Predictive Coding with Neural Transmission Delays: A Real-Time Temporal Alignment Hypothesis

The perceptual timescape: Perceptual history on the sub-second scale

There is a high-capacity store of brief time span (∼1000 ms) which information enters from perceptual processing, often called iconic memory or sensory memory. It is proposed that a main function of this store is to hold recent perceptual information in a temporally segregated representation, named the perceptual timescape. The perceptual timescape is a continually active representation of change and continuity over time that endows the perceived present with a perceived history. This is accomplished primarily by two kinds of time marking information: time distance information, which marks all items of information in the perceptual timescape according to how far in the past they occurred, and ordinal temporal information, which organises items of information in terms of their temporal order. Added to that is information about connectivity of perceptual objects over time. These kinds of information connect individual items over a brief span of time so as to represent change, persistence, and continuity over time. It is argued that there is a one-way street of information flow from perceptual processing either to the perceived present or directly into the perceptual timescape, and thence to working memory. Consistent with that, the information structure of the perceptual timescape supports postdictive reinterpretations of recent perceptual information. Temporal integration on a time scale of hundreds of milliseconds takes place in perceptual processing and does not draw on information in the perceptual timescape, which is concerned with temporal segregation, not integration.

Stimulus expectations do not modulate visual event-related potentials in probabilistic cueing designs

Humans and other animals can learn and exploit repeating patterns that occur within their environments. These learned patterns can be used to form expectations about future sensory events. Several influential predictive coding models have been proposed to explain how learned expectations influence the activity of stimulus-selective neurons in the visual system. These models specify reductions in neural response measures when expectations are fulfilled (termed expectation suppression) and increases following surprising sensory events. However, there is currently scant evidence for expectation suppression in the visual system when confounding factors are taken into account. Effects of surprise have been observed in blood oxygen level dependent (BOLD) signals, but not when using electrophysiological measures. To provide a strong test for expectation suppression and surprise effects we performed a predictive cueing experiment while recording electroencephalographic (EEG) data. Participants (n=48) learned cue-face associations during a training session and were then exposed to these cue-face pairs in a subsequent experiment. Using univariate analyses of face-evoked event-related potentials (ERPs) we did not observe any differences across expected (90% probability), neutral (50%) and surprising (10%) face conditions. Across these comparisons, Bayes factors consistently favoured the null hypothesis throughout the time-course of the stimulus-evoked response. When using multivariate pattern analysis we did not observe above-chance classification of expected and surprising face-evoked ERPs. By contrast, we found robust within- and across-trial stimulus repetition effects. Our findings do not support predictive coding-based accounts that specify reduced prediction error signalling when perceptual expectations are fulfilled. They instead highlight the utility of other types of predictive processing models that describe expectation-related phenomena in the visual system without recourse to prediction error signalling.

Computational assessment of visual coding across mouse brain areas and behavioural states

Introduction

Our brain is bombarded by a diverse range of visual stimuli, which are converted into corresponding neuronal responses and processed throughout the visual system. The neural activity patterns that result from these external stimuli vary depending on the object or scene being observed, but they also change as a result of internal or behavioural states. This raises the question of to what extent it is possible to predict the presented visual stimuli from neural activity across behavioural states, and how this varies in different brain regions.

Methods

To address this question, we assessed the computational capacity of decoders to extract visual information in awake behaving mice, by analysing publicly available standardised datasets from the Allen Brain Institute. We evaluated how natural movie frames can be distinguished based on the activity of units recorded in distinct brain regions and under different behavioural states. This analysis revealed the spectrum of visual information present in different brain regions in response to binary and multiclass classification tasks.

Results

Visual cortical areas showed highest classification accuracies, followed by thalamic and midbrain regions, with hippocampal regions showing close to chance accuracy. In addition, we found that behavioural variability led to a decrease in decoding accuracy, whereby large behavioural changes between train and test sessions reduced the classification performance of the decoders. A generalised linear model analysis suggested that this deterioration in classification might be due to an independent modulation of neural activity by stimulus and behaviour. Finally, we reconstructed the natural movie frames from optimal linear classifiers, and observed a strong similarity between reconstructed and actual movie frames. However, the similarity was significantly higher when the decoders were trained and tested on sessions with similar behavioural states.

