Predictive processing offers a unifying view of neural computation, proposing that brains continuously anticipate sensory input and update internal models based on prediction errors. In this talk, I will present converging evidence for the computational mechanisms underlying this framework across human neuroscience and deep neural networks. I will begin with recent work showing that large-scale distributed prediction-error encoding in the human brain directly predicts how sensory representations reorganize through predictive learning. I will then turn to PredNet, a popular predictive coding inspired deep network that has been widely used to model real-world biological vision systems. Using dynamic stimuli generated with our Spatiotemporal Style Transfer algorithm, we demonstrate that PredNet relies primarily on low-level spatiotemporal structure and remains insensitive to high-level content, revealing limits in its generalization capacity. Finally, I will discuss new recurrent vision models that integrate top-down feedback connections with intrinsic neural variability, uncovering a dual mechanism for robust sensory coding in which neural variability decorrelates unit responses, while top-down feedback stabilizes network dynamics. Together, these results outline how prediction error signaling and top-down feedback pathways shape adaptive sensory processing in biological and artificial systems.

The application of Virtual Reality (VR) environments allows us to experimentally dissociate social input and responses, opening powerful avenues of inquiry into the dynamics of social influence and the physiological and neural mechanisms of collective behaviour. A key task for the nervous system is to make sense of complex streams of potentially-informative sensory input, allowing appropriate, relatively low-dimensional, motor actions to be taken, sometimes under conditions of considerable time constraint. The student will employ fully immersive ‘holographic’ VR to investigate the behavioural mechanisms by which freely-swimming zebrafish obtain both social and non-social sensory information from their surroundings, and how they use this to inform movement decisions. Immersive VR allows extremely precise control over the appearance, body postural changes, and motion, allowing photorealistic virtual individuals to interact dynamically with unrestrained real animals. Similar to a method that has transformed neuroscience — the dynamic patch clamp paradigm in which inputs to neurons can be based on fast closed-loop measurements of their present behaviour — VR creates the possibility for a ‘dynamic social patch clamp’ paradigm in which we can develop, and interrogate, decision-making models by integrating virtual organisms in the same environment as real individuals. This tool will help us to infer the sensory basis of social influence, the causality of influence in (small) social networks, to provide highly repeatable stimuli (allowing us to evaluate inter-individual and within-individual variation) and to interrogate the feedback loops inherent in social dynamics. For more information see: https://www.smartnets-etn.eu/using-immersive-virtual-reality-vr-to-determine-causal-relationships-in-animal-social-networks/

The nervous system orchestrates adaptive behaviors by intricately coordinating responses to internal cues and environmental stimuli. This involves integrating sensory input, managing competing motivational states, and drawing on past experiences to anticipate future outcomes. While traditional models attribute this complexity to interactions between the mesocorticolimbic system and hypothalamic centers, the specific nodes of integration have remained elusive. Recent research, including our own, sheds light on the midline thalamus's overlooked role in this process. We propose that the midline thalamus integrates internal states with memory and emotional signals to guide adaptive behaviors. Our investigations into midline thalamic neuronal circuits have provided crucial insights into the neural mechanisms behind flexibility and adaptability. Understanding these processes is essential for deciphering human behavior and conditions marked by impaired motivation and emotional processing. Our research aims to contribute to this understanding, paving the way for targeted interventions and therapies to address such impairments.

The nervous system is fundamentally a closed loop control device: the output of actions continually influences the internal state and subsequent actions. This is true at the single cell and even the molecular level, where “actions” take the form of signals that are fed back to achieve a variety of functions, including homeostasis, excitability and various kinds of multistability that allow switching and storage of memory. It is also true at the behavioural level, where an animal’s motor actions directly influence sensory input on short timescales, and higher level information about goals and intended actions are continually updated on the basis of current and past actions. Studying the brain in a closed loop setting requires a multidisciplinary approach, leveraging engineering and theory as well as advances in measuring and manipulating the nervous system. I will describe our recent attempts to achieve this fusion of approaches at multiple levels in the nervous system, from synaptic signalling to closed loop brain machine interfaces.

