Neural tuning and representational geometry

Irrational choices via a curvilinear representational geometry for value

We make decisions by comparing values, but it is not yet clear how value is represented in the brain. Many models assume, if only implicitly, that the representational geometry of value is linear. However, in part due to a historical focus on noisy single neurons, rather than neuronal populations, this hypothesis has not been rigorously tested. Here, we examine the representational geometry of value in the ventromedial prefrontal cortex (vmPFC), a part of the brain linked to economic decision-making, in two male rhesus macaques. We find that values are encoded along a curved manifold in vmPFC. This curvilinear geometry predicts a specific pattern of irrational decision-making: that decision-makers will make worse choices when an irrelevant, decoy option is worse in value, compared to when it is better. We observe this type of irrational choices in behavior. Together, these results not only suggest that the representational geometry of value is nonlinear, but that this nonlinearity could impose bounds on rational decision-making.

Quality space computations for consciousness

The quality space hypothesis about conscious experience proposes that conscious sensory states are experienced in relation to other possible sensory states. For instance, the colour red is experienced as being more like orange, and less like green or blue. Recent empirical findings suggest that subjective similarity space can be explained in terms of similarities in neural activation patterns. Here, we consider how localist, workspace, and higher-order theories of consciousness can accommodate claims about the qualitative character of experience and functionally support a quality space. We review existing empirical evidence for each of these positions, and highlight novel experimental tools, such as altering local activation spaces via brain stimulation or behavioural training, that can distinguish these accounts.

Coordinated Response Modulations Enable Flexible Use of Visual Information

We use sensory information in remarkably flexible ways. We can generalize by ignoring task-irrelevant features, report different features of a stimulus, and use different actions to report a perceptual judgment. These forms of flexible behavior are associated with small modulations of the responses of sensory neurons. While the existence of these response modulations is indisputable, efforts to understand their function have been largely relegated to theory, where they have been posited to change information coding or enable downstream neurons to read out different visual and cognitive information using flexible weights. Here, we tested these ideas using a rich, flexible behavioral paradigm, multi-neuron, multi-area recordings in primary visual cortex (V1) and mid-level visual area V4. We discovered that those response modulations in V4 (but not V1) contain the ingredients necessary to enable flexible behavior, but not via those previously hypothesized mechanisms. Instead, we demonstrated that these response modulations are precisely coordinated across the population such that downstream neurons have ready access to the correct information to flexibly guide behavior without making changes to information coding or synapses. Our results suggest a novel computational role for task-dependent response modulations: they enable flexible behavior by changing the information that gets out of a sensory area, not by changing information coding within it.

Computational role of structure in neural activity and connectivity

One major challenge of neuroscience is identifying structure in seemingly disorganized neural activity. Different types of structure have different computational implications that can help neuroscientists understand the functional role of a particular brain area. Here, we outline a unified approach to characterize structure by inspecting the representational geometry and the modularity properties of the recorded activity and show that a similar approach can also reveal structure in connectivity. We start by setting up a general framework for determining geometry and modularity in activity and connectivity and relating these properties with computations performed by the network. We then use this framework to review the types of structure found in recent studies of model networks performing three classes of computations.

The interplay between spatial and non-spatial grouping cues over approximate number perception

Humans and animals share the cognitive ability to quickly extract approximate number information from sets. Main psychophysical models suggest that visual approximate numerosity relies on segmented units, which can be affected by Gestalt rules. Indeed, arrays containing spatial grouping cues, such as connectedness, closure, and even symmetry, are underestimated compared to ungrouped arrays with equal low-level features. Recent evidence suggests that non-spatial cues, such as color-similarity, also trigger numerosity underestimation. However, in natural vision, several grouping cues may coexist in the scene. Notably, conjunction of grouping cues (color and closure) reduces perceived numerosity following an additive rule. To test whether the conjunction-effect holds for other Gestalt cues, we investigated the effect of connectedness and symmetry over numerosity perception both in isolation and, critically, in conjunction with luminance similarity. Participants performed a comparison-task between a reference and a test stimulus varying in numerosity. In Experiment 1, test stimuli contained two isolated groupings (connectedness or luminance), a conjunction (connectedness and luminance), and a neutral condition (no groupings). Results show that point of subjective equality was higher in both isolated grouping conditions compared to the neutral condition. Furthermore, in the conjunction condition, the biases from isolated grouping cues added linearly, resulting in a numerosity underestimation equal to the sum of the isolated biases. In Experiment 2 we found that conjunction of symmetry and luminance followed the same additive rule. These findings strongly suggest that both spatial and non-spatial isolated cues affect numerosity perception. Crucially, we show that their conjunction effect extends to symmetry and connectedness.

Representation and computation in visual working memory

The ability to sustain internal representations of the sensory environment beyond immediate perception is a fundamental requirement of cognitive processing. In recent years, debates regarding the capacity and fidelity of the working memory (WM) system have advanced our understanding of the nature of these representations. In particular, there is growing recognition that WM representations are not merely imperfect copies of a perceived object or event. New experimental tools have revealed that observers possess richer information about the uncertainty in their memories and take advantage of environmental regularities to use limited memory resources optimally. Meanwhile, computational models of visuospatial WM formulated at different levels of implementation have converged on common principles relating capacity to variability and uncertainty. Here we review recent research on human WM from a computational perspective, including the neural mechanisms that support it.

A dynamic neural resource model bridges sensory and working memory

Probing memory of a complex visual image within a few hundred milliseconds after its disappearance reveals significantly greater fidelity of recall than if the probe is delayed by as little as a second. Classically interpreted, the former taps into a detailed but rapidly decaying visual sensory or 'iconic' memory (IM), while the latter relies on capacity-limited but comparatively stable visual working memory (VWM). While iconic decay and VWM capacity have been extensively studied independently, currently no single framework quantitatively accounts for the dynamics of memory fidelity over these time scales. Here, we extend a stationary neural population model of VWM with a temporal dimension, incorporating rapid sensory-driven accumulation of activity encoding each visual feature in memory, and a slower accumulation of internal error that causes memorized features to randomly drift over time. Instead of facilitating read-out from an independent sensory store, an early cue benefits recall by lifting the effective limit on VWM signal strength imposed when multiple items compete for representation, allowing memory for the cued item to be supplemented with information from the decaying sensory trace. Empirical measurements of human recall dynamics validate these predictions while excluding alternative model architectures. A key conclusion is that differences in capacity classically thought to distinguish IM and VWM are in fact contingent upon a single resource-limited WM store.