Conclusion

Our analysis provides a systematic assessment of visual coding in the mouse brain, and sheds light on the spectrum of visual information present across brain areas and behavioural states.

Learning heterogeneous delays in a layer of spiking neurons for fast motion detection

The precise timing of spikes emitted by neurons plays a crucial role in shaping the response of efferent biological neurons. This temporal dimension of neural activity holds significant importance in understanding information processing in neurobiology, especially for the performance of neuromorphic hardware, such as event-based cameras. Nonetheless, many artificial neural models disregard this critical temporal dimension of neural activity. In this study, we present a model designed to efficiently detect temporal spiking motifs using a layer of spiking neurons equipped with heterogeneous synaptic delays. Our model capitalizes on the diverse synaptic delays present on the dendritic tree, enabling specific arrangements of temporally precise synaptic inputs to synchronize upon reaching the basal dendritic tree. We formalize this process as a time-invariant logistic regression, which can be trained using labeled data. To demonstrate its practical efficacy, we apply the model to naturalistic videos transformed into event streams, simulating the output of the biological retina or event-based cameras. To evaluate the robustness of the model in detecting visual motion, we conduct experiments by selectively pruning weights and demonstrate that the model remains efficient even under significantly reduced workloads. In conclusion, by providing a comprehensive, event-driven computational building block, the incorporation of heterogeneous delays has the potential to greatly improve the performance of future spiking neural network algorithms, particularly in the context of neuromorphic chips.

Spike-timing dependent plasticity partially compensates for neural delays in a multi-layered network of motion-sensitive neurons

The ability of the brain to represent the external world in real-time is impacted by the fact that neural processing takes time. Because neural delays accumulate as information progresses through the visual system, representations encoded at each hierarchical level are based upon input that is progressively outdated with respect to the external world. This 'representational lag' is particularly relevant to the task of localizing a moving object-because the object's location changes with time, neural representations of its location potentially lag behind its true location. Converging evidence suggests that the brain has evolved mechanisms that allow it to compensate for its inherent delays by extrapolating the position of moving objects along their trajectory. We have previously shown how spike-timing dependent plasticity (STDP) can achieve motion extrapolation in a two-layer, feedforward network of velocity-tuned neurons, by shifting the receptive fields of second layer neurons in the opposite direction to a moving stimulus. The current study extends this work by implementing two important changes to the network to bring it more into line with biology: we expanded the network to multiple layers to reflect the depth of the visual hierarchy, and we implemented more realistic synaptic time-courses. We investigate the accumulation of STDP-driven receptive field shifts across several layers, observing a velocity-dependent reduction in representational lag. These results highlight the role of STDP, operating purely along the feedforward pathway, as a developmental strategy for delay compensation.

Visual Information Is Predictively Encoded in Occipital Alpha/Low-Beta Oscillations

