Biology subjects

Summerfield, C.

Publications and source records attributed to Summerfield, C..

17 recordsLinked to original sources

bioRxiv2025

Distinct roles of hippocampus and neocortex in symbolic compositional generalization

Humans can combine symbols to generate new meanings. Here, we studied the regional neural mechanisms that might make this possible. We asked participants to combine two discrete, symbolic features (a shape and a colour) to make a novel spatial inference. BOLD data suggested that the hippocampus encoded elementary visual attributes in a high-dimensional, parallel format that permitted flexible individuation. In ventromedial prefrontal cortex (vmPFC), posterior parietal cortex (PPC) and primary visual cortex (V1), neural patterns for novel stimuli (composites) could be predicted as a linear combination of signals for familiar stimuli (elements). In vmPFC, this composition occurred in a high-dimensional format, but in PPC and V1, it took place in a low-dimensional, spatial, response-consistent frame of reference. These data offer new insights into the neural circuit underlying compositional generalization.

neuroscience↗

bioRxiv2023

Predictive and Interpretable: Combining Artificial Neural Networks and Classic Cognitive Models to Understand Human Learning and Decision Making

Quantitative models of behavior are a fundamental tool in cognitive science. Typically, models are hand-crafted to implement specific cognitive mechanisms. Such "classic" models are interpretable by design, but may provide poor fit to experimental data. Artificial neural networks (ANNs), on the contrary, can fit arbitrary datasets at the cost of opaque mechanisms. Here, we adopt a hybrid approach, combining the predictive power of ANNs with the interpretability of classic models. We apply this approach to Reinforcement Learning (RL), beginning with classic RL models and replacing their components one-by-one with ANNs. We find that hybrid models can provide similar fit to fully-general ANNs, while retaining the interpretability of classic cognitive models: They reveal reward-based learning mechanisms in humans that are strikingly similar to classic RL. They also reveal mechanisms not contained in classic models, including separate rewardblind mechanisms, and the specific memory contents relevant to reward-based and reward-blind mechanisms.

neuroscience↗

bioRxiv2023

Goal-seeking compresses neural codes for space in the human hippocampus and orbitofrontal cortex

Humans can navigate flexibly to meet their goals. Here, we asked how the neural representation of allocentric space is distorted by goal-directed behaviour. Participants navigated an agent to two successive goal locations in a grid world environment comprising four interlinked rooms, with a contextual cue indicating the conditional dependence of one goal location on another. Examining the neural geometry by which room and context were encoded in fMRI signals, we found that map-like representations of the environment emerged in both hippocampus and neocortex. Cognitive maps in hippocampus and orbitofrontal cortices were compressed so that locations cued as goals were coded together in neural state space, and these distortions predicted successful learning. This effect was captured by a computational model in which current and prospective locations are jointly encoded in a place code, providing a theory of how goals warp the neural representation of space in macroscopic neural signals.

neuroscience↗

bioRxiv2021

Neural knowledge assembly in humans and deep networks

Human understanding of the world can change rapidly when new information comes to light, such as when a plot twist occurs in a work of fiction. This flexible "knowledge assembly" requires few-shot reorganisation of neural codes for relations among objects and events. However, existing computational theories are largely silent about how this could occur. Here, participants learned a transitive ordering among novel objects within two distinct contexts, before exposure to new knowledge that revealed how they were linked. BOLD signals in dorsal frontoparietal cortical areas revealed that objects were rapidly and dramatically rearranged on the neural manifold after minimal exposure to linking information. We then adapt stochastic online gradient descent to permit similar rapid knowledge assembly in a neural network model.

neuroscience↗

bioRxiv2021

Generalising state transition structure in the human medial temporal lobe

Effective planning involves knowing where different actions will take us. However natural environments are rich and complex, leading to an exponential increase in memory demand as a plan grows in depth. One potential solution to this problem is to share the neural state transition functions used for planning between similar contexts. Here, we asked human participants to perform a sequential decision making task designed so that knowledge could be shared between some contexts but not others. Computational modelling showed that participants shared a model of state transitions between contexts where appropriate. fMRI data identified the medial temporal lobe as a locus for learning of state transitions, and within the same region, correlated BOLD patterns were observed in contexts where state transition information were shared. Finally, we show that the transition model is updated more strongly following the receipt of positive compared to negative outcomes, a finding that challenges conventional theories of planning which assume knowledge about our environment is updated independently of outcomes received. Together, these findings propose a computational and neural account of how information relevant for planning can be shared between contexts.

