In this paper, I draw a parallel between the stability of physical systems and that of economic ones, such as the US financial system. I argue that the use of equilibrium assumptions is central to the analysis of dynamic behavior for both kinds of systems, and that we ought to interpret such idealizing strategies as footholds for causal exploration and explanation. Our considerations suggest multi-scale modeling as a natural home for such reasoning strategies, which can provide a (...) backdrop for the assessment and prevention of financial crises. Equilibrium assumptions are critical elements of the epistemic scaffolding that make up multi-scale models, which we should understand as a means of constructing explanations of dynamic behavior. (shrink)

A common reductionist assumption is that macro-scale behaviors can be described "bottom-up" if only sufficient details about lower-scale processes are available. The view that an "ideal" or "fundamental" physics would be sufficient to explain all macro-scale phenomena has been met with criticism from philosophers of biology. Specifically, scholars have pointed to the impossibility of deducing biological explanations from physical ones, and to the irreducible nature of distinctively biological processes such as gene regulation and evolution. This paper takes (...) a step back in asking whether bottom-up modeling is feasible even when modeling simple physical systems across scales. By comparing examples of multi-scale modeling in physics and biology, we argue that the “tyranny of scales” problem presents a challenge to reductive explanations in both physics and biology. The problem refers to the scale-dependency of physical and biological behaviors that forces researchers to combine different models relying on different scale-specific mathematical strategies and boundary conditions. Analyzing the ways in which different models are combined in multi-scale modeling also has implications for the relation between physics and biology. Contrary to the assumption that physical science approaches provide reductive explanations in biology, we exemplify how inputs from physics often reveal the importance of macro-scale models and explanations. We illustrate this through an examination of the role of biomechanics modeling in developmental biology. In such contexts, the relation between models at different scales and from different disciplines is neither reductive nor completely autonomous, but interdependent. (shrink)

Systems biologists often distance themselves from reductionist approaches and formulate their aim as understanding living systems “as a whole.” Yet, it is often unclear what kind of reductionism they have in mind, and in what sense their methodologies would offer a superior approach. To address these questions, we distinguish between two types of reductionism which we call “modular reductionism” and “bottom-up reductionism.” Much knowledge in molecular biology has been gained by decomposing living systems into functional modules or through detailed studies (...) of molecular processes. We ask whether systems biology provides novel ways to recompose these findings in the context of the system as a whole via computational simulations. As an example of computational integration of modules, we analyze the first whole-cell model of the bacterium M. genitalium. Secondly, we examine the attempt to recompose processes across different spatial scales via multi-scale cardiac models. Although these models rely on a number of idealizations and simplifying assumptions as well, we argue that they provide insight into the limitations of reductionist approaches. Whole-cell models can be used to discover properties arising at the interfaces of dynamically coupled processes within a biological system, thereby making more apparent what is lost through decomposition. Similarly, multi-scale modeling highlights the relevance of macroscale parameters and models and challenges the view that living systems can be understood “bottom-up.” Specifically, we point out that system-level properties constrain lower-scale processes. Thus, large-scale modeling reveals how living systems at the same time are more and less than the sum of the parts. (shrink)

As a philosophical paradigm, differential heterogenesis offers us a novel descriptive vantage with which to inscribe Deleuze’s virtuality within the terrain of “differential becoming,” conjugating “pure saliences” so as to parse economies, microhistories, insurgencies, and epistemological evolutionary processes that can be conceived of independently from their representational form. Unlike Gestalt theory’s oppositional constructions, the advantage of this aperture is that it posits a dynamic context to both media and its analysis, rendering them functionally tractable and set in relation to other (...) objects, rather than as sedentary identities. Surveying the genealogy of differential heterogenesis with particular interest in the legacy of Lautman’s dialectic, I make the case for a reading of the Deleuzean virtual that departs from an event-oriented approach, galvanizing Sarti and Citti’s dynamic a priori vis-à-vis Deleuze’s philosophy of difference. Specifically, I posit differential heterogenesis as frame with which to examine our contemporaneous epistemic shift as it relates to multi-scalar computational modeling while paying particular attention to neuro-inferential modes of inductive learning and homologous cognitive architecture. Carving a bricolage between Mark Wilson’s work on the “greediness of scales” and Deleuze’s “scales of reality”, this project threads between static ecologies and active externalism vis-à-vis endocentric frames of reference and syntactical scaffolding. (shrink)