Representational structures as a unifying framework for attention

Our visual system consciously processes only a subset of the incoming information. Selective attention allows us to prioritize relevant inputs, and can be allocated to features, locations, and objects. Recent advances in feature-based attention suggest that several selection principles are shared across these domains and that many differences between the effects of attention on perceptual processing can be explained by differences in the underlying representational structures. Moving forward, it can thus be useful to assess how attention changes the structure of the representational spaces over which it operates, which include the spatial organization, feature maps, and object-based coding in visual cortex. This will ultimately add to our understanding of how attention changes the flow of visual information processing more broadly.

The representational geometry of emotional states in basolateral amygdala

Sensory stimuli associated with aversive outcomes cause multiple behavioral responses related to an animal's evolving emotional state, but neural mechanisms underlying these processes remain unclear. Here aversive stimuli were presented to mice, eliciting two responses reflecting fear and flight to safety: tremble and ingress into a virtual burrow. Inactivation of basolateral amygdala (BLA) eliminated differential responses to aversive and neutral stimuli without eliminating responses themselves, suggesting BLA signals valence, not motor commands. However, two-photon imaging revealed that neurons typically exhibited mixed selectivity for stimulus identity, valence, tremble and/or ingress. Despite heterogeneous selectivity, BLA representational geometry was lower-dimensional when encoding valence, tremble and safety, enabling generalization of emotions across conditions. Further, tremble and valence coding directions were orthogonal, allowing linear readouts to specialize. Thus BLA representational geometry confers two computational properties that identify specialized neural circuits encoding variables describing emotional states: generalization across conditions, and readouts lacking interference from other readouts.

Topological structure of population activity in mouse visual cortex encodes densely sampled stimulus rotations

The primary visual cortex is one of the most well understood regions supporting the processing involved in sensory computation. Following the popularization of high-density neural recordings, it has been observed that the activity of large neural populations is often constrained to low dimensional manifolds. In this work, we quantify the structure of such neural manifolds in the visual cortex. We do this by analyzing publicly available two-photon optical recordings of mouse primary visual cortex in response to visual stimuli with a densely sampled rotation angle. Using a geodesic metric along with persistent homology, we discover that population activity in response to such stimuli generates a circular manifold, encoding the angle of rotation. Furthermore, we observe that this circular manifold is expressed differently in subpopulations of neurons with differing orientation and direction selectivity. Finally, we discuss some of the obstacles to reliably retrieving the truthful topology generated by a neural population.

Representational formats of human memory traces

Neural representations are internal brain states that constitute the brain's model of the external world or some of its features. In the presence of sensory input, a representation may reflect various properties of this input. When perceptual information is no longer available, the brain can still activate representations of previously experienced episodes due to the formation of memory traces. In this review, we aim at characterizing the nature of neural memory representations and how they can be assessed with cognitive neuroscience methods, mainly focusing on neuroimaging. We discuss how multivariate analysis techniques such as representational similarity analysis (RSA) and deep neural networks (DNNs) can be leveraged to gain insights into the structure of neural representations and their different representational formats. We provide several examples of recent studies which demonstrate that we are able to not only measure memory representations using RSA but are also able to investigate their multiple formats using DNNs. We demonstrate that in addition to slow generalization during consolidation, memory representations are subject to semantization already during short-term memory, by revealing a shift from visual to semantic format. In addition to perceptual and conceptual formats, we describe the impact of affective evaluations as an additional dimension of episodic memories. Overall, these studies illustrate how the analysis of neural representations may help us gain a deeper understanding of the nature of human memory.

Response sub-additivity and variability quenching in visual cortex

Sub-additivity and variability are ubiquitous response motifs in the primary visual cortex (V1). Response sub-additivity enables the construction of useful interpretations of the visual environment, whereas response variability indicates the factors that limit the precision with which the brain can do this. There is increasing evidence that experimental manipulations that elicit response sub-additivity often also quench response variability. Here, we provide an overview of these phenomena and suggest that they may have common origins. We discuss empirical findings and recent model-based insights into the functional operations, computational objectives and circuit mechanisms underlying V1 activity. These different modelling approaches all predict that response sub-additivity and variability quenching often co-occur. The phenomenology of these two response motifs, as well as many of the insights obtained about them in V1, generalize to other cortical areas. Thus, the connection between response sub-additivity and variability quenching may be a canonical motif across the cortex.

Representational maps in the brain: concepts, approaches, and applications

Neural systems have evolved to process sensory stimuli in a way that allows for efficient and adaptive behavior in a complex environment. Recent technological advances enable us to investigate sensory processing in animal models by simultaneously recording the activity of large populations of neurons with single-cell resolution, yielding high-dimensional datasets. In this review, we discuss concepts and approaches for assessing the population-level representation of sensory stimuli in the form of a representational map. In such a map, not only are the identities of stimuli distinctly represented, but their relational similarity is also mapped onto the space of neuronal activity. We highlight example studies in which the structure of representational maps in the brain are estimated from recordings in humans as well as animals and compare their methodological approaches. Finally, we integrate these aspects and provide an outlook for how the concept of representational maps could be applied to various fields in basic and clinical neuroscience.