Hierarchical predictive coding networks are a general model of sensory processing in the brain. Under neural delays, these networks have been suggested to naturally generate oscillatory activity in approximately the α frequency range (∼8-12 Hz). This suggests that α oscillations, a prominent feature of EEG recordings, may be a spectral "fingerprint" of predictive sensory processing. Here, we probed this possibility by investigating whether oscillations over the visual cortex predictively encode visual information. Specifically, we examined whether their power carries information about the position of a moving stimulus, in a temporally predictive fashion. In two experiments (N = 32, 18 female; N = 34, 17 female), participants viewed an apparent-motion stimulus moving along a circular path while EEG was recorded. To investigate the encoding of stimulus-position information, we developed a method of deriving probabilistic spatial maps from oscillatory power estimates. With this method, we demonstrate that it is possible to reconstruct the trajectory of a moving stimulus from α/low-β oscillations, tracking its position even across unexpected motion reversals. We also show that future position representations are activated in the absence of direct visual input, demonstrating that temporally predictive mechanisms manifest in α/β band oscillations. In a second experiment, we replicate these findings and show that the encoding of information in this range is not driven by visual entrainment. By demonstrating that occipital α/β oscillations carry stimulus-related information, in a temporally predictive fashion, we provide empirical evidence of these rhythms as a spectral "fingerprint" of hierarchical predictive processing in the human visual system.SIGNIFICANCE STATEMENT "Hierarchical predictive coding" is a general model of sensory information processing in the brain. When in silico predictive coding models are constrained by neural transmission delays, their activity naturally oscillates in roughly the α range (∼8-12 Hz). Using time-resolved EEG decoding, we show that neural rhythms in this approximate range (α/low-β) over the human visual cortex predictively encode the position of a moving stimulus. From the amplitude of these oscillations, we are able to reconstruct the stimulus' trajectory, revealing signatures of temporally predictive processing. This provides direct neural evidence linking occipital α/β rhythms to predictive visual processing, supporting the emerging view of such oscillations as a potential spectral "fingerprint" of hierarchical predictive processing in the human visual system.

Robust and efficient representations of dynamic stimuli in hierarchical neural networks via temporal smoothing

Introduction

Efficient coding that minimizes informational redundancy of neural representations is a widely accepted neural coding principle. Despite the benefit, maximizing efficiency in neural coding can make neural representation vulnerable to random noise. One way to achieve robustness against random noise is smoothening neural responses. However, it is not clear whether the smoothness of neural responses can hold robust neural representations when dynamic stimuli are processed through a hierarchical brain structure, in which not only random noise but also systematic error due to temporal lag can be induced.

Methods

In the present study, we showed that smoothness via spatio-temporally efficient coding can achieve both efficiency and robustness by effectively dealing with noise and neural delay in the visual hierarchy when processing dynamic visual stimuli.

Results

The simulation results demonstrated that a hierarchical neural network whose bidirectional synaptic connections were learned through spatio-temporally efficient coding with natural scenes could elicit neural responses to visual moving bars similar to those to static bars with the identical position and orientation, indicating robust neural responses against erroneous neural information. It implies that spatio-temporally efficient coding preserves the structure of visual environments locally in the neural responses of hierarchical structures.

Discussion

The present results suggest the importance of a balance between efficiency and robustness in neural coding for visual processing of dynamic stimuli across hierarchical brain structures.

The lateralized flash-lag illusion: A psychophysical and pupillometry study

The flash-lag illusion (FLI) is a visual phenomenon where a flashed object, either co-localized or in physical alignment with another continuously moving object, is perceived to lag behind the path of the moving object. In the present study, we reveal an anisotropy of the FLI between the lateral visual fields that was expressed psychophysically as different points of subjective equality, depending on the hemifield in which the stimuli appeared. Specifically, the study confirmed that, as seen in two previous studies, the FLI was significantly larger in the left visual field (LVF) than in the right (RVF). In addition, pupil dilations were larger in the RVF than in the LVF as well as returning to baseline levels more rapidly in the LVF. We interpret these findings as converging on revealing more efficient spatial and attentional processing and, in turn, extrapolation of motion in the LVF/right hemisphere than in the RVF/left hemisphere.

How the brain stays in sync with the real world

The brain can predict the location of a moving object to compensate for the delays caused by the processing of neural signals.

Position representations of moving objects align with real-time position in the early visual response

When interacting with the dynamic world, the brain receives outdated sensory information, due to the time required for neural transmission and processing. In motion perception, the brain may overcome these fundamental delays through predictively encoding the position of moving objects using information from their past trajectories. In the present study, we evaluated this proposition using multivariate analysis of high temporal resolution electroencephalographic data. We tracked neural position representations of moving objects at different stages of visual processing, relative to the real-time position of the object. During early stimulus-evoked activity, position representations of moving objects were activated substantially earlier than the equivalent activity evoked by unpredictable flashes, aligning the earliest representations of moving stimuli with their real-time positions. These findings indicate that the predictability of straight trajectories enables full compensation for the neural delays accumulated early in stimulus processing, but that delays still accumulate across later stages of cortical processing.