Interactions in service ecosystems, as opposed to the service dyad, have recently gained much attention from research. However, it is still unclear how they influence a customer’s experiential value and trigger desired prosocial behavior. The purpose of this study is to identify which elements of the multi-actor service ecosystem contribute to a customer’s experiential value and to investigate its relation to a customer’s interaction attitude and inter-customer helping behavior. The authors adopted a scale development procedure from the existing (...) literature. Service, brand, retail and tourism management research as well as expert feedback is used to generate a pool of 33 items. Exploratory and confirmatory factor analysis were conducted. The scale was validated based on more than 468 responses to a CASI at one of the world’s largest trade shows. Scale-development procedure was followed by structural equation modeling. CFA supports that experiential value in multi-actor ecosystems comprises five dimensions. The functional value of personnel, the perception of other customers’ appearance, the perception of other customers’ behavior, multisensory stimuli, and a customer’s enjoyment. Experiential value positively and directly relates to a customer’s interaction attitude and inter-customer helping behavior. Furthermore, the effect of experiential value on inter-customer helping behavior is partially mediated by interaction attitude. Managers interested in getting more out of interactions with customers will develop an understanding for the interplay between the physical environment and individuals within a multi-actor ecosystem. Social scientists and managers can use the scale to assess experiential value, encourage a customer’s interaction attitude and utilize the customers’ influence on their peers. This paper synthesizes insights from existing research on experiential value, from various fields, in one scale. This holistic approach is the first to simultaneously account for a customer’s interactions with the multisensory physical environment, personal interactions with employees and interactions between customers in a multi-actor service ecosystem. (shrink)

Multi-level institutional designs with distributed power and authority arrangements among federal, state, regional, and local government agencies could lead to the emergence of differential patterns of socioeconomic and infrastructure development pathways in complex social–ecological systems. Both exogenous drivers and endogenous processes in social–ecological systems can lead to changes in the number of “basins of attraction,” changes in the positions of the basins within the state space, and changes in the positions of the thresholds between basins. In an effort to (...) advance the theory and practice of the governance of policy systems, this study addresses a narrower empirical question: how do intergovernmental institutional rules set by federal, state, and regional government agencies generate and sustain basins of attraction in funding infrastructure projects? A pattern-oriented, agent-based model (ABM) of an intergovernmental network has been developed to simulate real-world transportation policy implementation processes across the federal, the state of Vermont, regional, and local governments for prioritizing transportation projects. The ABM simulates baseline and alternative intergovernmental institutional rule structures and assesses their impacts on financial investment flows. The ABM was calibrated with data from multiple focus groups, individual interviews, and analysis of federal, state, and regional scale transportation projects and programs. The results from experimental simulations are presented to test system-wide effects of alternative multi-level institutional designs, in particular different power and authority arrangements between state and regional governments, on the emergence of roadway project prioritization patterns and funding allocations across regions and towns. (shrink)

Cancer is a complex disease, necessitating research on many different levels; at the subcellular level to identify genes, proteins and signaling pathways associated with the disease; at the cellular level to identify, for example, cell-cell adhesion and communication mechanisms; at the tissue level to investigate disruption of homeostasis and interaction with the tissue of origin or settlement of metastasis; and finally at the systems level to explore its global impact, e.g. through the mechanism of cachexia. Mathematical models have been proposed (...) to identify key mechanisms that underlie dynamics and events at every scale of interest, and increasing effort is now being paid to multi-scale models that bridge the different scales. With more biological data becoming available and with increased interdisciplinary efforts, theoretical models are rendering suitable tools to predict the origin and course of the disease. The ultimate aims of cancer models, however, are to enlighten our concept of the carcinogenesis process and to assist in the designing of treatment protocols that can reduce mortality and improve patient quality of life. Conventional treatment of cancer is surgery combined with radiotherapy or chemotherapy for localized tumors or systemic treatment of advanced cancers, respectively. Although radiation is widely used as treatment, most scheduling is based on empirical knowledge and less on the predictions of sophisticated growth dynamical models of treatment response. Part of the failure to translate modeling research to the clinic may stem from language barriers, exacerbated by often esoteric model renderings with inaccessible parameterization. Here we discuss some ideas for combining tractable dynamical tumor growth models with radiation response models using biologically accessible parameters to provide a more intuitive and exploitable framework for understanding the complexity of radiotherapy treatment and failure. (shrink)