The sense of agency refers to the subjective feeling of controlling one's own behavior and, through them, external events. Why is this subjective feeling important for humans? Is it just a by-product of our actions? Previous studies have shown that the sense of agency can affect the intensity of sensory input because we predict the input from our motor intention. However, my research has found that the sense of agency plays more roles than just predictions. It enhances perceptual processes of sensory input and potentially helps to harvest more information about the link between the external world and the self. Furthermore, our recent research found both indirect and direct evidence that the sense of agency is important for people's exploratory behaviors, and this may be linked to proximal exploitations of one's control in the environment. In this talk, I will also introduce the paradigms we use to study the sense of agency as a result of perceptual processes, and our findings of individual differences in this sense and the implications.

Neurons in the rodent hippocampus appear to encode the position of the animal in physical space during movement. Individual ``place cells'' fire in restricted sub-regions of an environment, a feature often taken as evidence that the hippocampus encodes a map of space that subserves navigation. But these same neurons exhibit complex responses to many other variables that defy explanation by position alone, and the hippocampus is known to be more broadly critical for memory formation. Here we elaborate and test a theory of hippocampal coding which produces place cells as a general consequence of efficient memory coding. We constructed neural networks that actively exploit the correlations between memories in order to learn compressed representations of experience. Place cells readily emerged in the trained model, due to the correlations in sensory input between experiences at nearby locations. Notably, these properties were highly sensitive to the compressibility of the sensory environment, with place field size and population coding level in dynamic opposition to optimally encode the correlations between experiences. The effects of learning were also strongly biphasic: nearby locations are represented more similarly following training, while locations with intermediate similarity become increasingly decorrelated, both distance-dependent effects that scaled with the compressibility of the input features. Using virtual reality and 2-photon functional calcium imaging in head-fixed mice, we recorded the simultaneous activity of thousands of hippocampal neurons during virtual exploration to test these predictions. Varying the compressibility of sensory information in the environment produced systematic changes in place cell properties that reflected the changing input statistics, consistent with the theory. We similarly identified representational plasticity during learning, which produced a distance-dependent exchange between compression and pattern separation. These results motivate a more domain-general interpretation of hippocampal computation, one that is naturally compatible with earlier theories on the circuit's importance for episodic memory formation. Work done in collaboration with James Priestley, Lorenzo Posani, Marcus Benna, Attila Losonczy.

We hardly notice when there is a speck on our glasses, the obstructed visual information seems to be magically filled in. The mechanistic basis for this fundamental perceptual phenomenon has, however, remained obscure. What enables neurons in the visual system to respond to context when the stimulus is not available? While feedforward information drives the activity in cortex, feedback information is thought to provide contextual signals that are merely modulatory. We have made the discovery that mouse primary visual cortical neurons are strongly driven by feedback projections from higher visual areas when their feedforward sensory input from the retina is missing. This drive is so strong that it makes visual cortical neurons fire as much as if they were receiving a direct sensory input. These signals are likely used to predict input from the feedforward pathway. Preliminary results show that these feedback projections are strongly influenced by experience and learning.

We hardly notice when there is a speck on our glasses, the obstructed visual information seems to be magically filled in. The visual system uses visual context to predict the content of the stimulus. What enables neurons in the visual system to respond to context when the stimulus is not available? In cortex, sensory processing is based on a combination of feedforward information arriving from sensory organs, and feedback information that originates in higher-order areas. Whereas feedforward information drives the activity in cortex, feedback information is thought to provide contextual signals that are merely modulatory. We have made the exciting discovery that mouse primary visual cortical neurons are strongly driven by feedback projections from higher visual areas, in particular when their feedforward sensory input from the retina is missing. This drive is so strong that it makes visual cortical neurons fire as much as if they were receiving a direct sensory input.

