Search

We use efficient coding principles borrowed from sensory neuroscience to derive the optimal neural population to encode a reward distribution. We show that the responses of dopaminergic reward prediction error neurons in mouse and macaque are similar to those of the efficient code in the following ways: the neurons have a broad distribution of midpoints covering the reward distribution; neurons with higher thresholds have higher gains, more convex tuning functions and lower slopes; and their slope is higher when the reward distribution is narrower. Furthermore, we derive learning rules that converge to the efficient code. The learning rule for the position of the neuron on the reward axis closely resembles distributional reinforcement learning. Thus, reward prediction error neuron responses may be optimized to broadcast an efficient reward signal, forming a connection between efficient coding and reinforcement learning, two of the most successful theories in computational neuroscience., (© 2024. The Author(s), under exclusive licence to Springer Nature America, Inc.)

People procrastinate, but why? One long-standing hypothesis is that temporal discounting drives procrastination: in a task with a distant future reward, the discounted future reward fails to provide sufficient motivation to initiate work early. However, empirical evidence for this hypothesis has been lacking. Here, we used a long-term real-world task and a novel measure of procrastination to examine the association between temporal discounting and real-world procrastination. To measure procrastination, we critically measured the entire time course of the work progress instead of a single endpoint, such as task completion day. This approach allowed us to compute a fine-grained metric of procrastination. We found a positive correlation between individuals' degree of future reward discounting and their level of procrastination, suggesting that temporal discounting is a cognitive mechanism underlying procrastination. We found no evidence of a correlation when we, instead, measured procrastination by task completion day or by survey. This association between temporal discounting and procrastination offers empirical support for targeted interventions that could mitigate procrastination, such as modifying incentive systems to reduce the delay to a reward and lowering discount rates., (© 2024. The Author(s).)

When developing models in cognitive science, researchers typically start with their own intuitions about human behavior in a given task and then build in mechanisms that explain additional aspects of the data. This refinement step is often hindered by how difficult it is to distinguish the unpredictable randomness of people's decisions from meaningful deviations between those decisions and the model. One solution for this problem is to compare the model against deep neural networks trained on behavioral data, which can detect almost any pattern given sufficient data. Here, we apply this method to the domain of planning with a heuristic search model for human play in 4-in-a-row, a combinatorial game where participants think multiple steps into the future. Using a data set consisting of 10,874,547 games, we train deep neural networks to predict human moves and find that they accurately do so while capturing meaningful patterns in the data. Thus, deviations between the model and the best network allow us to identify opportunities for model improvement despite starting with a model that has undergone substantial testing in previous work. Based on this analysis, we add three extensions to the model that range from a simple opening bias to specific adjustments regarding endgame planning. Overall, our work demonstrates the advantages of model comparison with a high-performance deep neural network as well as the feasibility of scaling cognitive models to massive data sets for systematically investigating the processes underlying human sequential decision-making., (© 2023. The Author(s).)

The study of the brain's representations of uncertainty is a central topic in neuroscience. Unlike most quantities of which the neural representation is studied, uncertainty is a property of an observer's beliefs about the world, which poses specific methodological challenges. We analyze how the literature on the neural representations of uncertainty addresses those challenges and distinguish between 'code-driven' and 'correlational' approaches. Code-driven approaches make assumptions about the neural code for representing world states and the associated uncertainty. By contrast, correlational approaches search for relationships between uncertainty and neural activity without constraints on the neural representation of the world state that this uncertainty accompanies. To compare these two approaches, we apply several criteria for neural representations: sensitivity, specificity, invariance and functionality. Our analysis reveals that the two approaches lead to different but complementary findings, shaping new research questions and guiding future experiments., (© 2023. Springer Nature America, Inc.)

A hallmark of human intelligence is the ability to plan multiple steps into the future 1,2 . Despite decades of research 3-5 , it is still debated whether skilled decision-makers plan more steps ahead than novices 6-8 . Traditionally, the study of expertise in planning has used board games such as chess, but the complexity of these games poses a barrier to quantitative estimates of planning depth. Conversely, common planning tasks in cognitive science often have a lower complexity 9,10 and impose a ceiling for the depth to which any player can plan. Here we investigate expertise in a complex board game that offers ample opportunity for skilled players to plan deeply. We use model fitting methods to show that human behaviour can be captured using a computational cognitive model based on heuristic search. To validate this model, we predict human choices, response times and eye movements. We also perform a Turing test and a reconstruction experiment. Using the model, we find robust evidence for increased planning depth with expertise in both laboratory and large-scale mobile data. Experts memorize and reconstruct board features more accurately. Using complex tasks combined with precise behavioural modelling might expand our understanding of human planning and help to bridge the gap with progress in artificial intelligence., (© 2023. The Author(s), under exclusive licence to Springer Nature Limited.)