Mesoscale modeling is often considered merely as a practical strategy used when information on lower-scale details is lacking, or when there is a need to make models cognitively or computationally tractable. Without dismissing the importance of practical constraints for modeling choices, we argue that mesoscale models should not just be considered as abbreviations or placeholders for more “complete” models. Because many systems exhibit different behaviors at various spatial and temporal scales, bottom-up approaches are almost always doomed to (...) fail. Mesoscale models capture aspects of multi-scale systems that cannot be parameterized by simple averaging of lower-scale details. To understand the behavior of multi-scale systems, it is essential to identify mesoscale parameters that “code for” lower-scale details in a way that relate phenomena intermediate between microscopic and macroscopic features. We illustrate this point using examples of modeling of multi-scale systems in materials science and biology, where identification of material parameters such as stiffness or strain is a central step. The examples illustrate important aspects of a so-called “middle-out” modeling strategy. Rather than attempting to model the system bottom-up, one starts at intermediate scales where systems exhibit behaviors distinct from those at the atomic and continuum scales. One then seeks to upscale and downscale to gain a more complete understanding of the multi-scale system. The cases highlight how parameterization of lower-scale details not only enables tractable modeling but is also central to understanding functional and organizational features of multi-scale systems. (shrink)

Accounts of mental disorders focusing either on the brain as neurophysiological substrate or on systematic connections between symptoms are insufficient to account for the multifactorial nature of mental illnesses. Recently, multiplexes have been suggested to provide a holistic view of psychopathology that integrates data from different factors, at different scales, or across time. Intuitively, these multi-layered network structures present quite appealing models of mental disorders that can be constructed by powerful computational machinery based on increasing amounts of real-world data. (...) In this paper, I systematically examine what challenges psychopathology models face and to what extent different species of psychopathology models can address them. My analysis highlights that while multiplexes, as they are usually conceived, appear promising, they suffer from the same problems as other approaches. To remedy this, I suggest, we must go a step further and combine different kinds of multiplexes into 4D models. Once we embrace 4D multiplexes and identify appropriate ways to constrain them, we might unlock the true potential of multiplexes for making headway in psychopathology research. (shrink)

We review some of the main implications of the free-energy principle (FEP) for the study of the self-organization of living systems – and how the FEP can help us to understand (and model) biotic self-organization across the many temporal and spatial scales over which life exists. In order to maintain its integrity as a bounded system, any biological system - from single cells to complex organisms and societies - has to limit the disorder or dispersion (i.e., the long-run entropy) of (...) its constituent states. We review how this can be achieved by living systems that minimize their variational free energy. Variational free energy is an information theoretic construct, originally introduced into theoretical neuroscience and biology to explain perception, action, and learning. It has since been extended to explain the evolution, development, form, and function of entire organisms, providing a principled model of biotic self-organization and autopoiesis. It has provided insights into biological systems across spatiotemporal scales, ranging from microscales (e.g., sub- and multicellular dynamics), to intermediate scales (e.g., groups of interacting animals and culture), through to macroscale phenomena (the evolution of entire species). A crucial corollary of the FEP is that an organism just is (i.e., embodies or entails) an implicit model of its environment. As such, organisms come to embody causal relationships of their ecological niche, which, in turn, is influenced by their resulting behaviors. Crucially, free-energy minimization can be shown to be equivalent to the maximization of Bayesian model evidence. This allows us to cast natural selection in terms of Bayesian model selection, providing a robust theoretical account of how organisms come to match or accommodate the spatiotemporal complexity of their surrounding niche. In line with the theme of this volume; namely, biological complexity and self-organization, this chapter will examine a variational approach to self-organization across multiple dynamical scales. (shrink)