Neural circuits exhibit complex activity patterns, both spontaneously and in response to external stimuli. Information encoding and learning in neural circuits depend on the ability of time-varying stimuli to control spontaneous network activity. In particular, variability arising from the sensitivity to initial conditions of recurrent cortical circuits can limit the information conveyed about the sensory input. Spiking and firing rate network models can exhibit such sensitivity to initial conditions that are reflected in their dynamic entropy rate and attractor dimensionality computed from their full Lyapunov spectrum. I will show how chaos in both spiking and rate networks depends on biophysical properties of neurons and the statistics of time-varying stimuli. In spiking networks, increasing the input rate or coupling strength aids in controlling the driven target circuit, which is reflected in both a reduced trial-to-trial variability and a decreased dynamic entropy rate. With sufficiently strong input, a transition towards complete network state control occurs. Surprisingly, this transition does not coincide with the transition from chaos to stability but occurs at even larger values of external input strength. Controllability of spiking activity is facilitated when neurons in the target circuit have a sharp spike onset, thus a high speed by which neurons launch into the action potential. I will also discuss chaos and controllability in firing-rate networks in the balanced state. For these, external control of recurrent dynamics strongly depends on correlations in the input. This phenomenon was studied with a non-stationary dynamic mean-field theory that determines how the activity statistics and the largest Lyapunov exponent depend on frequency and amplitude of the input, recurrent coupling strength, and network size. This shows that uncorrelated inputs facilitate learning in balanced networks. The results highlight the potential of Lyapunov spectrum analysis as a diagnostic for machine learning applications of recurrent networks. They are also relevant in light of recent advances in optogenetics that allow for time-dependent stimulation of a select population of neurons.

The pupil provides a rich, non-invasive measure of the neural bases of perception and cognition, and has been of particular value in uncovering the role of arousal-linked neuromodulation, which alters cortical processing as well as pupil size. But pupil size is subject to a multitude of influences, which complicates unique interpretation. We measured pupils of observers experiencing perceptual multistability -- an ever-changing subjective percept in the face of unchanging but inconclusive sensory input. In separate conditions the endogenously generated perceptual changes were either task-relevant or not, allowing a separation between perception-related and task-related pupil signals. Perceptual changes were marked by a complex pupil response that could be decomposed into two components: a dilation tied to task execution and plausibly indicative of an arousal-linked noradrenaline surge, and an overlapping constriction tied to the perceptual transient and plausibly a marker of altered visual cortical representation. Constriction, but not dilation, amplitude systematically depended on the time interval between perceptual changes, possibly providing an overt index of neural adaptation. These results show that the pupil provides a simultaneous reading on interacting but dissociable neural processes during perceptual multistability, and suggest that arousal-linked neuromodulation shapes action but not perception in these circumstances. This presentation covers work that was published in e-life

Predictive coding represents a promising framework for understanding brain function, postulating that the brain continuously inhibits predictable sensory input, ensuring a preferential processing of surprising elements. A central aspect of this view on cortical computation is its hierarchical connectivity, involving recurrent message passing between excitatory bottom-up signals and inhibitory top-down feedback. Here we use computational modelling to demonstrate that such architectural hard-wiring is not necessary. Rather, predictive coding is shown to emerge as a consequence of energy efficiency, a fundamental requirement of neural processing. When training recurrent neural networks to minimise their energy consumption while operating in predictive environments, the networks self-organise into prediction and error units with appropriate inhibitory and excitatory interconnections and learn to inhibit predictable sensory input. We demonstrate that prediction units can reliably be identified through biases in their median preactivation, pointing towards a fundamental property of prediction units in the predictive coding framework. Moving beyond the view of purely top-down driven predictions, we demonstrate via virtual lesioning experiments that networks perform predictions on two timescales: fast lateral predictions among sensory units and slower prediction cycles that integrate evidence over time. Our results, which replicate across two separate data sets, suggest that predictive coding can be interpreted as a natural consequence of energy efficiency. More generally, they raise the question which other computational principles of brain function can be understood as a result of physical constraints posed by the brain, opening up a new area of bio-inspired, machine learning-powered neuroscience research.

Reflex circuits in the nervous system integrate changes in the environment with physiology. Compact clusters of brain neuron cell bodies, termed nuclei, are essential for receiving sensory input and for transmitting motor outputs to the body. These nucelii are critical relay stations which process incoming information and convert these signals to outgoing action potentials which regulate immune system functions. Thus, reflex neural circuits maintain parameters of immunological physiology within a narrow range optimal for health. Advances in neuroscience and immunology using optogenetics, pharmacogenetics, and functional mapping offer a new understanding of the importance of neural circuitry underlying immunity, and offer direct paths to new therapies.

