A theory of cortical responses

This article concerns the nature of evoked brain responses and the principles underlying their generation. We start with the premise that the sensory brain has evolved to represent or infer the causes of changes in its sensory inputs. The problem of inference is well formulated in statistical terms. The statistical fundaments of inference may therefore afford important constraints on neuronal implementation. By formulating the original ideas of Helmholtz on perception, in terms of modern-day statistical theories, one arrives at a model of perceptual inference and learning that can explain a remarkable range of neurobiological facts.It turns out that the problems of inferring the causes of sensory input (perceptual inference) and learning the relationship between input and cause (perceptual learning) can be resolved using exactly the same principle. Specifically, both inference and learning rest on minimizing the brain's free energy, as defined in statistical physics. Furthermore, inference and learning can proceed in a biologically plausible fashion. Cortical responses can be seen as the brain's attempt to minimize the free energy induced by a stimulus and thereby encode the most likely cause of that stimulus. Similarly, learning emerges from changes in synaptic efficacy that minimize the free energy, averaged over all stimuli encountered. The underlying scheme rests on empirical Bayes and hierarchical models of how sensory input is caused. The use of hierarchical models enables the brain to construct prior expectations in a dynamic and context-sensitive fashion. This scheme provides a principled way to understand many aspects of cortical organization and responses. The aim of this article is to encompass many apparently unrelated anatomical, physiological and psychophysical attributes of the brain within a single theoretical perspective.In terms of cortical architectures, the theoretical treatment predicts that sensory cortex should be arranged hierarchically, that connections should be reciprocal and that forward and backward connections should show a functional asymmetry (forward connections are driving, whereas backward connections are both driving and modulatory). In terms of synaptic physiology, it predicts associative plasticity and, for dynamic models, spike-timing-dependent plasticity. In terms of electrophysiology, it accounts for classical and extra classical receptive field effects and long-latency or endogenous components of evoked cortical responses. It predicts the attenuation of responses encoding prediction error with perceptual learning and explains many phenomena such as repetition suppression, mismatch negativity (MMN) and the P300 in electroencephalography. In psychophysical terms, it accounts for the behavioural correlates of these physiological phenomena, for example, priming and global precedence. The final focus of this article is on perceptual learning as measured with the MMN and the implications for empirical studies of coupling among cortical areas using evoked sensory responses.     

Neural processing as causal inference

Perception is about making sense, that is, understanding what events in the outside world caused the sensory observations. Consistent with this intuition, many aspects of human behavior confronting noise and ambiguity are well explained by principles of causal inference. Extending these insights, recent studies have applied the same powerful set of tools to perceptual processing at the neural level. According to these approaches, microscopic neural structures solve elementary probabilistic tasks and can be combined to construct hierarchical predictive models of the sensory input. This framework suggests that variability in neural responses reflects the inherent uncertainty associated with sensory interpretations and that sensory neurons are active predictors rather than passive filters of their inputs. Causal inference can account parsimoniously and quantitatively for non-linear dynamical properties in single synapses, single neurons and sensory receptive fields.     

Action and behavior: a free-energy formulation

We have previously tried to explain perceptual inference and learning under a free-energy principle that pursues Helmholtz’s agenda to understand the brain in terms of energy minimization. It is fairly easy to show that making inferences about the causes of sensory data can be cast as the minimization of a free-energy bound on the likelihood of sensory inputs, given an internal model of how they were caused. In this article, we consider what would happen if the data themselves were sampled to minimize this bound. It transpires that the ensuing active sampling or inference is mandated by ergodic arguments based on the very existence of adaptive agents. Furthermore, it accounts for many aspects of motor behavior; from retinal stabilization to goal-seeking. In particular, it suggests that motor control can be understood as fulfilling prior expectations about proprioceptive sensations. This formulation can explain why adaptive behavior emerges in biological agents and suggests a simple alternative to optimal control theory. We illustrate these points using simulations of oculomotor control and then apply to same principles to cued and goal-directed movements. In short, the free-energy formulation may provide an alternative perspective on the motor control that places it in an intimate relationship with perception.     