In the nation of Belize, and in particular the south side of Belize City, the main metropolitan area of the nation, significant economic disparities have led to child and adolescent exposure to high rates of violent crime, gang activity, unsafe neighborhoods, sexual, and physical violence. Problems associated with poor Social-Emotional Character Development are especially prevalent among boys. Consequently, valid culture-relevant measures are required that identify problematic behavior for policy-based intervention and evaluation of educational programs designed to ameliorate this problem. The (...) present study demonstrates the application of Exploratory Structural Equation Modeling to existing measures through the investigation of structural validity and generalizability of the Social-Emotional and Character Development Scale with a large sample of children from schools in Belize. Exploratory structural equation modeling results demonstrate the original factor correlations were reduced, providing less biased estimates than confirmatory factor analysis. Moreover, a multi-group Exploratory Structural Equation Modeling analysis illuminates significant differences between latent factor scores of males and females for most factors. Using this newer factor analytic procedure, original factors are reconceptualized to better situate the Social Emotional Character Development Scales into the larger body of Social-Emotional Learning competencies literature. (shrink)

The forms of words as they appear in text and speech are central to theories and models of lexical processing. Nonetheless, current methods for simulating their learning and representation fail to approach the scale and heterogeneity of real wordform lexicons. A connectionist architecture termed thesequence encoderis used to learn nearly 75,000 wordform representations through exposure to strings of stress‐marked phonemes or letters. First, the mechanisms and efficacy of the sequence encoder are demonstrated and shown to overcome problems with traditional (...) slot‐based codes. Then, two large‐scale simulations are reported that learned to represent lexicons of either phonological or orthographic wordforms. In doing so, the models learned the statistics of their lexicons as shown by better processing of well‐formed pseudowords as opposed to ill‐formed (scrambled) pseudowords, and by accounting for variance in well‐formedness ratings. It is discussed how the sequence encoder may be integrated into broader models of lexical processing. (shrink)

The forms of words as they appear in text and speech are central to theories and models of lexical processing. Nonetheless, current methods for simulating their learning and representation fail to approach the scale and heterogeneity of real wordform lexicons. A connectionist architecture termed thesequence encoderis used to learn nearly 75,000 wordform representations through exposure to strings of stress‐marked phonemes or letters. First, the mechanisms and efficacy of the sequence encoder are demonstrated and shown to overcome problems with traditional (...) slot‐based codes. Then, two large‐scale simulations are reported that learned to represent lexicons of either phonological or orthographic wordforms. In doing so, the models learned the statistics of their lexicons as shown by better processing of well‐formed pseudowords as opposed to ill‐formed (scrambled) pseudowords, and by accounting for variance in well‐formedness ratings. It is discussed how the sequence encoder may be integrated into broader models of lexical processing. (shrink)

Many researchers have documented how AI and data driven technologies have the potential to have profound effects on our lives—in ways that make these technologies stand out from those that went before. Around the world, we are seeing a significant growth in interest and investment in AI in healthcare. This has been coupled with rising concerns about the ethical implications of these technologies and an array of ethical guidelines for the use of AI and data in healthcare has arisen. Nevertheless, (...) the question of if and how AI and data technologies can be ethical remains open to debate. This paper aims to contribute to this debate by considering the wide range of implications that have been attributed to these technologies and asking whether current ethical guidelines take these factors into account. In particular, the paper argues that while current ethics guidelines for AI in healthcare effectively account for the four key issues identified in the ethics literature (transparency; fairness; responsibility and privacy), they have largely neglected wider issues relating to the way in which these technologies shape institutional and social arrangements. This, I argue, has given current ethics guidelines a strong focus on evaluating the impact of these technologies on the individual, while not accounting for the powerful social shaping effects of these technologies. To address this, the paper proposes a Multiscale Ethics Framework, which aims to help technology developers and ethical evaluations to consider the wider implications of these technologies. (shrink)