Neural responses are variable: even under identical experimental conditions, single neuron and population responses typically differ from trial to trial and across time. Recent work has demonstrated that this variability has predictable structure, can be modulated by sensory input and behaviour, and bears critical signatures of the underlying network dynamics and computations. However, current methods for characterising neural variability are primarily geared towards sensory coding in the laboratory: they require trials with repeatable experimental stimuli and behavioural covariates. In addition, they make strong assumptions about the parametric form of variability, rely on assumption-free but data-inefficient histogram-based approaches, or are altogether ill-suited for capturing variability modulation by covariates. Here we present a universal probabilistic spike count model that eliminates these shortcomings. Our method uses scalable Bayesian machine learning techniques to model arbitrary spike count distributions (SCDs) with flexible dependence on observed as well as latent covariates. Without requiring repeatable trials, it can flexibly capture covariate-dependent joint SCDs, and provide interpretable latent causes underlying the statistical dependencies between neurons. We apply the model to recordings from a canonical non-sensory neural population: head direction cells in the mouse. We find that variability in these cells defies a simple parametric relationship with mean spike count as assumed in standard models, its modulation by external covariates can be comparably strong to that of the mean firing rate, and slow low-dimensional latent factors explain away neural correlations. Our approach paves the way to understanding the mechanisms and computations underlying neural variability under naturalistic conditions, beyond the realm of sensory coding with repeatable stimuli.

Many models of attention assume that attentional selection takes place at a specific moment in time which demarcates the critical transition from pre-attentive to attentive processing of sensory input. We argue that this intuitively appealing account is not only inaccurate, but has led to substantial conceptual confusion (to the point where some attention researchers offer to abandon the term ‘attention’ altogether). As an alternative, we offer a “diachronic” framework that describes attentional selectivity as a process that unfolds over time. Key to this view is the concept of attentional episodes, brief periods of intense attentional amplification of sensory representations that regulate access to working memory and response-related processes. We describe how attentional episodes are linked to earlier attentional mechanisms and to recurrent processing at the neural level. We present data showing that multiple sequential events can be involuntarily encoded in working memory when they appear during the same attentional episode, whether they are relevant or not. We also discuss the costs associated with processing multiple events within a single episode. Finally, we argue that breaking down the dichotomy between pre-attentive and attentive (as well as early vs. late selection) offers new solutions to old problems in attention research that have never been resolved. It can provide a unified and conceptually coherent account of the network of cognitive and neural processes that produce the goal-directed selectivity in perceptual processing that is commonly referred to as “attention”.

A recent formulation of predictive coding theory proposes that a subset of neurons in each cortical area encodes sensory prediction errors, the difference between predictions relayed from higher cortex and the sensory input. Here, we test for evidence of prediction error responses in spiking responses and local field potentials (LFP) recorded in primary visual cortex and area V4 of macaque monkeys, and in complementary electroencephalographic (EEG) scalp recordings in human participants. We presented a fixed sequence of visual stimuli on most trials, and violated the expected ordering on a small subset of trials. Under predictive coding theory, pattern-violating stimuli should trigger robust prediction errors, but we found that spiking, LFP and EEG responses to expected and pattern-violating stimuli were nearly identical. Our results challenge the assertion that a fundamental computational motif in sensory cortex is to signal prediction errors, at least those based on predictions derived from temporal patterns of visual stimulation.

Perception of object motion is fundamentally multisensory, yet little is known about similarities and differences in the computations that give rise to our experience across senses. Insight can be provided by examining auditory motion processing in early blind individuals. In those who become blind early in life, the ‘visual’ motion area hMT+ responds to auditory motion. Meanwhile, the planum temporale, associated with auditory motion in sighted individuals, shows reduced selectivity for auditory motion, suggesting competition between cortical areas for functional role. According to the metamodal hypothesis of cross-modal plasticity developed by Pascual-Leone, the recruitment of hMT+ is driven by it being a metamodal structure containing “operators that execute a given function or computation regardless of sensory input modality”. Thus, the metamodal hypothesis predicts that the computations underlying auditory motion processing in early blind individuals should be analogous to visual motion processing in sighted individuals - relying on non-separable spatiotemporal filters. Inconsistent with the metamodal hypothesis, evidence suggests that the computational algorithms underlying auditory motion processing in early blind individuals fail to undergo a qualitative shift as a result of cross-modal plasticity. Auditory motion filters, in both blind and sighted subjects, are separable in space and time, suggesting that the recruitment of hMT+ to extract motion information from auditory input includes a significant modification of its normal computational operations.