Top-Down Feedback in an HMAX-Like Cortical Model of Object Perception Based on Hierarchical Bayesian Networks and Belief Propagation

Hierarchical generative models, such as Bayesian networks, and belief propagation have been shown to provide a theoretical framework that can account for perceptual processes, including feedforward recognition and feedback modulation. The framework explains both psychophysical and physiological experimental data and maps well onto the hierarchical distributed cortical anatomy. However, the complexity required to model cortical processes makes inference, even using approximate methods, very computationally expensive. Thus, existing object perception models based on this approach are typically limited to tree-structured networks with no loops, use small toy examples or fail to account for certain perceptual aspects such as invariance to transformations or feedback reconstruction. In this study we develop a Bayesian network with an architecture similar to that of HMAX, a biologically-inspired hierarchical model of object recognition, and use loopy belief propagation to approximate the model operations (selectivity and invariance). Crucially, the resulting Bayesian network extends the functionality of HMAX by including top-down recursive feedback. Thus, the proposed model not only achieves successful feedforward recognition invariant to noise, occlusions, and changes in position and size, but is also able to reproduce modulatory effects such as illusory contour completion and attention. Our novel and rigorous methodology covers key aspects such as learning using a layerwise greedy algorithm, combining feedback information from multiple parents and reducing the number of operations required. Overall, this work extends an established model of object recognition to include high-level feedback modulation, based on state-of-the-art probabilistic approaches. The methodology employed, consistent with evidence from the visual cortex, can be potentially generalized to build models of hierarchical perceptual organization that include top-down and bottom-up interactions, for example, in other sensory modalities.     

Predictive Coding or Evidence Accumulation? False Inference and Neuronal Fluctuations

Perceptual decisions can be made when sensory input affords an inference about what generated that input. Here, we report findings from two independent perceptual experiments conducted during functional magnetic resonance imaging (fMRI) with a sparse event-related design. The first experiment, in the visual modality, involved forced-choice discrimination of coherence in random dot kinematograms that contained either subliminal or periliminal motion coherence. The second experiment, in the auditory domain, involved free response detection of (non-semantic) near-threshold acoustic stimuli. We analysed fluctuations in ongoing neural activity, as indexed by fMRI, and found that neuronal activity in sensory areas (extrastriate visual and early auditory cortex) biases perceptual decisions towards correct inference and not towards a specific percept. Hits (detection of near-threshold stimuli) were preceded by significantly higher activity than both misses of identical stimuli or false alarms, in which percepts arise in the absence of appropriate sensory input. In accord with predictive coding models and the free-energy principle, this observation suggests that cortical activity in sensory brain areas reflects the precision of prediction errors and not just the sensory evidence or prediction errors per se.     

Predicting human perceptual decisions by decoding neuronal information profiles

Perception relies on the response of populations of neurons in sensory cortex. How the response profile of a neuronal population gives rise to perception and perceptual discrimination has been conceptualized in various ways. Here we suggest that neuronal population responses represent information about our environment explicitly as Fisher information (FI), which is a local measure of the variance estimate of the sensory input. We show how this sensory information can be read out and combined to infer from the available information profile which stimulus value is perceived during a fine discrimination task. In particular, we propose that the perceived stimulus corresponds to the stimulus value that leads to the same information for each of the alternative directions, and compare the model prediction to standard models considered in the literature (population vector, maximum likelihood, maximum-a-posteriori Bayesian inference). The models are applied to human performance in a motion discrimination task that induces perceptual misjudgements of a target direction of motion by task irrelevant motion in the spatial surround of the target stimulus (motion repulsion). By using the neurophysiological insight that surround motion suppresses neuronal responses to the target motion in the center, all models predicted the pattern of perceptual misjudgements. The variation of discrimination thresholds (error on the perceived value) was also explained through the changes of the total FI content with varying surround motion directions. The proposed FI decoding scheme incorporates recent neurophysiological evidence from macaque visual cortex showing that perceptual decisions do not rely on the most active neurons, but rather on the most informative neuronal responses. We statistically compare the prediction capability of the FI decoding approach and the standard decoding models. Notably, all models reproduced the variation of the perceived stimulus values for different surrounds, but with different neuronal tuning characteristics underlying perception. Compared to the FI approach the prediction power of the standard models was based on neurons with far wider tuning width and stronger surround suppression. Our study demonstrates that perceptual misjudgements can be based on neuronal populations encoding explicitly the available sensory information, and provides testable neurophysiological predictions on neuronal tuning characteristics underlying human perceptual decisions.     