The purpose of this study was to examine the cross-cultural validity of the Organizational Stressor Indicator for Sport Performers scale by investigating its psychometric properties with Ghanaian footballers. The study particularly sought to assess in the Ghanaian context, 1, the convergence validity and reliability of the OSI-SP scale, 2, the discriminant validity of the OSI-SP scale to understand the applicability of its factor structure, and 3, whether the OSI-SP hypothesized model fits the data collected within the study (...) context. The intensity aspect of the OSI-SP questionnaire was administered to 424 Ghana Premier League male footballers who took part in the 2020/2021 season. Quality control strategies were put in place to ensure consistency across interpreters and as well improve the validity of the data. The results from a multi-factor first-order confirmatory factor analysis showed some level of convergence validity of the OSI-SP scale in the Ghanaian context using football players. Out of the 23 items on the original scale, 20 met the factor loadings criterion. In assessing the discriminant validity of the OSI-SP scale using Heterotrait-monotrait, 50% of the correlation ratios met the criterion for the original 23-item instrument. Comparing the new model with the original model using the Akaike Information Criterion value, the model fit indices for the modified model appeared better than the original model with 23-items. Generally, there was minimal support for the applicability of the OSI-SP instrument across the sample of Ghanaian footballers. The implications of these findings are discussed in detail. (shrink)

The present paper describes a joint action paradigm in which individuals or pairs utilized two computer keys to keep a dot stimulus moving inside a larger rectangle. Members of a pair could neither see nor hear each other. This paradigm allowed us to combine the discrete-trial type dependent variables commonly utilized by representational theorists, with the continuous, temporal dependence variables utilized by dynamical theorists. Analysis revealed that individuals kept the dot in the rectangle longer than dyads and did so by (...) moving it back and forth within the rectangle. Dyads, however, pressed their individual buttons as quickly as possible in order to keep the dot near the middle of the rectangle. These findings indicate that joint action constitutes a multi-scale phenomena that is best investigated via multiple, complementary methodologies versus single-measure, competing theories. (shrink)

I discuss arguments about the relationship between different “levels” of explanation in the light of examples involving multi-scale analysis. I focus on arguments about causal competition between properties at different levels, such as Jaegwon Kim’s “supervenience argument.” A central feature of Kim’s argument is that higher-level properties can in general be identified with “micro-based” properties. I argue that explanations from multi-scale analysis give examples of explanations that are problematic for accounts such as Kim’s. I argue that (...) these difficulties suggest that some standard assumptions about causal competition need to be revised. (shrink)

The paper begins with a generic discussion of modelling, focusing on some of its practices and problems. I then move on to a philosophical discussion about emergence and multi-scale modelling; more specifically, the reasons why what looks like a promising strategy for dealing with emergence is sometimes incapable of delivering interesting results. This becomes especially evident when we look more closely at turbulence and what I take to be the main ontological feature of emergent behavior—universality. Finally, I conclude (...) by showing why, despite displaying multi-scale behaviour and some of the characteristics we identify with emergence, turbulence fails to fit neatly into the latter category and is not successfully captured using multi-scale modelling. The complex nature of turbulence illustrates the difficulties in characterizing emergence and why specific criteria are needed in order to prevent every complex behaviour we don’t understand being classified as emergent. (shrink)

The purpose of this cross-sectional investigation was to examine the item characteristics, factor structure, reliability, convergent validity, and construct validity of the Psychobiosocial States in Physical Education (PBS-SPE) scale. In Study 1, a sample of 1,030 students (582 girls and 448 boys, 10 to 19-year-olds), drawn from middle or high schools, rated the intensity of the 20 items version of the PBS-SPE scale thinking about the feelings they had usually experienced in physical education classes. In Study 2, an (...) additional sample of 1,025 students (578 girls, 447 boys, 10 to 19-year-olds), rated the 16 items of the final version of the scale. Two subsamples also completed an affective-related measure (i.e., the Physical Activity Enjoyment Scale) and two motivation scales often used in the physical education domain (i.e., the Teacher-Initiated Motivational Climate in Physical Education Questionnaire and the Situational Motivation Scale). Exploratory structural equation modeling and confirmatory factor analyses of the data showed that a two-factor, 16-item solution (i.e., 8 pleasant/functional items and 8 unpleasant/dysfunctional items) of the PBS-SPE scale reached satisfactory fit indices. Multi-group comparisons provided support for measurement and structural invariance across samples, gender, and age. Convergent and construct validity was also upheld. Overall, the findings offer support for the use of a new instrument in the assessment of psychobiosocial states in physical education settings. (shrink)

