Predictive coding is a consequence of energy efficiency in recurrent neural networks

Induced zero-phase synchronization as a potential neural code for optimized visuomotor integration

BACKGROUND

Goal-directed behavior requires the integration of information from the outside world and internal (somatosensory) sources about our own actions. Expectations (or 'internal models') are generated from prior knowledge and constantly updated based on sensory feedback. This optimized information integration ('predictive coding') results in a global behavioral advantage of anticipated action in the presence of uncertainty. Our goal was to probe the effect of phase entrainment of the sensorimotor mu-rhythm on visuomotor integration.

METHODS

Participants received transcranial alternating current stimulation over bilateral motor cortices (M1) while performing a visually-guided force adjustment task during functional magnetic resonance imaging.

RESULTS

Inter-hemispheric zero-phase entrainment resulted in effector-specific modulation of performance precision and effector-generic minimization of force signal complexity paralleled by BOLD activation changes in bilateral caudate and increased functional connectivity between the right M1 and contralateral putamen, inferior parietal, and medial temporal regions. While effector-specific changes in performance precision were associated with contralateral caudate and hippocampal activation decreases, only the global reduction in force signal complexity was associated with increased functional M1 connectivity with bilateral striatal regions.

CONCLUSION

We propose that zero-phase synchronization represents a neural mode of optimized information integration related to internal model updating within the recursive perception-action continuum associated with predictive coding.

Finding order in chaos: influences of environmental complexity and predictability on development

Environments are dynamic and complex. Some children experience more predictable early life environments than others. Here, we consider how moment-by-moment complexity and predictability in our early environments influence development. New studies using wearable sensors are quantifying this environmental variability at a fine temporal resolution across hierarchically structured physical and social features. We identify three types of predictability: periodicities ('at X time intervals, Y happens'), stability ('given statex, statex+1 is known'), and contingency ('when I do X, Y happens'). We discuss how the temporal dynamics of environments may differ between individuals and the diverse developmental neural pathways through which this may influence outcomes, such as central nervous system (CNS) arousal and executive control. Finally, we discuss practical consequences and directions for future research.

Coordinating multiple mental faculties during learning

Complex behavior is supported by the coordination of multiple brain regions. How do brain regions coordinate absent a homunculus? We propose coordination is achieved by a controller-peripheral architecture in which peripherals (e.g., the ventral visual stream) aim to supply needed inputs to their controllers (e.g., the hippocampus and prefrontal cortex) while expending minimal resources. We developed a formal model within this framework to address how multiple brain regions coordinate to support rapid learning from a few example images. The model captured how higher-level activity in the controller shaped lower-level visual representations, affecting their precision and sparsity in a manner that paralleled brain measures. In particular, the peripheral encoded visual information to the extent needed to support the smooth operation of the controller. Alternative models optimized by gradient descent irrespective of architectural constraints could not account for human behavior or brain responses, and, typical of standard deep learning approaches, were unstable trial-by-trial learners. While previous work offered accounts of specific faculties, such as perception, attention, and learning, the controller-peripheral approach is a step toward addressing next generation questions concerning how multiple faculties coordinate.

Dynamics of specialization in neural modules under resource constraints

The brain is structurally and functionally modular, although recent evidence has raised questions about the extent of both types of modularity. Using a simple, toy artificial neural network setup that allows for precise control, we find that structural modularity does not in general guarantee functional specialization (across multiple measures of specialization). Further, in this setup (1) specialization only emerges when features of the environment are meaningfully separable, (2) specialization preferentially emerges when the network is strongly resource-constrained, and (3) these findings are qualitatively similar across several different variations of network architectures. Finally, we show that functional specialization varies dynamically across time, and these dynamics depend on both the timing and bandwidth of information flow in the network. We conclude that a static notion of specialization is likely too simple a framework for understanding intelligence in situations of real-world complexity, from biology to brain-inspired neuromorphic systems.

Model-Based Approaches to Investigating Mismatch Responses in Schizophrenia

Alterations of mismatch responses (ie, neural activity evoked by unexpected stimuli) are often considered a potential biomarker of schizophrenia. Going beyond establishing the type of observed alterations found in diagnosed patients and related cohorts, computational methods can yield valuable insights into the underlying disruptions of neural mechanisms and cognitive function. Here, we adopt a typology of model-based approaches from computational cognitive neuroscience, providing an overview of the study of mismatch responses and their alterations in schizophrenia from four complementary perspectives: (a) connectivity models, (b) decoding models, (c) neural network models, and (d) cognitive models. Connectivity models aim at inferring the effective connectivity patterns between brain regions that may underlie mismatch responses measured at the sensor level. Decoding models use multivariate spatiotemporal mismatch response patterns to infer the type of sensory violations or to classify participants based on their diagnosis. Neural network models such as deep convolutional neural networks can be used for improved classification performance as well as for a systematic study of various aspects of empirical data. Finally, cognitive models quantify mismatch responses in terms of signaling and updating perceptual predictions over time. In addition to describing the available methodology and reviewing the results of recent computational psychiatry studies, we offer suggestions for future work applying model-based techniques to advance the study of mismatch responses in schizophrenia.