neuroscience↗

bioRxiv2021

Rich and lazy learning of task representations in brains and neural networks

How do neural populations code for multiple, potentially conflicting tasks? Here, we used computational simulations involving neural networks to define "lazy" and "rich" coding solutions to this multitasking problem, which trade off learning speed for robustness. During lazy learning the input dimensionality is expanded by random projections to the network hidden layer, whereas in rich learning hidden units acquire structured representations that privilege relevant over irrelevant features. For context-dependent decision-making, one rich solution is to project task representations onto low-dimensional and orthogonal manifolds. Using behavioural testing and neuroimaging in humans, and analysis of neural signals from macaque prefrontal cortex, we report evidence for neural coding patterns in biological brains whose dimensionality and neural geometry are consistent with the rich learning regime.

neuroscience↗

bioRxiv2020

Hippocampal place cells encode global location but not changes in environmental connectivity in a 4-room navigation task

Flexible navigation relies on a cognitive map of space, thought to be implemented by hippocampal place cells: neurons that exhibit location-specific firing. In connected environments, optimal navigation requires keeping track of ones location and of the available connections between subspaces. We examined whether the dorsal CA1 place cells of rats encode environmental connectivity in four geometrically-identical boxes arranged in a square. Rats moved between boxes by pushing saloon-type doors that could be locked in one or both directions. While rats demonstrated knowledge of environmental connectivity, their place cells did not respond to connectivity changes, nor did they represent doorways differently from other locations. Importantly, place cells coded the space in a global frame, expressing minimal repetitive fields despite the repetitive geometry (global coding). These results suggest that CA1 place cells provide a spatial map that does not explicitly include connectivity.

neuroscience↗

bioRxiv2020

Neural state space alignment for magnitude generalization in humans and recurrent networks

A prerequisite for intelligent behaviour is to understand how stimuli are related and to generalise this knowledge across contexts. Generalisation can be challenging when relational patterns are shared across contexts but exist on different physical scales. Here, we studied neural representations in humans and recurrent neural networks performing a magnitude comparison task, for which it was advantageous to generalise concepts of "more" or "less" between contexts. Using multivariate analysis of human brain signals and of neural network hidden unit activity, we observed that both systems developed parallel neural "number lines" for each context. In both model systems, these number state spaces were aligned in a way that explicitly facilitated generalisation of relational concepts (more and less). These findings suggest a previously overlooked role for neural normalisation in supporting transfer of a simple form of abstract relational knowledge (magnitude) in humans and machine learning systems.

neuroscience↗

bioRxiv2020

A normative account of confirmatory biases during reinforcement learning

Reinforcement learning involves updating estimates of the value of states and actions on the basis of experience. Previous work has shown that in humans, reinforcement learning exhibits a confirmatory bias: when updating the value of a chosen option, estimates are revised more radically following positive than negative reward prediction errors, but the converse is observed when updating the unchosen option value estimate. Here, we simulate performance on a multi-arm bandit task to examine the consequences of a confirmatory bias for reward harvesting. We report a paradoxical finding: that confirmatory biases allow the agent to maximise reward relative to an unbiased updating rule. This principle holds over a wide range of experimental settings and is most influential when decisions are corrupted by noise. We show that this occurs because on average, confirmatory biases lead to overestimating the value of more valuable bandits, and underestimating the value of less valuable bandits, rendering decisions overall more robust in the face of noise. Our results show how apparently suboptimal learning policies can in fact be reward-maximising if decisions are made with finite computational precision.