There are a lot of research studies that look at "fake news" from an Arabic online source, but they don't look at what makes those fake news spread. The threat grows, and at some point, it gets out of hand. That's why this paper is trying to figure out how to predict the features that make Arabic online fake news spread. It's using Naive Bayes, Logistic Regression, and Random forest of Machine Learning to do this. Online news stories that were (...) made up were used. They are found by using Term Frequency-Inverse Document Frequency. The best partition for testing and validating the prediction was chosen at random and used in the analysis. So, all three machine learning classifications for predicting fake news in Arabic online were done. The results of the experiment show that Random Forest Classifier outperformed the other two algorithms. It had the best TF-IDF with an accuracy of 86 percent. Naive Bayes had an accuracy rate of 84%, and Logistic Regression had an accuracy rate of 85%, so they all did well. As such, the model shows that the features in TF-IDF are the most essential point about the content of an online Arabic fake news. (shrink)

The purpose of crowd counting is to estimate the number of pedestrians in crowd images. Crowd counting or density estimation is an extremely challenging task in computer vision, due to large scale variations and dense scene. Current methods solve these issues by compounding multi-scale Convolutional Neural Network with different receptive fields. In this paper, a novel end-to-end architecture based on Multi-Scale Adversarial Convolutional Neural Network (MSA-CNN) is proposed to generate crowd density and estimate the amount (...) of crowd. Firstly, a multi-scale network is used to extract the globally relevant features in the crowd image, and then fractionally-strided convolutional layers are designed for up-sampling the output to recover the loss of crucial details caused by the earlier max pooling layers. An adversarial loss is directly employed to shrink the estimated value into the realistic subspace to reduce the blurring effect of density estimation. Joint training is performed in an end-to-end fashion using a combination of Adversarial loss and Euclidean loss. The two losses are integrated via a joint training scheme to improve density estimation performance.We conduct some extensive experiments on available datasets to show the significant improvements and supremacy of the proposed approach over the available state-of-the-art approaches. (shrink)

This chapter takes inspiration from Wittgenstein’s thinking to formulate a non-reductive toolbox for the study of religion associated with generative modelling, specifically as applied in complex adaptive systems theory. It converges on a communal perspective on religion as multiscale active inference that contrasts starkly with common ‘straw person’ perspectives on religion that reduce it to ‘erroneous’ theorising generated by the brain. In contrast, we argue, religious practices at the enculturated level of description involve implicit and explicit meanings, experienced both individually (...) and collectively. Firstly, we characterise these in terms of Wittgenstein’s notions of rule-following and language games. Secondly, we discuss how the collective and perspectival aspects of religion in enculturation with morals, doctrines, rituals, and expressions can be formalised in terms of deep active inference in multi-agent systems. This approach describes how religions are incommensurable with scientific methods due to the epistemic separation between the different kinds of language games of religious and scientific practices. Their touchpoints tend to give rise to perpetual confusion and unproductive discussions, requiring perspicuous presentations that can be formulated with our toolbox. In this sense, our account has implications for tackling complex societal challenges such as ethical policy-making. (shrink)

The debate between ΛCDM and MOND is often cast in terms of competing gravitational theories. However, recent philosophical discussion suggests that the ΛCDM–MOND debate demonstrates the challenges of multiscale modeling in the context of cosmological scales. I extend this discussion and explore what happens when the debate is thought to be about modeling rather than about theory, offering a model-focused interpretation of the ΛCDM–MOND debate. This analysis shows how a model-focused interpretation of the debate provides a better understanding (...) of challenges asso- ciated with extension to a different scale or domain, which are tied to commitments about explanatory fit. (shrink)