Examining predictive coding accounts of typical and autistic neurocognitive development

Predictive coding has emerged as a prominent theoretical framework for understanding perception and its neural underpinnings. There has been a recent surge of interest in the predictive coding framework across the mind sciences. However, comparatively little of the research in this field has investigated the neural underpinnings of predictive coding in young neurotypical and autistic children. This paper provides an overview of predictive coding accounts of typical and autistic neurocognitive development and includes a review of the current electrophysiological evidence supporting these accounts. Based on the current evidence, it is clear that more research in pediatrics is needed to evaluate predictive coding accounts of neurocognitive development fully. If supported, these accounts could have wide-ranging practical implications for pedagogy, parenting, artificial intelligence, and clinical approaches to helping autistic children manage the barrage of everyday sensory information.

Predictive learning shapes the representational geometry of the human brain

Predictive coding theories propose that the brain constantly updates internal models to minimize prediction errors and optimize sensory processing. However, the neural mechanisms that link prediction error encoding and optimization of sensory representations remain unclear. Here, we provide evidence how predictive learning shapes the representational geometry of the human brain. We recorded magnetoencephalography (MEG) in humans listening to acoustic sequences with different levels of regularity. We found that the brain aligns its representational geometry to match the statistical structure of the sensory inputs, by clustering temporally contiguous and predictable stimuli. Crucially, the magnitude of this representational shift correlates with the synergistic encoding of prediction errors in a network of high-level and sensory areas. Our findings suggest that, in response to the statistical regularities of the environment, large-scale neural interactions engaged in predictive processing modulate the representational content of sensory areas to enhance sensory processing.

Predicting subjective ratings of affect and comprehensibility with text features: a reader response study of narrative poetry

Literary reading is an interactive process between a reader and a text that depends on a balance between cognitive effort and emotional rewards. By studying both the crucial features of the text and of the subjective reader reception, a better understanding of this interactive process can be reached. In the present study, subjects (N=31) read and rated a work of narrative fiction that was written in a poetic style, thereby offering the readers two pathways to cognitive rewards: Aesthetic appreciation and narrative immersion. Using purely text-based quantitative descriptors, we were able to independently and accurately predict the subjective ratings in the dimensions comprehensibility, valence, arousal, and liking across roughly 140 pages of naturalistic text. The specific text features that were most important in predicting each rating dimension are discussed in detail. In addition, the implications of the findings are discussed more generally in the context of existing models of literary processing and future research avenues for empirical literary studies.

Shaping dynamical neural computations using spatiotemporal constraints

Dynamics play a critical role in computation. The principled evolution of states over time enables both biological and artificial networks to represent and integrate information to make decisions. In the past few decades, significant multidisciplinary progress has been made in bridging the gap between how we understand biological versus artificial computation, including how insights gained from one can translate to the other. Research has revealed that neurobiology is a key determinant of brain network architecture, which gives rise to spatiotemporally constrained patterns of activity that underlie computation. Here, we discuss how neural systems use dynamics for computation, and claim that the biological constraints that shape brain networks may be leveraged to improve the implementation of artificial neural networks. To formalize this discussion, we consider a natural artificial analog of the brain that has been used extensively to model neural computation: the recurrent neural network (RNN). In both the brain and the RNN, we emphasize the common computational substrate atop which dynamics occur-the connectivity between neurons-and we explore the unique computational advantages offered by biophysical constraints such as resource efficiency, spatial embedding, and neurodevelopment.