Multisensory Oddity Detection as Bayesian Inference

A key goal for the perceptual system is to optimally combine information from all the senses that may be available in order to develop the most accurate and unified picture possible of the outside world. The contemporary theoretical framework of ideal observer maximum likelihood integration (MLI) has been highly successful in modelling how the human brain combines information from a variety of different sensory modalities. However, in various recent experiments involving multisensory stimuli of uncertain correspondence, MLI breaks down as a successful model of sensory combination. Within the paradigm of direct stimulus estimation, perceptual models which use Bayesian inference to resolve correspondence have recently been shown to generalize successfully to these cases where MLI fails. This approach has been known variously as model inference, causal inference or structure inference. In this paper, we examine causal uncertainty in another important class of multi-sensory perception paradigm – that of oddity detection and demonstrate how a Bayesian ideal observer also treats oddity detection as a structure inference problem. We validate this approach by showing that it provides an intuitive and quantitative explanation of an important pair of multi-sensory oddity detection experiments – involving cues across and within modalities – for which MLI previously failed dramatically, allowing a novel unifying treatment of within and cross modal multisensory perception. Our successful application of structure inference models to the new ‘oddity detection’ paradigm, and the resultant unified explanation of across and within modality cases provide further evidence to suggest that structure inference may be a commonly evolved principle for combining perceptual information in the brain.     

Variability of perceptual multistability: from brain state to individual trait

Few phenomena are as suitable as perceptual multistability to demonstrate that the brain constructively interprets sensory input. Several studies have outlined the neural circuitry involved in generating perceptual inference but only more recently has the individual variability of this inferential process been appreciated. Studies of the interaction of evoked and ongoing neural activity show that inference itself is not merely a stimulus-triggered process but is related to the context of the current brain state into which the processing of external stimulation is embedded. As brain states fluctuate, so does perception of a given sensory input. In multistability, perceptual fluctuation rates are consistent for a given individual but vary considerably between individuals. There has been some evidence for a genetic basis for these individual differences and recent morphometric studies of parietal lobe regions have identified neuroanatomical substrates for individual variability in spontaneous switching behaviour. Moreover, disrupting the function of these latter regions by transcranial magnetic stimulation yields systematic interference effects on switching behaviour, further arguing for a causal role of these regions in perceptual inference. Together, these studies have advanced our understanding of the biological mechanisms by which the brain constructs the contents of consciousness from sensory input.     

Structure Learning in a Sensorimotor Association Task

Learning is often understood as an organism''s gradual acquisition of the association between a given sensory stimulus and the correct motor response. Mathematically, this corresponds to regressing a mapping between the set of observations and the set of actions. Recently, however, it has been shown both in cognitive and motor neuroscience that humans are not only able to learn particular stimulus-response mappings, but are also able to extract abstract structural invariants that facilitate generalization to novel tasks. Here we show how such structure learning can enhance facilitation in a sensorimotor association task performed by human subjects. Using regression and reinforcement learning models we show that the observed facilitation cannot be explained by these basic models of learning stimulus-response associations. We show, however, that the observed data can be explained by a hierarchical Bayesian model that performs structure learning. In line with previous results from cognitive tasks, this suggests that hierarchical Bayesian inference might provide a common framework to explain both the learning of specific stimulus-response associations and the learning of abstract structures that are shared by different task environments.     

Learning to smell the roses: experience-dependent neural plasticity in human piriform and orbitofrontal cortices

It is widely presumed that odor quality is a direct outcome of odorant structure, but human studies indicate that molecular knowledge of an odorant is not always sufficient to predict odor quality. Indeed, the same olfactory input may generate different odor percepts depending on prior learning and experience. Combining functional magnetic resonance imaging with an olfactory paradigm of perceptual learning, we examined how sensory experience modifies odor perception and odor quality coding in the human brain. Prolonged exposure to a target odorant enhanced perceptual differentiation for odorants related in odor quality or functional group, an effect that was paralleled by learning-induced response increases in piriform cortex and orbitofrontal cortex (OFC). Critically, the magnitude of OFC activation predicted subsequent improvement in behavioral differentiation. Our findings suggest that neural representations of odor quality can be rapidly updated through mere perceptual experience, a mechanism that may underlie the development of odor perception.     

A Hierarchy of Time-Scales and the Brain

In this paper, we suggest that cortical anatomy recapitulates the temporal hierarchy that is inherent in the dynamics of environmental states. Many aspects of brain function can be understood in terms of a hierarchy of temporal scales at which representations of the environment evolve. The lowest level of this hierarchy corresponds to fast fluctuations associated with sensory processing, whereas the highest levels encode slow contextual changes in the environment, under which faster representations unfold. First, we describe a mathematical model that exploits the temporal structure of fast sensory input to track the slower trajectories of their underlying causes. This model of sensory encoding or perceptual inference establishes a proof of concept that slowly changing neuronal states can encode the paths or trajectories of faster sensory states. We then review empirical evidence that suggests that a temporal hierarchy is recapitulated in the macroscopic organization of the cortex. This anatomic-temporal hierarchy provides a comprehensive framework for understanding cortical function: the specific time-scale that engages a cortical area can be inferred by its location along a rostro-caudal gradient, which reflects the anatomical distance from primary sensory areas. This is most evident in the prefrontal cortex, where complex functions can be explained as operations on representations of the environment that change slowly. The framework provides predictions about, and principled constraints on, cortical structure–function relationships, which can be tested by manipulating the time-scales of sensory input.     