Efficient coding of natural images in the mouse visual cortex

How the activity of neurons gives rise to natural vision remains a matter of intense investigation. The mid-level visual areas along the ventral stream are selective to a common class of natural images-textures-but a circuit-level understanding of this selectivity and its link to perception remains unclear. We addressed these questions in mice, first showing that they can perceptually discriminate between textures and statistically simpler spectrally matched stimuli, and between texture types. Then, at the neural level, we found that the secondary visual area (LM) exhibited a higher degree of selectivity for textures compared to the primary visual area (V1). Furthermore, textures were represented in distinct neural activity subspaces whose relative distances were found to correlate with the statistical similarity of the images and the mice's ability to discriminate between them. Notably, these dependencies were more pronounced in LM, where the texture-related subspaces were smaller than in V1, resulting in superior stimulus decoding capabilities. Together, our results demonstrate texture vision in mice, finding a linking framework between stimulus statistics, neural representations, and perceptual sensitivity-a distinct hallmark of efficient coding computations.

Learning attentional templates for value-based decision-making

Attention filters sensory inputs to enhance task-relevant information. It is guided by an "attentional template" that represents the stimulus features that are currently relevant. To understand how the brain learns and uses templates, we trained monkeys to perform a visual search task that required them to repeatedly learn new attentional templates. Neural recordings found that templates were represented across the prefrontal and parietal cortex in a structured manner, such that perceptually neighboring templates had similar neural representations. When the task changed, a new attentional template was learned by incrementally shifting the template toward rewarded features. Finally, we found that attentional templates transformed stimulus features into a common value representation that allowed the same decision-making mechanisms to deploy attention, regardless of the identity of the template. Altogether, our results provide insight into the neural mechanisms by which the brain learns to control attention and how attention can be flexibly deployed across tasks.

Neuronal travelling waves explain rotational dynamics in experimental datasets and modelling

Spatiotemporal properties of neuronal population activity in cortical motor areas have been subjects of experimental and theoretical investigations, generating numerous interpretations regarding mechanisms for preparing and executing limb movements. Two competing models, representational and dynamical, strive to explain the relationship between movement parameters and neuronal activity. A dynamical model uses the jPCA method that holistically characterizes oscillatory activity in neuron populations by maximizing the data rotational dynamics. Different rotational dynamics interpretations revealed by the jPCA approach have been proposed. Yet, the nature of such dynamics remains poorly understood. We comprehensively analyzed several neuronal-population datasets and found rotational dynamics consistently accounted for by a traveling wave pattern. For quantifying rotation strength, we developed a complex-valued measure, the gyration number. Additionally, we identified parameters influencing rotation extent in the data. Our findings suggest that rotational dynamics and traveling waves are typically the same phenomena, so reevaluation of the previous interpretations where they were considered separate entities is needed.

The neuronal implementation of representational geometry in primate prefrontal cortex

Modern neuroscience has seen the rise of a population-doctrine that represents cognitive variables using geometrical structures in activity space. Representational geometry does not, however, account for how individual neurons implement these representations. Leveraging the principle of sparse coding, we present a framework to dissect representational geometry into biologically interpretable components that retain links to single neurons. Applied to extracellular recordings from the primate prefrontal cortex in a working memory task with interference, the identified components revealed disentangled and sequential memory representations including the recovery of memory content after distraction, signals hidden to conventional analyses. Each component was contributed by small subpopulations of neurons with distinct spiking properties and response dynamics. Modeling showed that such sparse implementations are supported by recurrently connected circuits as in prefrontal cortex. The perspective of neuronal implementation links representational geometries to their cellular constituents, providing mechanistic insights into how neural systems encode and process information.

Topological analysis of sharp-wave ripple waveforms reveals input mechanisms behind feature variations

The reactivation of experience-based neural activity patterns in the hippocampus is crucial for learning and memory. These reactivation patterns and their associated sharp-wave ripples (SWRs) are highly variable. However, this variability is missed by commonly used spectral methods. Here, we use topological and dimensionality reduction techniques to analyze the waveform of ripples recorded at the pyramidal layer of CA1. We show that SWR waveforms distribute along a continuum in a low-dimensional space, which conveys information about the underlying layer-specific synaptic inputs. A decoder trained in this space successfully links individual ripples with their expected sinks and sources, demonstrating how physiological mechanisms shape SWR variability. Furthermore, we found that SWR waveforms segregated differently during wakefulness and sleep before and after a series of cognitive tasks, with striking effects of novelty and learning. Our results thus highlight how the topological analysis of ripple waveforms enables a deeper physiological understanding of SWRs.

Signatures of task learning in neural representations

While neural plasticity has long been studied as the basis of learning, the growth of large-scale neural recording techniques provides a unique opportunity to study how learning-induced activity changes are coordinated across neurons within the same circuit. These distributed changes can be understood through an evolution of the geometry of neural manifolds and latent dynamics underlying new computations. In parallel, studies of multi-task and continual learning in artificial neural networks hint at a tradeoff between non-interference and compositionality as guiding principles to understand how neural circuits flexibly support multiple behaviors. In this review, we highlight recent findings from both biological and artificial circuits that together form a new framework for understanding task learning at the population level.

Information content in continuous attractor neural networks is preserved in the presence of moderate disordered background connectivity

Continuous attractor neural networks (CANN) form an appealing conceptual model for the storage of information in the brain. However a drawback of CANN is that they require finely tuned interactions. We here study the effect of quenched noise in the interactions on the coding of positional information within CANN. Using the replica method we compute the Fisher information for a network with position-dependent input and recurrent connections composed of a short-range (in space) and a disordered component. We find that the loss in positional information is small for not too large disorder strength, indicating that CANN have a regime in which the advantageous effects of local connectivity on information storage outweigh the detrimental ones. Furthermore, a substantial part of this information can be extracted with a simple linear readout.

Circular and unified analysis in network neuroscience

Genuinely new discovery transcends existing knowledge. Despite this, many analyses in systems neuroscience neglect to test new speculative hypotheses against benchmark empirical facts. Some of these analyses inadvertently use circular reasoning to present existing knowledge as new discovery. Here, I discuss that this problem can confound key results and estimate that it has affected more than three thousand studies in network neuroscience over the last decade. I suggest that future studies can reduce this problem by limiting the use of speculative evidence, integrating existing knowledge into benchmark models, and rigorously testing proposed discoveries against these models. I conclude with a summary of practical challenges and recommendations.