animal behavior and cognition↗

bioRxiv2019

Selective integration during sequential sampling in posterior neural signals

Decisions are typically made after integrating information about multiple attributes of alternatives in a choice set. The computational mechanisms by which this integration occurs have been a focus of extensive research in humans and other animals. Where observers are obliged to consider attributes in turn, a framework known as \"selective integration\" can capture salient biases in human choices. The model proposes that successive attributes compete for processing resources and integration is biased towards the alternative with the locally preferred attribute. Quantitative analysis shows that this model, although it discards choice-relevant information, is optimal when the observers decisions are corrupted by noise that occurs beyond the sensory stage. Here, we used scalp electroencephalographic (EEG) recordings to test a neural prediction of the model: that locally preferred attributes should be encoded with higher gain in neural signals over posterior cortex. Over two sessions, human observers (of either sex) judged which of two simultaneous streams of bars had the higher (or lower) average height. The selective integration model fit the data better than a rival model without bias. Single-trial analysis showed that neural signals contralateral to the preferred attribute covaried more steeply with the decision information conferred by locally preferred attributes. These findings provide neural evidence in support of selective integration, complementing existing behavioural work.\n\nSignificance StatementWe often make choices about stimuli with multiple attributes, such as when deciding which car to buy on the basis of price, performance and fuel economy. A model of the choice process, known as selective integration, proposes that rather than taking all of the decision-relevant information equally into account when making choices, we discard or overlook a portion of it. Although information is discarded, this strategy can lead to better decisions when memory is limited. Here, we test and confirm predictions of the model about the brain signals that occur when different stimulus attributes of stimulus are being evaluated. Our work provides the first neural support for the selective integration model.

neuroscience↗

bioRxiv2018

Causal inference in the multisensory brain

When combining information across different senses humans need to flexibly select cues of a common origin whilst avoiding distraction from irrelevant inputs. The brain could solve this challenge using a hierarchical principle, by deriving rapidly a fused sensory estimate for computational expediency and, later and if required, filtering out irrelevant signals based on the inferred sensory cause(s). Analysing time- and source-resolved human magnetoencephalographic data we unveil a systematic spatio-temporal cascade of the relevant computations, starting with early segregated unisensory representations, continuing with sensory fusion in parietal-temporal regions and culminating as causal inference in the frontal lobe. Our results reconcile previous computational accounts of multisensory perception by showing that prefrontal cortex guides flexible integrative behaviour based on candidate representations established in sensory and association cortices, thereby framing multisensory integration in the generalised context of adaptive behaviour.

neuroscience↗

bioRxiv2018

Neural structure mapping in human probabilistic reward learning

Humans can learn abstract concepts that describe invariances over relational patterns in data. One such concept, known as magnitude, allows stimuli to be compactly represented by a single dimension (i.e. on a mental line), for example according to their cardinality, size or value. Here, we measured representations of magnitude in humans by recording neural signals whilst they viewed symbolic numbers. During a subsequent reward-guided learning task, the neural patterns elicited by novel complex visual images reflected their pay-out probability in a way that suggested they were encoded onto the same mental number line. Our findings suggest that in humans, learning about values is accompanied by structural alignment of value representations with neural codes for the concept of magnitude.

neuroscience↗

bioRxiv2018

A network for computing value homeostasis in the human medial prefrontal cortex

Humans and other animals make decisions in order to satisfy their goals. However, it remains unknown how neural circuits compute which of multiple possible goals should be pursued (e.g. when balancing hunger and thirst) and combine these signals with estimates of available reward alternatives. Here, humans undergoing functional magnetic resonance imaging (fMRI) accumulated two distinct assets over a sequence of trials. Financial outcomes depended on the minimum cumulate of either asset, creating a need to maintain \"value homeostasis\" by redressing any imbalance among the assets. BOLD signals in the dorsal anterior cingulate cortex (dACC) tracked the level of homeostatic imbalance among goals, whereas the ventromedial prefrontal cortex (vmPFC) signalled the level of homeostatic redress incurred by a choice, rather than the overall amount received. These results suggest that a network of medial frontal brain regions compute a value signal that maintains homeostatic balance among internal goals.

neuroscience↗

bioRxiv2018

Human noise blindness drives suboptimal cognitive inference

Humans typically make near-optimal sensorimotor judgments but show systematic biases when making more cognitive judgments. Here we test the hypothesis that, while humans are sensitive to the noise present during early sensory processing, the \"optimality gap\" arises because they are blind to noise introduced by later cognitive integration of variable or discordant pieces of information. In six psychophysical experiments, human observers judged the average orientation of an array of contrast gratings. We varied the stimulus contrast (encoding noise) and orientation variability (integration noise) of the array. Participants adapted near-optimally to changes in encoding noise, but, under increased integration noise, displayed a range of suboptimal behaviours: they ignored stimulus base rates, reported excessive confidence in their choices, and refrained from opting out of objectively difficult trials. These overconfident behaviours were captured by a Bayesian model which is blind to integration noise. Our study provides a computationally grounded explanation of suboptimal cognitive inferences.