Precise Spiking Motifs in Neurobiological and Neuromorphic Data

Why do neurons communicate through spikes? By definition, spikes are all-or-none neural events which occur at continuous times. In other words, spikes are on one side binary, existing or not without further details, and on the other, can occur at any asynchronous time, without the need for a centralized clock. This stands in stark contrast to the analog representation of values and the discretized timing classically used in digital processing and at the base of modern-day neural networks. As neural systems almost systematically use this so-called event-based representation in the living world, a better understanding of this phenomenon remains a fundamental challenge in neurobiology in order to better interpret the profusion of recorded data. With the growing need for intelligent embedded systems, it also emerges as a new computing paradigm to enable the efficient operation of a new class of sensors and event-based computers, called neuromorphic, which could enable significant gains in computation time and energy consumption-a major societal issue in the era of the digital economy and global warming. In this review paper, we provide evidence from biology, theory and engineering that the precise timing of spikes plays a crucial role in our understanding of the efficiency of neural networks.

A quantitative model reveals a frequency ordering of prediction and prediction-error signals in the human brain

The human brain is proposed to harbor a hierarchical predictive coding neuronal network underlying perception, cognition, and action. In support of this theory, feedforward signals for prediction error have been reported. However, the identification of feedback prediction signals has been elusive due to their causal entanglement with prediction-error signals. Here, we use a quantitative model to decompose these signals in electroencephalography during an auditory task, and identify their spatio-spectral-temporal signatures across two functional hierarchies. Two prediction signals are identified in the period prior to the sensory input: a low-level signal representing the tone-to-tone transition in the high beta frequency band, and a high-level signal for the multi-tone sequence structure in the low beta band. Subsequently, prediction-error signals dependent on the prior predictions are found in the gamma band. Our findings reveal a frequency ordering of prediction signals and their hierarchical interactions with prediction-error signals supporting predictive coding theory.

A computational framework can extract spatio-spectro-temporal neural signatures corresponding to hierarchical prediction and prediction errors in a local-global auditory task.

Spatio-Temporally Efficient Coding Assigns Functions to Hierarchical Structures of the Visual System

Hierarchical structures constitute a wide array of brain areas, including the visual system. One of the important questions regarding visual hierarchical structures is to identify computational principles for assigning functions that represent the external world to hierarchical structures of the visual system. Given that visual hierarchical structures contain both bottom-up and top-down pathways, the derived principles should encompass these bidirectional pathways. However, existing principles such as predictive coding do not provide an effective principle for bidirectional pathways. Therefore, we propose a novel computational principle for visual hierarchical structures as spatio-temporally efficient coding underscored by the efficient use of given resources in both neural activity space and processing time. This coding principle optimises bidirectional information transmissions over hierarchical structures by simultaneously minimising temporal differences in neural responses and maximising entropy in neural representations. Simulations demonstrated that the proposed spatio-temporally efficient coding was able to assign the function of appropriate neural representations of natural visual scenes to visual hierarchical structures. Furthermore, spatio-temporally efficient coding was able to predict well-known phenomena, including deviations in neural responses to unlearned inputs and bias in preferred orientations. Our proposed spatio-temporally efficient coding may facilitate deeper mechanistic understanding of the computational processes of hierarchical brain structures.

Perception in real-time: predicting the present, reconstructing the past

We feel that we perceive events in the environment as they unfold in real-time. However, this intuitive view of perception is impossible to implement in the nervous system due to biological constraints such as neural transmission delays. I propose a new way of thinking about real-time perception: at any given moment, instead of representing a single timepoint, perceptual mechanisms represent an entire timeline. On this timeline, predictive mechanisms predict ahead to compensate for delays in incoming sensory input, and reconstruction mechanisms retroactively revise perception when those predictions do not come true. This proposal integrates and extends previous work to address a crucial gap in our understanding of a fundamental aspect of our everyday life: the experience of perceiving the present.