Early selection of task-relevant features through population gating

Brains can gracefully weed out irrelevant stimuli to guide behavior. This feat is believed to rely on a progressive selection of task-relevant stimuli across the cortical hierarchy, but the specific across-area interactions enabling stimulus selection are still unclear. Here, we propose that population gating, occurring within primary auditory cortex (A1) but controlled by top-down inputs from prelimbic region of medial prefrontal cortex (mPFC), can support across-area stimulus selection. Examining single-unit activity recorded while rats performed an auditory context-dependent task, we found that A1 encoded relevant and irrelevant stimuli along a common dimension of its neural space. Yet, the relevant stimulus encoding was enhanced along an extra dimension. In turn, mPFC encoded only the stimulus relevant to the ongoing context. To identify candidate mechanisms for stimulus selection within A1, we reverse-engineered low-rank RNNs trained on a similar task. Our analyses predicted that two context-modulated neural populations gated their preferred stimulus in opposite contexts, which we confirmed in further analyses of A1. Finally, we show in a two-region RNN how population gating within A1 could be controlled by top-down inputs from PFC, enabling flexible across-area communication despite fixed inter-areal connectivity.

Performing a vibrotactile discrimination task modulates finger representations in primary somatosensory cortex

It is well established that vibrotactile stimuli are represented in somatotopic maps. However, less is known about whether these somatotopic representations are modulated by task demands and maybe even in the absence of tactile input. Here, we used a vibrotactile discrimination task as a tool to investigate these questions in further detail. Participants were required to actively perceive and process tactile stimuli in comparison to a no-task control condition where identical stimuli were passively perceived (no-memory condition). Importantly, both vibrotactile stimuli were either applied to the right index or little finger, allowing us to investigate whether cognitive task demands shape finger representations in primary somatosensory cortex (S1). Using multivoxel pattern analysis and representational similarity analysis, we found that S1 finger representations were more distinct during the memory than the no-memory condition. Interestingly, this effect was not only observed while tactile stimuli were presented but also during the delay period (i.e., in the absence of tactile stimulation). Our findings imply that when individuals are required to focus on tactile stimuli, retain them in their memory, and engage in active processing of distinctive stimulus features, this exerts a modulatory effect on the finger representations present in S1.NEW & NOTEWORTHY Using multivoxel pattern analysis, we found that discrimination task demands shape finger representations in the contralateral primary somatosensory cortex (S1), and that somatotopic representations are modulated by task demands not only during tactile stimulation but also to a certain extent in the absence of tactile input.

A quantitative model of ensemble perception as summed activation in feature space

Ensemble perception is a process by which we summarize complex scenes. Despite the importance of ensemble perception to everyday cognition, there are few computational models that provide a formal account of this process. Here we develop and test a model in which ensemble representations reflect the global sum of activation signals across all individual items. We leverage this set of minimal assumptions to formally connect a model of memory for individual items to ensembles. We compare our ensemble model against a set of alternative models in five experiments. Our approach uses performance on a visual memory task for individual items to generate zero-free-parameter predictions of interindividual and intraindividual differences in performance on an ensemble continuous-report task. Our top-down modelling approach formally unifies models of memory for individual items and ensembles and opens a venue for building and comparing models of distinct memory processes and representations.

Probabilistic Learning and Psychological Similarity

The notions of psychological similarity and probabilistic learning are key posits in cognitive, computational, and developmental psychology and in machine learning. However, their explanatory relationship is rarely made explicit within and across these research fields. This opinionated review critically evaluates how these notions can mutually inform each other within computational cognitive science. Using probabilistic models of concept learning as a case study, I argue that two notions of psychological similarity offer important normative constraints to guide modelers' interpretations of representational primitives. In particular, the two notions furnish probabilistic models of cognition with meaningful interpretations of what the associated subjective probabilities in the model represent and how they attach to experiences from which the agent learns. Similarity representations thereby provide probabilistic models with cognitive, as opposed to purely mathematical, content.

Common population codes produce extremely nonlinear neural manifolds

Populations of neurons represent sensory, motor, and cognitive variables via patterns of activity distributed across the population. The size of the population used to encode a variable is typically much greater than the dimension of the variable itself, and thus, the corresponding neural population activity occupies lower-dimensional subsets of the full set of possible activity states. Given population activity data with such lower-dimensional structure, a fundamental question asks how close the low-dimensional data lie to a linear subspace. The linearity or nonlinearity of the low-dimensional structure reflects important computational features of the encoding, such as robustness and generalizability. Moreover, identifying such linear structure underlies common data analysis methods such as Principal Component Analysis (PCA). Here, we show that for data drawn from many common population codes the resulting point clouds and manifolds are exceedingly nonlinear, with the dimension of the best-fitting linear subspace growing at least exponentially with the true dimension of the data. Consequently, linear methods like PCA fail dramatically at identifying the true underlying structure, even in the limit of arbitrarily many data points and no noise.

All-optical recreation of naturalistic neural activity with a multifunctional transgenic reporter mouse

Determining which features of the neural code drive behavior requires the ability to simultaneously read out and write in neural activity patterns with high precision across many neurons. All-optical systems that combine two-photon calcium imaging and targeted photostimulation enable the activation of specific, functionally defined groups of neurons. However, these techniques are unable to test how patterns of activity across a population contribute to computation because of an inability to both read and write cell-specific firing rates. To overcome this challenge, we make two advances: first, we introduce a genetic line of mice for Cre-dependent co-expression of a calcium indicator and a potent soma-targeted microbial opsin. Second, using this line, we develop a method for read-out and write-in of precise population vectors of neural activity by calibrating the photostimulation to each cell. These advances offer a powerful and convenient platform for investigating the neural codes of computation and behavior.