Umwelt theory is an expression of von Uexküll’s subjective biology and as such usually applied in analysis of individual animals, yet it is fundamentally relational and therefore also suitable for analysis of more complex wholes. Since the birth of the modern environmental movement in the 1960s, there has been growing scientific and political acknowledgement of there being a global environmental crisis, which today manifests itself as a climate change and biodiversity crisis. This calls for a multi-scale ecosemiotic approach (...) to analysis of human ecology at various levels and scales. In this article I explore to what extent ecosemiotic methodology, drawing on Umwelt theory and its consistently subjective perspective, can be applied in analysis of human ecology at different geographical and ecological scales ranging from the global to the local. The article incorporates a case study of human–animal relations in Mamirauá Sustainable Development Reserve in the Central Amazon. This is a seasonal floodplain forest area surrounded by rivers. I investigate aspects of the living conditions and ecology of the reserve, with a main focus on indigenous communities and the circumstances of two primate species, namely the red howler monkey and the black-headed squirrel monkey. I outline matrixes of levels of study in ecosemiotics, and scales in human ecology, and apply two scales to the Mamirauá case. These take an individual animal’s and an individual human being’s subjective experience as their respectively starting points. This allows for multi-scale studies of human ecology from complementary angles. (shrink)

Convolutional neural networks, which automatically learn features from raw data to approximate functions, are being increasingly applied to the end-to-end analysis of electroencephalographic signals, especially for decoding brain states in brain-computer interfaces. Nevertheless, CNNs introduce a large number of trainable parameters, may require long training times, and lack in interpretability of learned features. The aim of this study is to propose a CNN design for P300 decoding with emphasis on its lightweight design while guaranteeing high performance, on the effects of (...) different training strategies, and on the use of post-hoc techniques to explain network decisions. The proposed design, named MS-EEGNet, learned temporal features in two different timescales in an efficient and optimized way, and was validated on three P300 datasets. The CNN was trained using different strategies and was compared with several state-of-the-art algorithms. Furthermore, variants of the baseline MS-EEGNet were analyzed to evaluate the impact of different hyper-parameters on performance. Lastly, saliency maps were used to derive representations of the relevant spatio-temporal features that drove CNN decisions. MS-EEGNet was the lightest CNN compared with the tested SOA CNNs, despite its multiple timescales, and significantly outperformed the SOA algorithms. Post-hoc hyper-parameter analysis confirmed the benefits of the innovative aspects of MS-EEGNet. Furthermore, MS-EEGNet did benefit from transfer learning, especially using a low number of training examples, suggesting that the proposed approach could be used in BCIs to accurately decode the P300 event while reducing calibration times. Representations derived from the saliency maps matched the P300 spatio-temporal distribution, further validating the proposed decoding approach. This study, by specifically addressing the aspects of lightweight design, transfer learning, and interpretability, can contribute to advance the development of deep learning algorithms for P300-based BCIs. (shrink)

Objective: Tangent Space Mapping using the geometric structure of the covariance matrices is an effective method to recognize multiclass motor imagery. Compared with the traditional CSP method, the Riemann geometric method based on TSM takes into account the nonlinear information contained in the covariance matrix, and can extract more abundant and effective features. Moreover, the method is an unsupervised operation, which can reduce the time of feature extraction. However, EEG features induced by MI mental activities of different subjects are not (...) the same, so selection of subject-specific discriminative EEG frequency components play a vital role in the recognition of multiclass MI. In order to solve the problem, a discriminative and multi-scale filter bank tangent space mapping algorithm is proposed in this article to design the subject-specific Filter Bank so as to effectively recognize multiclass MI tasks. Methods: On the 4-class BCI competition IV-2a dataset, first, a non-parametric method of multivariate analysis of variance based on the sum of squared distances is used to select discriminative frequency bands for a subject; next, a multi-scale FB is generated according to the range of these frequency bands, and then decompose multi-channel EEG of the subject into multiple sub-bands combined with several time windows. Then TSM algorithm is used to estimate Riemannian tangent space features in each sub-band and finally a liner Support Vector Machines is used for classification. Main Results: The analysis results show that the proposed discriminative FB enhances the multi-scale TSM algorithm, improves the classification accuracy and reduces the execution time during training and testing. On the 4-class BCI competition IV-2a dataset, the average session to session classification accuracy of nine subjects reached 77.33 ± 12.3%. When the training time and the test time are similar, the average classification accuracy is 2.56% higher than the latest TSM method based on multi-scale filter bank analysis technology. When the classification accuracy is similar, the training speed is increased by more than three times, and the test speed is increased two times more. Compared with Supervised Fisher Geodesic Minimum Distance to the Mean, another new variant based on Riemann geometry classifier, the average accuracy is 3.36% higher, we also compared with the latest Deep Learning method, and the average accuracy of 10-fold cross validation improved by 2.58%. Conclusion: Research shows that the proposed DMFBTSM algorithm can improve the classification accuracy of MI tasks. Significance: Compared with the MFBTSM algorithm, the algorithm proposed in this article is expected to select frequency bands with good separability for specific subject to improve the classification accuracy of multiclass MI tasks and reduce the feature dimension to reduce training time and testing time. (shrink)