Surprised at all the entropy: hippocampal, caudate and midbrain contributions to learning from prediction errors

Influential concepts in neuroscientific research cast the brain a predictive machine that revises its predictions when they are violated by sensory input. This relates to the predictive coding account of perception, but also to learning. Learning from prediction errors has been suggested for take place in the hippocampal memory system as well as in the basal ganglia. The present fMRI study used an action-observation paradigm to investigate the contributions of the hippocampus, caudate nucleus and midbrain dopaminergic system to different types of learning: learning in the absence of prediction errors, learning from prediction errors, and responding to the accumulation of prediction errors in unpredictable stimulus configurations. We conducted analyses of the regions of interests' BOLD response towards these different types of learning, implementing a bootstrapping procedure to correct for false positives. We found both, caudate nucleus and the hippocampus to be activated by perceptual prediction errors. The hippocampal responses seemed to relate to the associative mismatch between a stored representation and current sensory input. Moreover, its response was significantly influenced by the average information, or Shannon entropy of the stimulus material. In accordance with earlier results, the habenula was activated by perceptual prediction errors. Lastly, we found that the substantia nigra was activated by the novelty of sensory input. In sum, we established that the midbrain dopaminergic system, the hippocampus, and the caudate nucleus were to different degrees significantly involved in the three different types of learning: acquisition of new information, learning from prediction errors and responding to unpredictable stimulus developments. We relate learning from perceptual prediction errors to the concept of predictive coding and related information theoretic accounts.     

A Bayesian Attractor Model for Perceptual Decision Making

Even for simple perceptual decisions, the mechanisms that the brain employs are still under debate. Although current consensus states that the brain accumulates evidence extracted from noisy sensory information, open questions remain about how this simple model relates to other perceptual phenomena such as flexibility in decisions, decision-dependent modulation of sensory gain, or confidence about a decision. We propose a novel approach of how perceptual decisions are made by combining two influential formalisms into a new model. Specifically, we embed an attractor model of decision making into a probabilistic framework that models decision making as Bayesian inference. We show that the new model can explain decision making behaviour by fitting it to experimental data. In addition, the new model combines for the first time three important features: First, the model can update decisions in response to switches in the underlying stimulus. Second, the probabilistic formulation accounts for top-down effects that may explain recent experimental findings of decision-related gain modulation of sensory neurons. Finally, the model computes an explicit measure of confidence which we relate to recent experimental evidence for confidence computations in perceptual decision tasks.     

Generative models for discovering sparse distributed representations.

We describe a hierarchical, generative model that can be viewed as a nonlinear generalization of factor analysis and can be implemented in a neural network. The model uses bottom-up, top-down and lateral connections to perform Bayesian perceptual inference correctly. Once perceptual inference has been performed the connection strengths can be updated using a very simple learning rule that only requires locally available information. We demonstrate that the network learns to extract sparse, distributed, hierarchical representations.     

Learning to use an invisible visual signal for perception

How does the brain construct a percept from sensory signals? One approach to this fundamental question is to investigate perceptual learning as induced by exposure to statistical regularities in sensory signals [1-7]. Recent studies showed that exposure to novel correlations between sensory signals can cause a signal to have new perceptual effects [2, 3]. In those studies, however, the signals were clearly visible. The automaticity of the learning was therefore difficult to determine. Here we investigate whether learning of this sort, which causes new effects on appearance, can be low level and automatic by employing a visual signal whose perceptual consequences were made invisible-a vertical disparity gradient masked by other depth cues. This approach excluded high-level influences such as attention or consciousness. Our stimulus for probing perceptual appearance was a rotating cylinder. During exposure, we introduced a new contingency between the invisible signal and the rotation direction of the cylinder. When subsequently presenting an ambiguously rotating version of the cylinder, we found that the invisible signal influenced the perceived rotation direction. This demonstrates that perception can rapidly undergo "structure learning" by automatically picking up novel contingencies between sensory signals, thus automatically recruiting signals for novel uses during the construction of a percept.     