Early and late evoked brain responses differentially reflect feature encoding and perception in the flash-lag illusion

In the flash-lag illusion (FLI), the position of a flash presented ahead of a moving bar is mislocalized, so the flash appears to lag the bar. Currently, it is not clear whether this effect is due to early perceptual-related neural processes such as motion extrapolation or reentrant processing, or due to later feedback processing relating to postdiction, i.e., retroactively altered perception. We presented 17 participants with the FLI paradigm while recording EEG. A central flash occurred either 51 ms ("early") or 16 ms ("late") before the bar moving from left to right reached the screen center. Participants judged whether the flash appeared to the right ("no flash lag illusion") or to the left ("flash-lag illusion") of the bar. Using single-trial linear modeling, we examined the influence of timing ("early" vs. "late") and perception ("illusion" vs. "no illusion") on flash-evoked brain responses and estimated the cortical sources underlying the FLI. An earlier frontal and occipital component (200-276 ms) differentiated time-locked early vs. late stimulus presentation, indicating that early evoked brain responses reflect feature encoding in the FLI. Perception of the FLI was associated with a late window (368-452 ms) in the ERP, with larger deflections for illusion than no illusion trials, localized to the left inferior occipital gyrus. This suggests a postdiction-related reconstruction of ambiguous sensory stimulation involving late processes in the occipito-temporal cortex, previously associated with temporal integration phenomena. Our findings indicate that perception of the FLI relies on an interplay between ongoing stimulus encoding of the moving bar and feedback processing of the flash, which takes place at later integration stages.

Resolving visual motion through perceptual gaps

Perceptual gaps can be caused by objects in the foreground temporarily occluding objects in the background or by eyeblinks, which briefly but frequently interrupt visual information. Resolving visual motion across perceptual gaps is particularly challenging, as object position changes during the gap. We examine how visual motion is maintained and updated through externally driven (occlusion) and internally driven (eyeblinks) perceptual gaps. Focusing on both phenomenology and potential mechanisms such as suppression, extrapolation, and integration, we present a framework for how perceptual gaps are resolved over space and time. We finish by highlighting critical questions and directions for future work.

Predictive Visual Motion Extrapolation Emerges Spontaneously and without Supervision at Each Layer of a Hierarchical Neural Network with Spike-Timing-Dependent Plasticity

The fact that the transmission and processing of visual information in the brain takes time presents a problem for the accurate real-time localization of a moving object. One way this problem might be solved is extrapolation: using an object's past trajectory to predict its location in the present moment. Here, we investigate how a simulated in silico layered neural network might implement such extrapolation mechanisms, and how the necessary neural circuits might develop. We allowed an unsupervised hierarchical network of velocity-tuned neurons to learn its connectivity through spike-timing-dependent plasticity (STDP). We show that the temporal contingencies between the different neural populations that are activated by an object as it moves causes the receptive fields of higher-level neurons to shift in the direction opposite to their preferred direction of motion. The result is that neural populations spontaneously start to represent moving objects as being further along their trajectory than where they were physically detected. Because of the inherent delays of neural transmission, this effectively compensates for (part of) those delays by bringing the represented position of a moving object closer to its instantaneous position in the world. Finally, we show that this model accurately predicts the pattern of perceptual mislocalization that arises when human observers are required to localize a moving object relative to a flashed static object (the flash-lag effect; FLE).SIGNIFICANCE STATEMENT Our ability to track and respond to rapidly changing visual stimuli, such as a fast-moving tennis ball, indicates that the brain is capable of extrapolating the trajectory of a moving object to predict its current position, despite the delays that result from neural transmission. Here, we show how the neural circuits underlying this ability can be learned through spike-timing-dependent synaptic plasticity and that these circuits emerge spontaneously and without supervision. This demonstrates how the neural transmission delays can, in part, be compensated to implement the extrapolation mechanisms required to predict where a moving object is at the present moment.

