Orthogonal representations for robust contextdependent task performance in brains and neural networks Highlights d We trained artificial neural networks and humans on contextdependent decision tasks d Initial weight variance determined the network's representational geometry d Human fronto-parietal representations were similar to those of low-variance networks d Theory of nonlinear gating explains how these are formed in neural networks and brains
Humans and other animals accumulate resources, or wealth, by making successive risky decisions. If and how risk attitudes vary with wealth remains an open question. Here humans accumulated reward by accepting or rejecting successive monetary gambles within arbitrarily defined temporal contexts. Risk preferences changed substantially toward risk aversion as reward accumulated within a context, and blood oxygen level dependent (BOLD) signals in the ventromedial prefrontal cortex (PFC) tracked the latent growth of cumulative economic outcomes. Risky behavior was captured by a computational model in which reward prompts an adaptive update to the function that links utilities to choices. These findings can be understood if humans have evolved economic decision policies that fail to maximize overall expected value but reduce variance in cumulative outcomes, thereby ensuring that resources remain above a critical survival threshold.
When making economic choices, such as those between goods or gambles, humans act as if their internal representation of the value and probability of a prospect is distorted away from its true value. These distortions give rise to decisions which apparently fail to maximize reward, and preferences that reverse without reason. Why would humans have evolved to encode value and probability in a distorted fashion, in the face of selective pressure for reward-maximizing choices? Here, we show that under the simple assumption that humans make decisions with finite computational precision––in other words, that decisions are irreducibly corrupted by noise––the distortions of value and probability displayed by humans are approximately optimal in that they maximize reward and minimize uncertainty. In two empirical studies, we manipulate factors that change the reward-maximizing form of distortion, and find that in each case, humans adapt optimally to the manipulation. This work suggests an answer to the longstanding question of why humans make “irrational” economic choices.
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.
Anxiety disorders are the most common mental disorder worldwide. Although anxiety disorders differ in the nature of feared objects or situations, they share a common mechanism by which fear generalizes to related but innocuous objects, eliciting avoidance of objects and situations that pose no objective risk. This overgeneralization appears to be a crucial mechanism in the persistence of anxiety psychopathology. In this study we test whether an intervention that promotes discrimination learning reduces generalization of fear, in particular, harm expectancy and avoidance compared to an irrelevant (control) training. Healthy participants (N = 80) were randomly allocated to a training condition. Using a fear conditioning paradigm, participants first learned visual danger and safety signals (set 1). Baseline level of stimulus generalization was tested with ambiguous stimuli on a spectrum between the danger and safety signals. There were no differences between the training groups. Participants then received the stimulus discrimination training or a control training. After training, participants learned a new set of danger and safety signals (set 2), and the level of harm expectancy generalization and behavioural avoidance of ambiguous stimuli was tested. Although the training groups did not differ in fear generalization on a cognitive level (harm expectancy), the results showed a different pattern of avoidance of ambiguous stimuli, with the discrimination training group showing less avoidance of stimuli that resembled the safety signals. These results support the potential of interventions that promote discrimination learning in the treatment of anxiety disorders.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.