Statistical inference on representational geometries

Neuroscience has recently made much progress, expanding the complexity of both neural activity measurements and brain-computational models. However, we lack robust methods for connecting theory and experiment by evaluating our new big models with our new big data. Here, we introduce new inference methods enabling researchers to evaluate and compare models based on the accuracy of their predictions of representational geometries: A good model should accurately predict the distances among the neural population representations (e.g. of a set of stimuli). Our inference methods combine novel 2-factor extensions of crossvalidation (to prevent overfitting to either subjects or conditions from inflating our estimates of model accuracy) and bootstrapping (to enable inferential model comparison with simultaneous generalization to both new subjects and new conditions). We validate the inference methods on data where the ground-truth model is known, by simulating data with deep neural networks and by resampling of calcium-imaging and functional MRI data. Results demonstrate that the methods are valid and conclusions generalize correctly. These data analysis methods are available in an open-source Python toolbox (rsatoolbox.readthedocs.io).

Distinct ventral stream and prefrontal cortex representational dynamics during sustained conscious visual perception

Instances of sustained stationary sensory input are ubiquitous. However, previous work focused almost exclusively on transient onset responses. This presents a critical challenge for neural theories of consciousness, which should account for the full temporal extent of experience. To address this question, we use intracranial recordings from ten human patients with epilepsy to view diverse images of multiple durations. We reveal that, in sensory regions, despite dramatic changes in activation magnitude, the distributed representation of categories and exemplars remains sustained and stable. In contrast, in frontoparietal regions, we find transient content representation at stimulus onset. Our results highlight the connection between the anatomical and temporal correlates of experience. To the extent perception is sustained, it may rely on sensory representations and to the extent perception is discrete, centered on perceptual updating, it may rely on frontoparietal representations.

The dynamics and geometry of choice in premotor cortex

The brain represents sensory variables in the coordinated activity of neural populations, in which tuning curves of single neurons define the geometry of the population code. Whether the same coding principle holds for dynamic cognitive variables remains unknown because internal cognitive processes unfold with a unique time course on single trials observed only in the irregular spiking of heterogeneous neural populations. Here we show the existence of such a population code for the dynamics of choice formation in the primate premotor cortex. We developed an approach to simultaneously infer population dynamics and tuning functions of single neurons to the population state. Applied to spike data recorded during decision-making, our model revealed that populations of neurons encoded the same dynamic variable predicting choices, and heterogeneous firing rates resulted from the diverse tuning of single neurons to this decision variable. The inferred dynamics indicated an attractor mechanism for decision computation. Our results reveal a common geometric principle for neural encoding of sensory and dynamic cognitive variables.

A unifying perspective on neural manifolds and circuits for cognition

Two different perspectives have informed efforts to explain the link between the brain and behaviour. One approach seeks to identify neural circuit elements that carry out specific functions, emphasizing connectivity between neurons as a substrate for neural computations. Another approach centres on neural manifolds - low-dimensional representations of behavioural signals in neural population activity - and suggests that neural computations are realized by emergent dynamics. Although manifolds reveal an interpretable structure in heterogeneous neuronal activity, finding the corresponding structure in connectivity remains a challenge. We highlight examples in which establishing the correspondence between low-dimensional activity and connectivity has been possible, unifying the neural manifold and circuit perspectives. This relationship is conspicuous in systems in which the geometry of neural responses mirrors their spatial layout in the brain, such as the fly navigational system. Furthermore, we describe evidence that, in systems in which neural responses are heterogeneous, the circuit comprises interactions between activity patterns on the manifold via low-rank connectivity. We suggest that unifying the manifold and circuit approaches is important if we are to be able to causally test theories about the neural computations that underlie behaviour.

The neuroconnectionist research programme

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

Decoding and geometry of ten finger movements in human posterior parietal cortex and motor cortex

Objective. Enable neural control of individual prosthetic fingers for participants with upper-limb paralysis.Approach. Two tetraplegic participants were each implanted with a 96-channel array in the left posterior parietal cortex (PPC). One of the participants was additionally implanted with a 96-channel array near the hand knob of the left motor cortex (MC). Across tens of sessions, we recorded neural activity while the participants attempted to move individual fingers of the right hand. Offline, we classified attempted finger movements from neural firing rates using linear discriminant analysis with cross-validation. The participants then used the neural classifier online to control individual fingers of a brain-machine interface (BMI). Finally, we characterized the neural representational geometry during individual finger movements of both hands.Main Results. The two participants achieved 86% and 92% online accuracy during BMI control of the contralateral fingers (chance = 17%). Offline, a linear decoder achieved ten-finger decoding accuracies of 70% and 66% using respective PPC recordings and 75% using MC recordings (chance = 10%). In MC and in one PPC array, a factorized code linked corresponding finger movements of the contralateral and ipsilateral hands.Significance. This is the first study to decode both contralateral and ipsilateral finger movements from PPC. Online BMI control of contralateral fingers exceeded that of previous finger BMIs. PPC and MC signals can be used to control individual prosthetic fingers, which may contribute to a hand restoration strategy for people with tetraplegia.

A gradual transition from veridical to categorical representations along the visual hierarchy during working memory, but not perception

The ability to stably maintain visual information over brief delays is central to cognitive functioning. One possible way to achieve robust working memory maintenance is by having multiple concurrent mnemonic representations across multiple cortical loci. For example, early visual cortex might contribute to storage by representing information in a "sensory-like" format, while intraparietal sulcus uses a format transformed away from sensory driven responses. As an explicit test of mnemonic code transformations along the visual hierarchy, we quantitatively modeled the progression of veridical-to-categorical orientation representations in human participants. Participants directly viewed, or held in mind, an oriented grating pattern, and the similarity between fMRI activation patterns for different orientations was calculated throughout retinotopic cortex. During direct perception, similarity was clustered around cardinal orientations, while during working memory the obliques were represented more similarly. We modeled these similarity patterns based on the known distribution of orientation information in the natural world: The "veridical" model uses an efficient coding framework to capture hypothesized representations during visual perception. The "categorical" model assumes that different "psychological distances" between orientations result in orientation categorization relative to cardinal axes. During direct perception, the veridical model explained the data well in early visual areas, while the categorical model did worse. During working memory, the veridical model only explained some of the data, while the categorical model gradually gained explanatory power for increasingly anterior retinotopic regions. These findings suggest that directly viewed images are represented veridically, but once visual information is no longer tethered to the sensory world, there is a gradual progression to more categorical mnemonic formats along the visual hierarchy.