A link between synaptic plasticity and reorganization of brain activity in Parkinson's disease

The link between synaptic plasticity and reorganization of brain activity in health and disease remains a scientific challenge. We examined this question in Parkinson's disease (PD) where functional up-regulation of postsynaptic D₂ receptors has been documented while its significance at the neural activity level has never been identified. We investigated cortico-subcortical plasticity in PD using the oculomotor system as a model to study reorganization of dopaminergic networks. This model is ideal because this system reorganizes due to frontal-to-parietal shifts in blood oxygen level-dependent (BOLD) activity. We tested the prediction that functional activation plasticity is associated with postsynaptic dopaminergic modifications by combining positron emission tomography/functional magnetic resonance imaging to investigate striatal postsynaptic reorganization of dopamine D₂ receptors (using ¹¹C-raclopride) and neural activation in PD. We used covariance (connectivity) statistics at molecular and functional levels to probe striato-cortical reorganization in PD in on/off medication states to show that functional and molecular forms of reorganization are related. D₂ binding across regions defined by prosaccades showed increased molecular connectivity between both caudate/putamen and hyperactive parietal eye fields in PD in contrast with frontal eye fields in controls, in line with the shift model. Concerning antisaccades, parietal-striatal connectivity dominated in again in PD, unlike frontal regions. Concerning molecular-BOLD covariance, a striking sign reversal was observed: PD patients showed negative frontal-putamen functional-molecular associations, consistent with the reorganization shift, in contrast with the positive correlations observed in controls. Follow-up analysis in off-medication PD patients confirmed the negative BOLD-molecular correlation. These results provide a link among BOLD responses, striato-cortical synaptic reorganization, and neural plasticity in PD.

Evaluating the evidence for expectation suppression in the visual system

Reports of expectation suppression have shaped the development of influential predictive coding-based theories of visual perception. However recent work has highlighted confounding factors that may mimic or inflate expectation suppression effects. In this review, we describe four confounds that are prevalent across experiments that tested for expectation suppression: effects of surprise, attention, stimulus repetition and adaptation, and stimulus novelty. With these confounds in mind we then critically review the evidence for expectation suppression across probabilistic cueing, statistical learning, oddball, action-outcome learning and apparent motion designs. We found evidence for expectation suppression within a specific subset of statistical learning designs that involved weeks of sequence learning prior to neural activity measurement. Across other experimental contexts, whereby stimulus appearance probabilities were learned within one or two testing sessions, there was inconsistent evidence for genuine expectation suppression. We discuss how an absence of expectation suppression could inform models of predictive processing, repetition suppression and perceptual decision-making. We also provide suggestions for designing experiments that may better test for expectation suppression in future work.

Motion Extrapolation in Visual Processing: Lessons from 25 Years of Flash-Lag Debate

Because of the delays inherent in neural transmission, the brain needs time to process incoming visual information. If these delays were not somehow compensated, we would consistently mislocalize moving objects behind their physical positions. Twenty-five years ago, Nijhawan used a perceptual illusion he called the flash-lag effect (FLE) to argue that the brain's visual system solves this computational challenge by extrapolating the position of moving objects (Nijhawan, 1994). Although motion extrapolation had been proposed a decade earlier (e.g., Finke et al., 1986), the proposal that it caused the FLE and functioned to compensate for computational delays was hotly debated in the years that followed, with several alternative interpretations put forth to explain the effect. Here, I argue, 25 years later, that evidence from behavioral, computational, and particularly recent functional neuroimaging studies converges to support the existence of motion extrapolation mechanisms in the visual system, as well as their causal involvement in the FLE. First, findings that were initially argued to challenge the motion extrapolation model of the FLE have since been explained, and those explanations have been tested and corroborated by more recent findings. Second, motion extrapolation explains the spatial shifts observed in several FLE conditions that cannot be explained by alternative (temporal) models of the FLE. Finally, neural mechanisms that actually perform motion extrapolation have been identified at multiple levels of the visual system, in multiple species, and with multiple different methods. I outline key questions that remain, and discuss possible directions for future research.

