causal relationships, and support complex medical inference, such as future response prediction, automated learning, and rational decision making. Although many engines exist for creating Bayesian networks, most require a local installation and significant data manipulation to be practical for a general biologist or clinician. No software pipeline currently exists for interpretation and inference of dynamic Bayesian networks learned from biomedical and clinical data. RESULTS: miniTUBA is a web-based modeling system that allows clinical and biomedical researchers to perform complex medical/clinical inference and prediction using dynamic Bayesian network analysis with temporal datasets. The software allows users to choose different analysis parameters (e.g. Markov lags and prior topology), and continuously update their data and refine their results. miniTUBA can make temporal predictions to suggest interventions based on an automated learning process pipeline using all data provided. Preliminary tests using synthetic data and laboratory research data indicate that miniTUBA accurately identifies regulatory network structures from temporal data. AVAILABILITY: miniTUBA is available at http://www.minituba.org.     

BNFinder: exact and efficient method for learning Bayesian networks

MOTIVATION: Bayesian methods are widely used in many different areas of research. Recently, it has become a very popular tool for biological network reconstruction, due to its ability to handle noisy data. Even though there are many software packages allowing for Bayesian network reconstruction, only few of them are freely available to researchers. Moreover, they usually require at least basic programming abilities, which restricts their potential user base. Our goal was to provide software which would be freely available, efficient and usable to non-programmers. RESULTS: We present a BNFinder software, which allows for Bayesian network reconstruction from experimental data. It supports dynamic Bayesian networks and, if the variables are partially ordered, also static Bayesian networks. The main advantage of BNFinder is the use exact algorithm, which is at the same time very efficient (polynomial with respect to the number of observations).     

Cumulative disease progression models for cross‐sectional data: A review and comparison

A better understanding of disease progression is beneficial for early diagnosis and appropriate individual therapy. Many different approaches for statistical modelling of cumulative disease progression have been proposed in the literature, including simple path models up to complex restricted Bayesian networks. Important fields of application are diseases such as cancer and HIV. Tumour progression is measured by means of chromosome aberrations, whereas people infected with HIV develop drug resistances because of genetic changes of the HI‐virus. These two very different diseases have typical courses of disease progression, which can be modelled partly by consecutive and partly by independent steps. This paper gives an overview of the different progression models and points out their advantages and drawbacks. Different models are compared via simulations to analyse how they work if some of their assumptions are violated. In a simulation study, we evaluate how models perform in terms of fitting induced multivariate probability distributions and topological relationships. We often find that the true model class used for generating data is outperformed by either a less or a more complex model class. The more flexible conjunctive Bayesian networks can be used to fit oncogenetic trees, whereas mixtures of oncogenetic trees with three tree components can be well fitted by mixture models with only two tree components.     

From heterogeneous healthcare data to disease-specific biomarker networks: A hierarchical Bayesian network approach

In this work, we introduce an entirely data-driven and automated approach to reveal disease-associated biomarker and risk factor networks from heterogeneous and high-dimensional healthcare data. Our workflow is based on Bayesian networks, which are a popular tool for analyzing the interplay of biomarkers. Usually, data require extensive manual preprocessing and dimension reduction to allow for effective learning of Bayesian networks. For heterogeneous data, this preprocessing is hard to automatize and typically requires domain-specific prior knowledge. We here combine Bayesian network learning with hierarchical variable clustering in order to detect groups of similar features and learn interactions between them entirely automated. We present an optimization algorithm for the adaptive refinement of such group Bayesian networks to account for a specific target variable, like a disease. The combination of Bayesian networks, clustering, and refinement yields low-dimensional but disease-specific interaction networks. These networks provide easily interpretable, yet accurate models of biomarker interdependencies. We test our method extensively on simulated data, as well as on data from the Study of Health in Pomerania (SHIP-TREND), and demonstrate its effectiveness using non-alcoholic fatty liver disease and hypertension as examples. We show that the group network models outperform available biomarker scores, while at the same time, they provide an easily interpretable interaction network.     

Exploring the protein–protein interaction landscape in plants

Protein–protein interactions (PPIs) represent an essential aspect of plant systems biology. Identification of key protein players and their interaction networks provide crucial insights into the regulation of plant developmental processes and into interactions of plants with their environment. Despite the great advance in the methods for the discovery and validation of PPIs, still several challenges remain. First, the PPI networks are usually highly dynamic, and the in vivo interactions are often transient and difficult to detect. Therefore, the properties of the PPIs under study need to be considered to select the most suitable technique, because each has its own advantages and limitations. Second, besides knowledge on the interacting partners of a protein of interest, characteristics of the interaction, such as the spatial or temporal dynamics, are highly important. Hence, multiple approaches have to be combined to obtain a comprehensive view on the PPI network present in a cell. Here, we present the progress in commonly used methods to detect and validate PPIs in plants with a special emphasis on the PPI features assessed in each approach and how they were or can be used for the study of plant interactions with their environment.     

Evaluation of Bayesian networks for modelling habitat suitability and management of a protected area

To be effective, management of protected areas should be based on the best available evidence, including the scientific literature and expert knowledge. However, lack of such evidence in a suitable form to support decision-making may hinder effective management. Here we examine the use of Bayesian networks to support the management of protected areas, through the development of habitat suitability models for eight species of conservation concern. Bayesian networks were constructed on the basis of the scientific literature and expert knowledge, and were then tested using results from a field survey. Models of all species demonstrated very high discrimination between presence and absence sites, as indicated by AUC values >0.8, with values >0.9 obtained for four species, and Kappa values in the range of 0.4–0.9. The Bayesian networks were then used to examine the impact of different management interventions on habitat suitability of each species, including tree cutting, grazing and burning. Species differed in terms of their sensitivity to different management interventions, and model output provided evidence of both negative and positive interactions between types of intervention. These results highlight the trade-offs that must often be made when undertaking conservation management, and demonstrate the value of Bayesian networks in helping to make such trade-offs explicit. The identification of management impacts through analysis of available evidence also demonstrates the value of Bayesian networks for supporting evidence-based approaches to protected area management.     