Hippocampal spatial view cells for memory and navigation, and their underlying connectivity in humans

Hippocampal and parahippocampal gyrus spatial view neurons in primates respond to the spatial location being looked at. The representation is allocentric, in that the responses are to locations "out there" in the world, and are relatively invariant with respect to retinal position, eye position, head direction, and the place where the individual is located. The underlying connectivity in humans is from ventromedial visual cortical regions to the parahippocampal scene area, leading to the theory that spatial view cells are formed by combinations of overlapping feature inputs self-organized based on their closeness in space. Thus, although spatial view cells represent "where" for episodic memory and navigation, they are formed by ventral visual stream feature inputs in the parahippocampal gyrus in what is the parahippocampal scene area. A second "where" driver of spatial view cells are parietal inputs, which it is proposed provide the idiothetic update for spatial view cells, used for memory recall and navigation when the spatial view details are obscured. Inferior temporal object "what" inputs and orbitofrontal cortex reward inputs connect to the human hippocampal system, and in macaques can be associated in the hippocampus with spatial view cell "where" representations to implement episodic memory. Hippocampal spatial view cells also provide a basis for navigation to a series of viewed landmarks, with the orbitofrontal cortex reward inputs to the hippocampus providing the goals for navigation, which can then be implemented by hippocampal connectivity in humans to parietal cortex regions involved in visuomotor actions in space. The presence of foveate vision and the highly developed temporal lobe for object and scene processing in primates including humans provide a basis for hippocampal spatial view cells to be key to understanding episodic memory in the primate and human hippocampus, and the roles of this system in primate including human navigation.

Multi-scale goal distance representations in human hippocampus during virtual spatial navigation

Goal-directed navigation relies on both coarse and fine-grained coding of spatial distance between the current position of a navigating subject and a goal destination. However, the neural signatures underlying goal distance coding remain poorly understood. Using intracranial EEG recordings from the hippocampus of drug-resistant epilepsy patients who performed a virtual spatial navigation task, we found that the right hippocampal theta power was significantly modulated by goal distance and decreased with goal proximity. This modulation varied along the hippocampal longitudinal axis such that theta power in the posterior hippocampus decreased more strongly with goal proximity. Similarly, neural timescale, reflecting the duration across which information can be maintained, increased gradually from the posterior to anterior hippocampus. Taken together, this study provides empirical evidence for multi-scale spatial representations of goal distance in the human hippocampus and links the hippocampal processing of spatial information to its intrinsic temporal dynamics.

Unsupervised approach to decomposing neural tuning variability

Neural representation is often described by the tuning curves of individual neurons with respect to certain stimulus variables. Despite this tradition, it has become increasingly clear that neural tuning can vary substantially in accordance with a collection of internal and external factors. A challenge we are facing is the lack of appropriate methods to accurately capture the moment-to-moment tuning variability directly from the noisy neural responses. Here we introduce an unsupervised statistical approach, Poisson functional principal component analysis (Pf-PCA), which identifies different sources of systematic tuning fluctuations, moreover encompassing several current models (e.g.,multiplicative gain models) as special cases. Applying this method to neural data recorded from macaque primary visual cortex- a paradigmatic case for which the tuning curve approach has been scientifically essential- we discovered a simple relationship governing the variability of orientation tuning, which unifies different types of gain changes proposed previously. By decomposing the neural tuning variability into interpretable components, our method enables discovery of unexpected structure of the neural code, capturing the influence of the external stimulus drive and internal states simultaneously.

Feature-based attention warps the perception of visual features

Selective attention improves sensory processing of relevant information but can also impact the quality of perception. For example, attention increases visual discrimination performance and at the same time boosts apparent stimulus contrast of attended relative to unattended stimuli. Can attention also lead to perceptual distortions of visual representations? Optimal tuning accounts of attention suggest that processing is biased towards "off-tuned" features to maximize the signal-to-noise ratio in favor of the target, especially when targets and distractors are confusable. Here, we tested whether such tuning gives rise to phenomenological changes of visual features. We instructed participants to select a color among other colors in a visual search display and subsequently asked them to judge the appearance of the target color in a 2-alternative forced choice task. Participants consistently judged the target color to appear more dissimilar from the distractor color in feature space. Critically, the magnitude of these perceptual biases varied systematically with the similarity between target and distractor colors during search, indicating that attentional tuning quickly adapts to current task demands. In control experiments we rule out possible non-attentional explanations such as color contrast or memory effects. Overall, our results demonstrate that selective attention warps the representational geometry of color space, resulting in profound perceptual changes across large swaths of feature space. Broadly, these results indicate that efficient attentional selection can come at a perceptual cost by distorting our sensory experience.

Comparing representations and computations in single neurons versus neural networks

Single-neuron-level explanations have been the gold standard in neuroscience for decades. Recently, however, neural-network-level explanations have become increasingly popular. This increase in popularity is driven by the fact that the analysis of neural networks can solve problems that cannot be addressed by analyzing neurons independently. In this opinion article, I argue that while both frameworks employ the same general logic to link physical and mental phenomena, in many cases the neural network framework provides better explanatory objects to understand representations and computations related to mental phenomena. I discuss what constitutes a mechanistic explanation in neural systems, provide examples, and conclude by highlighting a number of the challenges and considerations associated with the use of analyses of neural networks to study brain function.