Top-down causation is often taken to be a metaphysically suspicious type of causation that is found in a few complex systems, such as in human mind-body relations. However, as Ellis and others have shown, top-down causation is ubiquitous in physics as well as in biology. Top-down causation occurs whenever specific dynamic behaviors are realized or selected among a broader set of possible lower-level states. Thus understood, the occurrence of dynamic and structural patterns in physical and biological systems presents a problem (...) for reductionist positions. We illustrate with examples of universality and functional equivalence classes how higher-level behaviors can be multiple realized by distinct lower-level systems or states. Multiple realizability in both contexts entails what Ellis calls “causal slack” between levels, or what others understand as relative explanatory autonomy. To clarify these notions further, we examine procedures for upscaling in multi-scale modeling. We argue that simple averaging strategies for upscaling only work for simplistic homogenous systems, because of the scale-dependency of characteristic behaviors in multi-scale systems. We suggest that this interpretation has implications for what Ellis calls mechanical top-down causation, as it presents a stronger challenge to reductionism than typically assumed. (shrink)

Standard semantic information processing models—information in; information processed; information out —lend themselves to standard models of the functioning of the brain in terms, e.g., of threshold-switch neurons connected via classical synapses. That is, in terms of sophisticated descendants of McCulloch and Pitts models. I argue that both the cognition and the brain sides of this framework are incorrect: cognition and thought are not constituted as forms of semantic information processing, and the brain does not function in terms of passive input (...) processing units organized as neural nets. An alternative framework is developed that models cognition and thought not in terms of semantic information processing, and, correspondingly, models brain functional processes also not in terms of semantic information processing. As alternative to such models: I outline a pragmatist oriented, interaction based, model of representation; derive from this model a fundamental framework of constraints on how the brain must function; show that such a framework is in fact found in the brain, and develop the outlines of a broader model of how mental processes can be realized within this alternative framework. Part I of this discussion focuses on some criticisms of standard modeling frameworks for representation and cognition, and outlines an alternative interactivist, pragmatist oriented, model. In part II, the focus is on the fact that the brain does not, in fact, function in accordance with standard passive input processing models—e.g., information processing models. Instead, there are multiple endogenously active processes at multiple spatial and temporal scales across multiple kinds of cells. A micro-functional model that accounts for, and even predicts, these multi-scale phenomena in generating emergent representation and cognition is outlined. That is, I argue that the interactivist model of representation outlined offers constraints on how the brain should function that are in fact empirically found, and, in reverse, that the multifarious details of brain functioning entail the pragmatist representational model—a very strong interrelationship. In the sequel paper, starting with part III, this model is extended to address macro-functioning in the CNS. In part IV, I offer a discussion of an approach to brain functioning that has some similarities with, as well as differences from, the model presented here: sometimes called the predictive brain approach. (shrink)

This paper explores some issues concerning how we should think about interventions (in the sense of unconfounded manipulations) of "upper-level" variables in contexts in which these supervene on but are not identical with lower-level realizers. It is argued that we should reject the demand that interventions on upper-level variables must leave their lower-level realizers unchanged– a requirement that within an interventionist framework would imply that upper-level variables are causally inert. Instead an intervention on an upper-level variable at the same time (...) changes its lower-level realizer in a way that is consistent with the change in the upper- level variable. The lower-level realizer should not be regarded as a potential "confounder" of upper-level causal relations, as in "exclusionist" arguments. Several proposals for making this precise are considered and several pro-exclusionist arguments are criticized. (shrink)