Motion extrapolation in the High-Phi illusion: Analogous but dissociable effects on perceived position and perceived motion

A range of visual illusions, including the much-studied flash-lag effect, demonstrate that neural signals coding for motion and position interact in the visual system. One interpretation of these illusions is that they are the consequence of motion extrapolation mechanisms in the early visual system. Here, we study the recently reported High-Phi illusion to investigate whether it might be caused by the same underlying mechanisms. In the High-Phi illusion, a rotating texture is abruptly replaced by a new, uncorrelated texture. This leads to the percept of a large illusory jump, which can be forward or backward depending on the duration of the initial motion sequence (the inducer). To investigate whether this motion illusion also leads to illusions of perceived position, in three experiments we asked observers to localize briefly flashed targets presented concurrently with the new texture. Our results replicate the original finding of perceived forward and backward jumps, and reveal an illusion of perceived position. Like the observed effects on illusory motion, these position shifts could be forward or backward, depending on the duration of the inducer: brief inducers caused forward mislocalization, and longer inducers caused backward mislocalization. Additionally, we found that both jumps and mislocalizations scaled in magnitude with the speed of the inducer. Interestingly, forward position shifts were observed at shorter inducer durations than forward jumps. We interpret our results as an interaction of extrapolation and correction-for-extrapolation, and discuss possible mechanisms in the early visual system that might carry out these computations.

A neural network trained for prediction mimics diverse features of biological neurons and perception

Recent work has shown that convolutional neural networks (CNNs) trained on image recognition tasks can serve as valuable models for predicting neural responses in primate visual cortex. However, these models typically require biologically-infeasible levels of labeled training data, so this similarity must at least arise via different paths. In addition, most popular CNNs are solely feedforward, lacking a notion of time and recurrence, whereas neurons in visual cortex produce complex time-varying responses, even to static inputs. Towards addressing these inconsistencies with biology, here we study the emergent properties of a recurrent generative network that is trained to predict future video frames in a self-supervised manner. Remarkably, the resulting model is able to capture a wide variety of seemingly disparate phenomena observed in visual cortex, ranging from single-unit response dynamics to complex perceptual motion illusions, even when subjected to highly impoverished stimuli. These results suggest potentially deep connections between recurrent predictive neural network models and computations in the brain, providing new leads that can enrich both fields.

Predictions drive neural representations of visual events ahead of incoming sensory information

The transmission of sensory information through the visual system takes time. As a result of these delays, the visual information available to the brain always lags behind the timing of events in the present moment. Compensating for these delays is crucial for functioning within dynamic environments, since interacting with a moving object (e.g., catching a ball) requires real-time localization of the object. One way the brain might achieve this is via prediction of anticipated events. Using time-resolved decoding of electroencephalographic (EEG) data, we demonstrate that the visual system represents the anticipated future position of a moving object, showing that predictive mechanisms activate the same neural representations as afferent sensory input. Importantly, this activation is evident before sensory input corresponding to the stimulus position is able to arrive. Finally, we demonstrate that, when predicted events do not eventuate, sensory information arrives too late to prevent the visual system from representing what was expected but never presented. Taken together, we demonstrate how the visual system can implement predictive mechanisms to preactivate sensory representations, and argue that this might allow it to compensate for its own temporal constraints, allowing us to interact with dynamic visual environments in real time.

How can we play together? Temporal inconsistencies in neural coding of music

If sensory organs encode environment, this code must be decoded to perception. The currently dominant theory of perception - predictive coding - assumes a "Bayesian decoder," a probability function, which will present (to whom?) an optimal guess, given previous encodings of the environment - old codes testing new codes. Such a process would delay perception noticeably. This is inconsistent with the perception of music, which for several reasons must be direct.