The medial temporal lobe supports the quality of visual short-term memory representation

The quality of short-term memory (STM) underlies our ability to recall the exact details of a recent event, yet how the human brain enables this core cognitive function remains poorly understood. Here we use multiple experimental approaches to test the hypothesis that the quality of STM, such as its precision or fidelity, relies on the medial temporal lobe (MTL), a region commonly associated with the ability to distinguish similar information remembered in long-term memory. First, with intracranial recordings, we find that delay-period MTL activity retains item-specific STM content that is predictive of subsequent recall precision. Second, STM recall precision is associated with an increase in the strength of intrinsic MTL-to-neocortical functional connections during a brief retention interval. Finally, perturbing the MTL through electrical stimulation or surgical removal can selectively reduce STM precision. Collectively, these findings provide converging evidence that the MTL is critically involved in the quality of STM representation.

The entorhinal-DG/CA3 pathway in the medial temporal lobe retains visual working memory of a simple surface feature

Classic models consider working memory (WM) and long-term memory as distinct mental faculties that are supported by different neural mechanisms. Yet, there are significant parallels in the computation that both types of memory require. For instance, the representation of precise item-specific memory requires the separation of overlapping neural representations of similar information. This computation has been referred to as pattern separation, which can be mediated by the entorhinal-DG/CA3 pathway of the medial temporal lobe (MTL) in service of long-term episodic memory. However, although recent evidence has suggested that the MTL is involved in WM, the extent to which the entorhinal-DG/CA3 pathway supports precise item-specific WM has remained elusive. Here, we combine an established orientation WM task with high-resolution fMRI to test the hypothesis that the entorhinal-DG/CA3 pathway retains visual WM of a simple surface feature. Participants were retrospectively cued to retain one of the two studied orientation gratings during a brief delay period and then tried to reproduce the cued orientation as precisely as possible. By modeling the delay-period activity to reconstruct the retained WM content, we found that the anterior-lateral entorhinal cortex (aLEC) and the hippocampal DG/CA3 subfield both contain item-specific WM information that is associated with subsequent recall fidelity. Together, these results highlight the contribution of MTL circuitry to item-specific WM representation.

Neural Mechanisms That Make Perceptual Decisions Flexible

Neural mechanisms of perceptual decision making have been extensively studied in experimental settings that mimic stable environments with repeating stimuli, fixed rules, and payoffs. In contrast, we live in an ever-changing environment and have varying goals and behavioral demands. To accommodate variability, our brain flexibly adjusts decision-making processes depending on context. Here, we review a growing body of research that explores the neural mechanisms underlying this flexibility. We highlight diverse forms of context dependency in decision making implemented through a variety of neural computations. Context-dependent neural activity is observed in a distributed network of brain structures, including posterior parietal, sensory, motor, and subcortical regions, as well as the prefrontal areas classically implicated in cognitive control. We propose that investigating the distributed network underlying flexible decisions is key to advancing our understanding and discuss a path forward for experimental and theoretical investigations.

A rubric for human-like agents and NeuroAI

Researchers across cognitive, neuro- and computer sciences increasingly reference 'human-like' artificial intelligence and 'neuroAI'. However, the scope and use of the terms are often inconsistent. Contributed research ranges widely from mimicking behaviour, to testing machine learning methods as neurally plausible hypotheses at the cellular or functional levels, or solving engineering problems. However, it cannot be assumed nor expected that progress on one of these three goals will automatically translate to progress in others. Here, a simple rubric is proposed to clarify the scope of individual contributions, grounded in their commitments to human-like behaviour, neural plausibility or benchmark/engineering/computer science goals. This is clarified using examples of weak and strong neuroAI and human-like agents, and discussing the generative, corroborate and corrective ways in which the three dimensions interact with one another. The author maintains that future progress in artificial intelligence will need strong interactions across the disciplines, with iterative feedback loops and meticulous validity tests-leading to both known and yet-unknown advances that may span decades to come. This article is part of a discussion meeting issue 'New approaches to 3D vision'.

Temporal context and latent state inference in the hippocampal splitter signal

The hippocampus is thought to enable the encoding and retrieval of ongoing experience, the organization of that experience into structured representations like contexts, maps, and schemas, and the use of these structures to plan for the future. A central goal is to understand what the core computations supporting these functions are, and how these computations are realized in the collective action of single neurons. A potential access point into this issue is provided by 'splitter cells', hippocampal neurons that fire differentially on the overlapping segment of trajectories that differ in their past and/or future. However, the literature on splitter cells has been fragmented and confusing, owing to differences in terminology, behavioral tasks, and analysis methods across studies. In this review, we synthesize consistent findings from this literature, establish a common set of terms, and translate between single-cell and ensemble perspectives. Most importantly, we examine the combined findings through the lens of two major theoretical ideas about hippocampal function: representation of temporal context and latent state inference. We find that unique signature properties of each of these models are necessary to account for the data, but neither theory, by itself, explains all of its features. Specifically, the temporal gradedness of the splitter signal is strong support for temporal context, but is hard to explain using state models, while its flexibility and task-dependence is naturally accounted for using state inference, but poses a challenge otherwise. These theories suggest a number of avenues for future work, and we believe their application to splitter cells is a timely and informative domain for testing and refining theoretical ideas about hippocampal function.

Emergent behaviour and neural dynamics in artificial agents tracking odour plumes

Tracking an odour plume to locate its source under variable wind and plume statistics is a complex task. Flying insects routinely accomplish such tracking, often over long distances, in pursuit of food or mates. Several aspects of this remarkable behaviour and its underlying neural circuitry have been studied experimentally. Here we take a complementary in silico approach to develop an integrated understanding of their behaviour and neural computations. Specifically, we train artificial recurrent neural network agents using deep reinforcement learning to locate the source of simulated odour plumes that mimic features of plumes in a turbulent flow. Interestingly, the agents' emergent behaviours resemble those of flying insects, and the recurrent neural networks learn to compute task-relevant variables with distinct dynamic structures in population activity. Our analyses put forward a testable behavioural hypothesis for tracking plumes in changing wind direction, and we provide key intuitions for memory requirements and neural dynamics in odour plume tracking.