Preparation of Parasagittal Slices for the Investigation of Dorsal-ventral Organization of the Rodent Medial Entorhinal Cortex

Computation in the brain relies on neurons responding appropriately to their synaptic inputs. Neurons differ in their complement and distribution of membrane ion channels that determine how they respond to synaptic inputs. However, the relationship between these cellular properties and neuronal function in behaving animals is not well understood. One approach to this problem is to investigate topographically organized neural circuits in which the position of individual neurons maps onto information they encode or computations they carry out¹. Experiments using this approach suggest principles for tuning of synaptic responses underlying information encoding in sensory and cognitive circuits^2,3.The topographical organization of spatial representations along the dorsal-ventral axis of the medial entorhinal cortex (MEC) provides an opportunity to establish relationships between cellular mechanisms and computations important for spatial cognition. Neurons in layer II of the rodent MEC encode location using grid-like firing fields^4-6. For neurons found at dorsal positions in the MEC the distance between the individual firing fields that form a grid is on the order of 30 cm, whereas for neurons at progressively more ventral positions this distance increases to greater than 1 m. Several studies have revealed cellular properties of neurons in layer II of the MEC that, like the spacing between grid firing fields, also differ according to their dorsal-ventral position, suggesting that these cellular properties are important for spatial computation^2,7-10.Here we describe procedures for preparation and electrophysiological recording from brain slices that maintain the dorsal-ventral extent of the MEC enabling investigation of the topographical organization of biophysical and anatomical properties of MEC neurons. The dorsal-ventral position of identified neurons relative to anatomical landmarks is difficult to establish accurately with protocols that use horizontal slices of MEC^7,8,11,12, as it is difficult to establish reference points for the exact dorsal-ventral location of the slice. The procedures we describe enable accurate and consistent measurement of location of recorded cells along the dorsal-ventral axis of the MEC as well as visualization of molecular gradients^2,10. The procedures have been developed for use with adult mice (> 28 days) and have been successfully employed with mice up to 1.5 years old. With adjustments they could be used with younger mice or other rodent species. A standardized system of preparation and measurement will aid systematic investigation of the cellular and microcircuit properties of this area.     

Metacognition in the rat

The ability to reflect on one's own mental processes, termed metacognition, is a defining feature of human existence [1, 2]. Consequently, a fundamental question in comparative cognition is whether nonhuman animals have knowledge of their own cognitive states [3]. Recent evidence suggests that people and nonhuman primates [4-8] but not less "cognitively sophisticated" species [3, 9, 10] are capable of metacognition. Here, we demonstrate for the first time that rats are capable of metacognition--i.e., they know when they do not know the answer in a duration-discrimination test. Before taking the duration test, rats were given the opportunity to decline the test. On other trials, they were not given the option to decline the test. Accurate performance on the duration test yielded a large reward, whereas inaccurate performance resulted in no reward. Declining a test yielded a small but guaranteed reward. If rats possess knowledge regarding whether they know the answer to the test, they would be expected to decline most frequently on difficult tests and show lowest accuracy on difficult tests that cannot be declined [4]. Our data provide evidence for both predictions and suggest that a nonprimate has knowledge of its own cognitive state.     

H-CORE: enabling genome-scale Bayesian analysis of biological systems without prior knowledge

The Bayesian network is a popular tool for describing relationships between data entities by representing probabilistic (in)dependencies with a directed acyclic graph (DAG) structure. Relationships have been inferred between biological entities using the Bayesian network model with high-throughput data from biological systems in diverse fields. However, the scalability of those approaches is seriously restricted because of the huge search space for finding an optimal DAG structure in the process of Bayesian network learning. For this reason, most previous approaches limit the number of target entities or use additional knowledge to restrict the search space. In this paper, we use the hierarchical clustering and order restriction (H-CORE) method for the learning of large Bayesian networks by clustering entities and restricting edge directions between those clusters, with the aim of overcoming the scalability problem and thus making it possible to perform genome-scale Bayesian network analysis without additional biological knowledge. We use simulations to show that H-CORE is much faster than the widely used sparse candidate method, whilst being of comparable quality. We have also applied H-CORE to retrieving gene-to-gene relationships in a biological system (The 'Rosetta compendium'). By evaluating learned information through literature mining, we demonstrate that H-CORE enables the genome-scale Bayesian analysis of biological systems without any prior knowledge.     

Prediction of contact maps by GIOHMMs and recurrent neural networks using lateral propagation from all four cardinal corners

MOTIVATION: Accurate prediction of protein contact maps is an important step in computational structural proteomics. Because contact maps provide a translation and rotation invariant topological representation of a protein, they can be used as a fundamental intermediary step in protein structure prediction. RESULTS: We develop a new set of flexible machine learning architectures for the prediction of contact maps, as well as other information processing and pattern recognition tasks. The architectures can be viewed as recurrent neural network implemantations of a class of Bayesian networks we call generalized input-output HMMs (GIOHMMs). For the specific case of contact maps, contextual information is propagated laterally through four hidden planes, one for each cardinal corner. We show that these architectures can be trained from examples and yield contact map predictors that outperform previously reported methods. While several extensions and improvements are in progress, the current version can accurately predict 60.5% of contacts at a distance cutoff of 8 A and 45% of distant contacts at 10 A, for proteins of length up to 300.