In this talk I outline the technical challenges and current solutions to layer fMRI. Specifically, I describe our acquisition strategies for maximizing resolution, spatial coverage, time efficiency as well as, perhaps most importantly, vascular specificity. Novel applications from our group, including mapping feedforward and feedback connections to M1 during task and sensory input modulation and S1 during a sensory prediction task are be shown. Layer specific activity in dorsal lateral prefrontal cortex during a working memory task is also demonstrated. Additionally, I’ll show preliminary work on mapping whole brain layer-specific resting state connectivity and hierarchy.

The Stabilized Supralinear Network is a model of recurrently connected excitatory (E) and inhibitory (I) neurons with a supralinear input-output relation. It can explain cortical computations such as response normalization and inhibitory stabilization. However, the network's connectivity is designed by hand, based on experimental measurements. How the recurrent synaptic weights can be learned from the sensory input statistics in a biologically plausible way is unknown. Earlier theoretical work on plasticity focused on single neurons and the balance of excitation and inhibition but did not consider the simultaneous plasticity of recurrent synapses and the formation of receptive fields. Here we present a recurrent E-I network model where all synaptic connections are simultaneously plastic, and E neurons self-stabilize by recruiting co-tuned inhibition. Motivated by experimental results, we employ a local Hebbian plasticity rule with multiplicative normalization for E and I synapses. We develop a theoretical framework that explains how plasticity enables inhibition balanced excitatory receptive fields that match experimental results. We show analytically that sufficiently strong inhibition allows neurons' receptive fields to decorrelate and distribute themselves across the stimulus space. For strong recurrent excitation, the network becomes stabilized by inhibition, which prevents unconstrained self-excitation. In this regime, external inputs integrate sublinearly. As in the Stabilized Supralinear Network, this results in response normalization and winner-takes-all dynamics: when two competing stimuli are presented, the network response is dominated by the stronger stimulus while the weaker stimulus is suppressed. In summary, we present a biologically plausible theoretical framework to model plasticity in fully plastic recurrent E-I networks. While the connectivity is derived from the sensory input statistics, the circuit performs meaningful computations. Our work provides a mathematical framework of plasticity in recurrent networks, which has previously only been studied numerically and can serve as the basis for a new generation of brain-inspired unsupervised machine learning algorithms.

Sensory experience and perceptual learning changes receptive field properties of cortical pyramidal neurons possibly mediated by long-term potentiation (LTP) of synapses. We have previously shown in the mouse somatosensory cortex (S1) that sensory-driven LTP in layer (L) 2/3 pyramidal neurons is dependent on higher order thalamic feedback from the posteromedial nucleus (POm), which is thought to convey contextual information from various cortical regions integrated with sensory input. We have followed up on this work by dissecting the cortical microcircuitry that underlies this form of LTP. We found that repeated pairing of Pom thalamocortical and intracortical pathway activity in brain slices induces NMDAr-dependent LTP of the L2/3 synapses that are driven by the intracortical pathway. Repeated pairing also recruits activity of vasoactive intestinal peptide (VIP) interneurons, whereas it reduces the activity of somatostatin (SST) interneurons. VIP interneuron-mediated inhibition of SST interneurons has been established as a motif for the disinhibition of pyramidal neurons. By chemogenetic interrogation we found that activation of this disinhibitory microcircuit motif by higher-order thalamic feedback is indispensable for eliciting LTP. Preliminary results in vivo suggest that VIP neuron activity also increases during sensory-evoked LTP. Together, this suggests that the higherorder thalamocortical feedback may help modifying the strength of synaptic circuits that process first-order sensory information in S1. To start characterizing the relationship between higher-order feedback and cortical plasticity during learning in vivo, we adapted a perceptual learning paradigm in which head-fixed mice have to discriminate two types of textures in order to obtain a reward. POm axons or L2/3 pyramidal neurons labeled with the genetically encoded calcium indicator GCaMP6s were imaged during the acquisition of this task as well as the subsequent learning of a new discrimination rule. We found that a subpopulation of the POm axons and L2/3 neurons dynamically represent textures. Moreover, upon a change in reward contingencies, a fraction of the L2/3 neurons re-tune their selectivity to the texture that is newly associated with the reward. Altogether, our data indicates that higher-order thalamic feedback can facilitate synaptic plasticity and may be implicated in dynamic sensory stimulus representations in S1, which depends on higher-order features that are associated with the stimuli.