Collective behavior from surprise minimization

Collective motion is ubiquitous in nature; groups of animals, such as fish, birds, and ungulates appear to move as a whole, exhibiting a rich behavioral repertoire that ranges from directed movement to milling to disordered swarming. Typically, such macroscopic patterns arise from decentralized, local interactions among constituent components (e.g., individual fish in a school). Preeminent models of this process describe individuals as self-propelled particles, subject to self-generated motion and "social forces" such as short-range repulsion and long-range attraction or alignment. However, organisms are not particles; they are probabilistic decision-makers. Here, we introduce an approach to modeling collective behavior based on active inference. This cognitive framework casts behavior as the consequence of a single imperative: to minimize surprise. We demonstrate that many empirically observed collective phenomena, including cohesion, milling, and directed motion, emerge naturally when considering behavior as driven by active Bayesian inference-without explicitly building behavioral rules or goals into individual agents. Furthermore, we show that active inference can recover and generalize the classical notion of social forces as agents attempt to suppress prediction errors that conflict with their expectations. By exploring the parameter space of the belief-based model, we reveal nontrivial relationships between the individual beliefs and group properties like polarization and the tendency to visit different collective states. We also explore how individual beliefs about uncertainty determine collective decision-making accuracy. Finally, we show how agents can update their generative model over time, resulting in groups that are collectively more sensitive to external fluctuations and encode information more robustly.

The cognitive neuroscience of self-awareness: Current framework, clinical implications, and future research directions

Self-awareness, the ability to take oneself as the object of awareness, has been an enigma for our species, with different answers to this question being provided by religion, philosophy, and, more recently, science. The current review aims to discuss the neurocognitive mechanisms underlying self-awareness. The multidimensional nature of self-awareness will be explored, suggesting how it can be thought of as an emergent property observed in different cognitive complexity levels, within a predictive coding approach. A presentation of alterations of self-awareness in neuropsychiatric conditions will ground a discussion on alternative frameworks to understand this phenomenon, in health and psychopathology, with future research directions being indicated to fill current gaps in the literature. This article is categorized under: Philosophy > Consciousness Psychology > Brain Function and Dysfunction Neuroscience > Cognition.

An active inference perspective for the amygdala complex

The amygdala is a heterogeneous network of subcortical nuclei with central importance in cognitive and clinical neuroscience. Various experimental designs in human psychology and animal model research have mapped multiple conceptual frameworks (e.g., valence/salience and decision making) to ever more refined amygdala circuitry. However, these predominantly bottom up-driven accounts often rely on interpretations tailored to a specific phenomenon, thus preventing comprehensive and integrative theories. We argue here that an active inference model of amygdala function could unify these fractionated approaches into an overarching framework for clearer empirical predictions and mechanistic interpretations. This framework embeds top-down predictive models, informed by prior knowledge and belief updating, within a dynamical system distributed across amygdala circuits in which self-regulation is implemented by continuously tracking environmental and homeostatic demands.

Active Predictive Coding: A Unifying Neural Model for Active Perception, Compositional Learning, and Hierarchical Planning

There is growing interest in predictive coding as a model of how the brain learns through predictions and prediction errors. Predictive coding models have traditionally focused on sensory coding and perception. Here we introduce active predictive coding (APC) as a unifying model for perception, action, and cognition. The APC model addresses important open problems in cognitive science and AI, including (1) how we learn compositional representations (e.g., part-whole hierarchies for equivariant vision) and (2) how we solve large-scale planning problems, which are hard for traditional reinforcement learning, by composing complex state dynamics and abstract actions from simpler dynamics and primitive actions. By using hypernetworks, self-supervised learning, and reinforcement learning, APC learns hierarchical world models by combining task-invariant state transition networks and task-dependent policy networks at multiple abstraction levels. We illustrate the applicability of the APC model to active visual perception and hierarchical planning. Our results represent, to our knowledge, the first proof-of-concept demonstration of a unified approach to addressing the part-whole learning problem in vision, the nested reference frames learning problem in cognition, and the integrated state-action hierarchy learning problem in reinforcement learning.

Going after the bigger picture: Using high-capacity models to understand mind and brain

Deep neural networks (DNNs) provide a unique opportunity to move towards a generic modelling framework in psychology. The high representational capacity of these models combined with the possibility for further extensions has already allowed us to investigate the forest, namely the complex landscape of representations and processes that underlie human cognition, without forgetting about the trees, which include individual psychological phenomena.

Shaping dynamical neural computations using spatiotemporal constraints

Dynamics play a critical role in computation. The principled evolution of states over time enables both biological and artificial networks to represent and integrate information to make decisions. In the past few decades, significant multidisciplinary progress has been made in bridging the gap between how we understand biological versus artificial computation, including how insights gained from one can translate to the other. Research has revealed that neurobiology is a key determinant of brain network architecture, which gives rise to spatiotemporally constrained patterns of activity that underlie computation. Here, we discuss how neural systems use dynamics for computation, and claim that the biological constraints that shape brain networks may be leveraged to improve the implementation of artificial neural networks. To formalize this discussion, we consider a natural artificial analog of the brain that has been used extensively to model neural computation: the recurrent neural network (RNN). In both the brain and the RNN, we emphasize the common computational substrate atop which dynamics occur-the connectivity between neurons-and we explore the unique computational advantages offered by biophysical constraints such as resource efficiency, spatial embedding, and neurodevelopment.