What we see is how we are: new paradigms in visual research

As most sensory modalities, the visual system needs to deal with very fast changes in the environment. Instead of processing all sensory stimuli, the brain is able to construct a perceptual experience by combining selected sensory input with an ongoing internal activity. Thus, the study of visual perception needs to be approached by examining not only the physical properties of stimuli, but also the brain's ongoing dynamical states onto which these perturbations are imposed. At least three different models account for this internal dynamics. One model is based on cardinal cells where the activity of few cells by itself constitutes the neuronal correlate of perception, while a second model is based on a population coding that states that the neuronal correlate of perception requires distributed activity throughout many areas of the brain. A third proposition, known as the temporal correlation hypothesis states that the distributed neuronal populations that correlate with perception, are also defined by synchronization of the activity on a millisecond time scale. This would serve to encode contextual information by defining relations between the features of visual objects. If temporal properties of neural activity are important to establish the neural mechanisms of perception, then the study of appropriate dynamical stimuli should be instrumental to determine how these systems operate. The use of natural stimuli and natural behaviors such as free viewing, which features fast changes of internal brain states as seen by motor markers, is proposed as a new experimental paradigm to study visual perception.     

Task-induced neural covariability as a signature of approximate Bayesian learning and inference

Perception is often characterized computationally as an inference process in which uncertain or ambiguous sensory inputs are combined with prior expectations. Although behavioral studies have shown that observers can change their prior expectations in the context of a task, robust neural signatures of task-specific priors have been elusive. Here, we analytically derive such signatures under the general assumption that the responses of sensory neurons encode posterior beliefs that combine sensory inputs with task-specific expectations. Specifically, we derive predictions for the task-dependence of correlated neural variability and decision-related signals in sensory neurons. The qualitative aspects of our results are parameter-free and specific to the statistics of each task. The predictions for correlated variability also differ from predictions of classic feedforward models of sensory processing and are therefore a strong test of theories of hierarchical Bayesian inference in the brain. Importantly, we find that Bayesian learning predicts an increase in so-called “differential correlations” as the observer’s internal model learns the stimulus distribution, and the observer’s behavioral performance improves. This stands in contrast to classic feedforward encoding/decoding models of sensory processing, since such correlations are fundamentally information-limiting. We find support for our predictions in data from existing neurophysiological studies across a variety of tasks and brain areas. Finally, we show in simulation how measurements of sensory neural responses can reveal information about a subject’s internal beliefs about the task. Taken together, our results reinterpret task-dependent sources of neural covariability as signatures of Bayesian inference and provide new insights into their cause and their function.     

Learning in mammalian sensory cortex

The improvement of perceptual capabilities with training can offer important insight into the physiological basis of learning in the cerebral cortex. The rapid time course and ease with which some perceptual capabilities can improve suggest that learning is an integral part of normal perception. Electrophysiological and neuroimaging studies of sensory systems in the cortex suggest that the changes underlying perceptual learning can occur in a variety of areas and are likely to involve multiple mechanisms. In particular, recent psychological and physiological studies suggest that perceptual learning might often involve the task-specific suppression of signals that interfere with performance.     

Reinforcement Learning or Active Inference?

This paper questions the need for reinforcement learning or control theory when optimising behaviour. We show that it is fairly simple to teach an agent complicated and adaptive behaviours using a free-energy formulation of perception. In this formulation, agents adjust their internal states and sampling of the environment to minimize their free-energy. Such agents learn causal structure in the environment and sample it in an adaptive and self-supervised fashion. This results in behavioural policies that reproduce those optimised by reinforcement learning and dynamic programming. Critically, we do not need to invoke the notion of reward, value or utility. We illustrate these points by solving a benchmark problem in dynamic programming; namely the mountain-car problem, using active perception or inference under the free-energy principle. The ensuing proof-of-concept may be important because the free-energy formulation furnishes a unified account of both action and perception and may speak to a reappraisal of the role of dopamine in the brain.     

Cortical contributions to olfaction: plasticity and perception

In most sensory systems, the sensory cortex is the place where sensation approaches perception. As described in this review, olfaction is no different. The olfactory system includes both primary and higher order cortical regions. These cortical structures perform computations that take highly analytical afferent input and synthesize it into configural odor objects. Cortical plasticity plays an important role in this synthesis and may underlie olfactory perceptual learning. Olfactory cortex is also involved in odor memory and association of odors with multimodal input and contexts. Finally, the olfactory cortex serves as an important sensory gate, modulating information throughput based on recent experience and behavioral state.