Regulation of information flow in the brain is critical for many forms of behavior. In the process of sensory based decision-making, decisions about future actions are held in memory until enacted, making them potentially vulnerable to distracting sensory input. Therefore, gating of information flow from sensory to motor areas could protect memory from interference during decision-making, but the underlying network mechanisms are not understood. I will present our recent experimental and modeling work describing how information flow from the sensory cortex can be gated by state-dependent frontal cortex dynamics during decision-making in mice. Our results show that communication between brain regions can be regulated via attractor dynamics, which control the degree of commitment to an action, and reveal a novel mechanism of gating of neural information.

Sequential temporal ordering and patterning are key features of natural signals, used by the brain to decode stimuli and perceive them as sensory objects. Touch is one sensory modality where temporal patterning carries key information, and the rodent whisker system is a prominent model for understanding neuronal coding and plasticity underlying touch sensation. Neurons in this system are precise encoders of fluctuations in whisker dynamics down to a timescale of milliseconds, but it is not clear whether they can refine their encoding abilities as a result of learning patterned stimuli. For example, can they enhance temporal integration to become better at distinguishing sequences? To explore how cortical coding plasticity underpins sequence discrimination, we developed a task in which mice distinguished between tactile ‘word’ sequences constructed from distinct vibrations delivered to the whiskers, assembled in different orders. Animals licked to report the presence of the target sequence. Optogenetic inactivation showed that the somatosensory cortex was necessary for sequence discrimination. Two-photon imaging in layer 2/3 of the primary somatosensory “barrel” cortex (S1bf) revealed that, in well-trained animals, neurons had heterogeneous selectivity to multiple task variables including not just sensory input but also the animal’s action decision and the trial outcome (presence or absence of the predicted reward). Many neurons were activated preceding goal-directed licking, thus reflecting the animal’s learnt action in response to the target sequence; these neurons were found as soon as mice learned to associate the rewarded sequence with licking. In contrast, learning evoked smaller changes in sensory response tuning: neurons responding to stimulus features were already found in naïve mice, and training did not generate neurons with enhanced temporal integration or categorical responses. Therefore, in S1bf sequence learning results in neurons whose activity reflects the learnt association between target sequence and licking, rather than a refined representation of sensory features. Taken together with results from other laboratories, our findings suggest that neurons in sensory cortex are involved in task-specific processing and that an animal does not sense the world independently of what it needs to feel in order to guide behaviour.

Exposure to proper environment during early development is essential for brain maturation. Impaired sensory input or abnormal experiences can have long-term negative consequences on brain health. We seek to define the precise synaptic aberrations caused by abnormal visual experiences early in life, and how these can be remedied through viral, genetic and environmental approaches. Resulting knowledge will contribute to the development of new approaches to mitigate nervous system damage caused by abnormal early life experience.

My research at the interface of neurobiology, biomechanics, and behavior seeks to understand how the timing precision of sensory input structures locomotor output. My lab studies the flight behavior of the fruit fly, Drosophila melanogaster, combining powerful genetic tools available for labeling and manipulating neural circuits with cutting-edge imaging in awake, behaving animals. This work has the potential to fundamentally reshape understanding of the evolution of insect flight, as well as highlight the tremendous importance of timing in the context of locomotion. Timing is crucial to the nervous system. The ability to rapidly detect and process subtle disturbances in the environment determines whether an animal can attain its next meal or successfully navigate complex, unpredictable terrain. While previous work on various animals has made tremendous strides uncovering the specialized neural circuits used to resolve timing differences with sub-microsecond resolution, it has focused on the detection of timing differences in sensory systems. Understanding of how the timing of motor output is structured by precise sensory input remains poor. My research focuses on an organ unique to fruit flies, called the haltere, that serves as a bridge for detecting and acting on subtle timing differences, helping flies execute rapid maneuvers. Understanding how this relatively simple insect canperform such impressive aerial feats demands an integrative approach that combines physics, muscle mechanics, neuroscience, and behavior. This unique, powerful approach will reveal the general principles that govern sensorimotor processing.