Prediction during language comprehension: what is next?

Prediction is often regarded as an integral aspect of incremental language comprehension, but little is known about the cognitive architectures and mechanisms that support it. We review studies showing that listeners and readers use all manner of contextual information to generate multifaceted predictions about upcoming input. The nature of these predictions may vary between individuals owing to differences in language experience, among other factors. We then turn to unresolved questions which may guide the search for the underlying mechanisms. (i) Is prediction essential to language processing or an optional strategy? (ii) Are predictions generated from within the language system or by domain-general processes? (iii) What is the relationship between prediction and memory? (iv) Does prediction in comprehension require simulation via the production system? We discuss promising directions for making progress in answering these questions and for developing a mechanistic understanding of prediction in language.

The interpersonal computational psychiatry of social coordination in schizophrenia

Impairments in social coordination form a core dimension of various psychiatric disorders, including schizophrenia. Advances in interpersonal and computational psychiatry support a major change in studying social coordination in schizophrenia. Although these developments provided novel perspectives to study how interpersonal activities shape coordination and to examine computational mechanisms, direct attempts to integrate the two methodologies have been sparse. Here, we propose an interpersonal computational framework that (1) leverages the active inference framework to model aberrant social coordination processes in schizophrenia and (2) incorporates dynamical system models to dissect intrapersonal and interpersonal synchronisation to inform a statistical model based on active inference. We discuss how this interpersonal computational psychiatry framework can elucidate the aberrant processes leading to psychopathology, with schizophrenia as an example, and highlight how it might aid clinical intervention and practice. Finally, we discuss challenges and opportunities for using the framework in studying social coordination impairments.

The neuroconnectionist research programme

Artificial neural networks (ANNs) inspired by biology are beginning to be widely used to model behavioural and neural data, an approach we call 'neuroconnectionism'. ANNs have been not only lauded as the current best models of information processing in the brain but also criticized for failing to account for basic cognitive functions. In this Perspective article, we propose that arguing about the successes and failures of a restricted set of current ANNs is the wrong approach to assess the promise of neuroconnectionism for brain science. Instead, we take inspiration from the philosophy of science, and in particular from Lakatos, who showed that the core of a scientific research programme is often not directly falsifiable but should be assessed by its capacity to generate novel insights. Following this view, we present neuroconnectionism as a general research programme centred around ANNs as a computational language for expressing falsifiable theories about brain computation. We describe the core of the programme, the underlying computational framework and its tools for testing specific neuroscientific hypotheses and deriving novel understanding. Taking a longitudinal view, we review past and present neuroconnectionist projects and their responses to challenges and argue that the research programme is highly progressive, generating new and otherwise unreachable insights into the workings of the brain.

Inconsistent illusory motion in predictive coding deep neural networks

Why do we perceive illusory motion in some static images? Several accounts point to eye movements, response latencies to different image elements, or interactions between image patterns and motion energy detectors. Recently PredNet, a recurrent deep neural network (DNN) based on predictive coding principles, was reported to reproduce the "Rotating Snakes" illusion, suggesting a role for predictive coding. We begin by replicating this finding, then use a series of "in silico" psychophysics and electrophysiology experiments to examine whether PredNet behaves consistently with human observers and non-human primate neural data. A pretrained PredNet predicted illusory motion for all subcomponents of the Rotating Snakes pattern, consistent with human observers. However, we found no simple response delays in internal units, unlike evidence from electrophysiological data. PredNet's detection of motion in gradients seemed dependent on contrast, but depends predominantly on luminance in humans. Finally, we examined the robustness of the illusion across ten PredNets of identical architecture, retrained on the same video data. There was large variation across network instances in whether they reproduced the Rotating Snakes illusion, and what motion, if any, they predicted for simplified variants. Unlike human observers, no network predicted motion for greyscale variants of the Rotating Snakes pattern. Our results sound a cautionary note: even when a DNN successfully reproduces some idiosyncrasy of human vision, more detailed investigation can reveal inconsistencies between humans and the network, and between different instances of the same network. These inconsistencies suggest that predictive coding does not reliably give rise to human-like illusory motion.

Energy efficiency as a normative account for predictive coding

In this issue of Patterns, Ali et al. demonstrate that predictive coding emerges in an artificial neural network optimized to be energy efficient. The results offer an explanation for why brains may implement predictive coding.