Exploring the role of structuralist methodology in the neuroscience of consciousness: a defense and analysis

Traditional contrastive analysis has been the foundation of consciousness science, but its limitations due to the lack of a reliable method for measuring states of consciousness have prompted the exploration of alternative approaches. Structuralist theories have gained attention as an alternative that focuses on the structural properties of phenomenal experience and seeks to identify their neural encoding via structural similarities between quality spaces and neural state spaces. However, the intertwining of philosophical assumptions about structuralism and structuralist methodology may pose a challenge to those who are skeptical of the former. In this paper, I offer an analysis and defense of structuralism as a methodological approach in consciousness science, which is partly independent of structuralist assumptions on the nature of consciousness. By doing so, I aim to make structuralist methodology more accessible to a broader scientific and philosophical audience. I situate methodological structuralism in the context of questions concerning mental representation, psychophysical measurement, holism, and functional relevance of neural processes. At last, I analyze the relationship between the structural approach and the distinction between conscious and unconscious states.

A space of goals: the cognitive geometry of informationally bounded agents

Traditionally, Euclidean geometry is treated by scientists as a priori and objective. However, when we take the position of an agent, the problem of selecting a best route should also factor in the abilities of the agent, its embodiment and particularly its cognitive effort. In this paper, we consider geometry in terms of travel between states within a world by incorporating information processing costs with the appropriate spatial distances. This induces a geometry that increasingly differs from the original geometry of the given world as information costs become increasingly important. We visualize this 'cognitive geometry' by projecting it onto two- and three-dimensional spaces showing distinct distortions reflecting the emergence of epistemic and information-saving strategies as well as pivot states. The analogies between traditional cost-based geometries and those induced by additional informational costs invite a generalization of the notion of geodesics as cheapest routes towards the notion of infodesics. In this perspective, the concept of infodesics is inspired by the property of geodesics that, travelling from a given start location to a given goal location along a geodesic, not only the goal, but all points along the way are visited at optimal cost from the start.

Deep problems with neural network models of human vision

Deep neural networks (DNNs) have had extraordinary successes in classifying photographic images of objects and are often described as the best models of biological vision. This conclusion is largely based on three sets of findings: (1) DNNs are more accurate than any other model in classifying images taken from various datasets, (2) DNNs do the best job in predicting the pattern of human errors in classifying objects taken from various behavioral datasets, and (3) DNNs do the best job in predicting brain signals in response to images taken from various brain datasets (e.g., single cell responses or fMRI data). However, these behavioral and brain datasets do not test hypotheses regarding what features are contributing to good predictions and we show that the predictions may be mediated by DNNs that share little overlap with biological vision. More problematically, we show that DNNs account for almost no results from psychological research. This contradicts the common claim that DNNs are good, let alone the best, models of human object recognition. We argue that theorists interested in developing biologically plausible models of human vision need to direct their attention to explaining psychological findings. More generally, theorists need to build models that explain the results of experiments that manipulate independent variables designed to test hypotheses rather than compete on making the best predictions. We conclude by briefly summarizing various promising modeling approaches that focus on psychological data.

Improving the accuracy of single-trial fMRI response estimates using GLMsingle

Advances in artificial intelligence have inspired a paradigm shift in human neuroscience, yielding large-scale functional magnetic resonance imaging (fMRI) datasets that provide high-resolution brain responses to thousands of naturalistic visual stimuli. Because such experiments necessarily involve brief stimulus durations and few repetitions of each stimulus, achieving sufficient signal-to-noise ratio can be a major challenge. We address this challenge by introducing GLMsingle, a scalable, user-friendly toolbox available in MATLAB and Python that enables accurate estimation of single-trial fMRI responses (glmsingle.org). Requiring only fMRI time-series data and a design matrix as inputs, GLMsingle integrates three techniques for improving the accuracy of trial-wise general linear model (GLM) beta estimates. First, for each voxel, a custom hemodynamic response function (HRF) is identified from a library of candidate functions. Second, cross-validation is used to derive a set of noise regressors from voxels unrelated to the experiment. Third, to improve the stability of beta estimates for closely spaced trials, betas are regularized on a voxel-wise basis using ridge regression. Applying GLMsingle to the Natural Scenes Dataset and BOLD5000, we find that GLMsingle substantially improves the reliability of beta estimates across visually-responsive cortex in all subjects. Comparable improvements in reliability are also observed in a smaller-scale auditory dataset from the StudyForrest experiment. These improvements translate into tangible benefits for higher-level analyses relevant to systems and cognitive neuroscience. We demonstrate that GLMsingle: (i) helps decorrelate response estimates between trials nearby in time; (ii) enhances representational similarity between subjects within and across datasets; and (iii) boosts one-versus-many decoding of visual stimuli. GLMsingle is a publicly available tool that can significantly improve the quality of past, present, and future neuroimaging datasets sampling brain activity across many experimental conditions.

The geometry of representational drift in natural and artificial neural networks

Neurons in sensory areas encode/represent stimuli. Surprisingly, recent studies have suggested that, even during persistent performance, these representations are not stable and change over the course of days and weeks. We examine stimulus representations from fluorescence recordings across hundreds of neurons in the visual cortex using in vivo two-photon calcium imaging and we corroborate previous studies finding that such representations change as experimental trials are repeated across days. This phenomenon has been termed "representational drift". In this study we geometrically characterize the properties of representational drift in the primary visual cortex of mice in two open datasets from the Allen Institute and propose a potential mechanism behind such drift. We observe representational drift both for passively presented stimuli, as well as for stimuli which are behaviorally relevant. Across experiments, the drift differs from in-session variance and most often occurs along directions that have the most in-class variance, leading to a significant turnover in the neurons used for a given representation. Interestingly, despite this significant change due to drift, linear classifiers trained to distinguish neuronal representations show little to no degradation in performance across days. The features we observe in the neural data are similar to properties of artificial neural networks where representations are updated by continual learning in the presence of dropout, i.e. a random masking of nodes/weights, but not other types of noise. Therefore, we conclude that a potential reason for the representational drift in biological networks is driven by an underlying dropout-like noise while continuously learning and that such a mechanism may be computational advantageous for the brain in the same way it is for artificial neural networks, e.g. preventing overfitting.