neuroscience↗

bioRxiv2018

Focused learning promotes continual task performance in humans

Humans can learn to perform multiple tasks in succession over the lifespan (\"continual\" learning), whereas current machine learning systems fail. Here, we investigated the cognitive mechanisms that permit successful continual learning in humans. Unlike neural networks, humans that were trained on temporally autocorrelated task objectives (focussed training) learned to perform new tasks more effectively, and performed better on a later test involving randomly interleaved tasks. Analysis of error patterns suggested that focussed learning permitted the formation of factorised task representations that were protected from mutual interference. Furthermore, individuals with a strong prior tendency to represent the task space in a factorised manner enjoyed greater benefit of focussed over interleaved training. Building artificial agents that learn to factorise tasks appropriately may be a promising route to solving continual task performance in machine learning.\n\nSignificance StatementHumans learn to perform many different tasks over the lifespan, such as speaking both French and Spanish. The brain has to represent task information without mutual interference. In machine learning, this \"continual learning\" is a major unsolved challenge. Here, we studied the patterns of errors made by humans and state-of-the-art deep networks whilst they learned new tasks from scratch and without instruction. Humans, but not machines, seem to benefit from training regimes that focussed on one task at a time, especially when they had a prior bias to represent stimuli in a way that facilitated task separation. Machines trained to exhibit the same prior bias suffered less interference between tasks, suggesting new avenues for solving continual learning in artificial systems.

neuroscience↗

bioRxiv2017

Adaptive gain control in the human dorsal anterior cingulate cortex

When making decisions, humans are often distracted by irrelevant information. Distraction has different impact on perceptual, cognitive and value-guided choices, giving rise to well-described behavioural phenomena such as the tilt illusion, conflict adaptation, or economic decoy effects. However, a single, unified model that can account for all these phenomena has yet to emerge. Here, we offer one such account, based on adaptive gain control, and additionally show that it successfully predicts a range of counterintuitive new behavioural phenomena on variants of a classic cognitive paradigm, the Eriksen flanker task. We also report that BOLD signals in a dorsal network prominently including the anterior cingulate cortex (dACC), index a gain-modulated decision variable predicted by the model. This work unifies the study of distraction across perceptual, cognitive and economic domains.

neuroscience↗

bioRxiv2017

Robust Averaging Protects Decisions from Noise in Neural Computations

An ideal observer will give equivalent weight to sources of information that are equally reliable. However, when averaging visual information, human observers tend to downweight or discount features that are relatively outlying or deviant ( robust averaging). Why humans adopt an integration policy that discards important decision information remains unknown. Here, observers were asked to judge the average tilt in a circular array of high-contrast gratings, relative to an orientation boundary defined by a central reference grating. Observers showed robust averaging of orientation, but the extent to which they did so was a positive predictor of their overall performance. Using computational simulations, we show that although robust averaging is suboptimal for a perfect integrator, it paradoxically enhances performance in the presence of \"late\" noise, i.e. which corrupts decisions during integration. In other words, robust decision strategies increase the brains resilience to noise arising in neural computations during decision-making.\n\nAuthor SummaryHumans often make decisions by averaging information from multiple sources. When all the sources are equally reliable, they should all have equivalent impact (or weight) on the decisions of an \"ideal\" observer, i.e. one with perfect memory. However, recent experiments have suggested that humans give unequal weight to sources that are deviant or unusual, a phenomenon called \"robust averaging\". Here, we use computer simulations to try to understand why humans do this. Our simulations show that under the assumption that information processing is limited by a source of internal uncertainty that we call \"late\" noise, robust averaging actually leads to improved performance. Using behavioural testing, we replicate the finding of robust averaging in a cohort of healthy humans, and show that those participants that engage in robust averaging perform better on the task. This study thus provides new information about the limitations on human decision-making.

neuroscience↗