Trust decisions are inherently uncertain, as people usually have incomplete information about the trustworthiness of the other person prior to their decision to trust or not trust. Therefore, it is typically beneficial to gather information about a trustee's past behaviour before deciding whether or not to trust them. However, elaborate inquiries about a trustee's behaviour may change the trustee's willingness to reciprocate, causing either a decrease due to the trustee's negative impressions of the investor or an increase because the investor appears to be highly betrayal-averse to the trustee. In turn, such a change could cause the investor to gather less or more information, respectively. Here, we examine how information acquisition is modulated by social context, monetary cost, and the trustee's trustworthiness. We gave participants the opportunity to sequentially sample information about a trustee's reciprocation history before they decided whether or not to invest. Participants sampled less when there was a monetary cost and when the gathered information was more conclusive. On some trials, we induced a social context by telling the participant that the trustee would learn how much the participant sampled ("overt sampling"). Crucially, when sampling was free, participants sampled less when sampling was overt than when it was covert, suggesting that they avoided leaving negative impressions. We find that the data were well accounted for by a Bayesian heuristic model, in which the agent continues sampling until uncertainty about trustworthiness-as measured by the width of the posterior belief-drops below a level that they find tolerable. This study opens the door to broader applications of the tools and models of information sampling to social decision-making.

Primary aldosteronism (PA) is mainly treated by mineralocorticoid receptor antagonists or laparoscopic adrenalectomy (LA), but the effectiveness of surgical versus medical treatment in patients with adrenal venous sampling (AVS)-proven unilateral PA is unclear. Fifty-one consecutive patients with AVS-proven PA were enrolled. We compared the therapeutic effects between the surgery group (n = 21) and medication group (n = 30) by evaluating the complete control rate (CCR) of hypertension, blood pressure (BP), and number of antihypertensive drugs after a long-term follow-up (>12 months). The CCR of hypertension was assessed using a multivariate adjusted Cox proportional hazards regression model. After a mean follow-up of 21.18 ± 5.35 months, the CCR was significantly higher in the surgery than medication group (85.7% vs. 13.3%, respectively; p < 0.001). Before adjustment for covariates, the CCR of hypertension in patients who underwent LA was 7.75 times higher than that in patients who underwent medical treatment (95% CI, 2.33-25.78; p = 0.001); significant results were also shown in the adjusted models. Systolic and diastolic BP were also lower in the surgery than medication group (120.3 ± 12.99 vs. 133.54 ± 16.60 and 79.00 ± 7.62 vs. 87.35 ± 12.36 mmHg, respectively; p = 0.01 for both), as was the number of antihypertensive drugs (0.19 ± 0.51 vs. 2.33 ± 0.78, respectively; p < 0.001). The rate of hypokalemia was not significantly different between the two groups (0.0% vs. 13.3%, respectively; p = 0.13). In conclusion, AVS plays an essential role in the subtype diagnosis of PA, and surgical candidates with AVS-proven unilateral PA should be highly suggested to undergo LA.

Bayesian models of behavior suggest that organisms represent uncertainty associated with sensory variables. However, the neural code of uncertainty remains elusive. A central hypothesis is that uncertainty is encoded in the population activity of cortical neurons in the form of likelihood functions. We tested this hypothesis by simultaneously recording population activity from primate visual cortex during a visual categorization task in which trial-to-trial uncertainty about stimulus orientation was relevant for the decision. We decoded the likelihood function from the trial-to-trial population activity and found that it predicted decisions better than a point estimate of orientation. This remained true when we conditioned on the true orientation, suggesting that internal fluctuations in neural activity drive behaviorally meaningful variations in the likelihood function. Our results establish the role of population-encoded likelihood functions in mediating behavior and provide a neural underpinning for Bayesian models of perception.

Sequential and persistent activity models are two prominent models of short-term memory in neural circuits. In persistent activity models, memories are represented in persistent or nearly persistent activity patterns across a population of neurons, whereas in sequential models, memories are represented dynamically by a sequential activity pattern across the population. Experimental evidence for both models has been reported previously. However, it has been unclear under what conditions these two qualitatively different types of solutions emerge in neural circuits. Here, we address this question by training recurrent neural networks on several short-term memory tasks under a wide range of circuit and task manipulations. We show that both sequential and nearly persistent solutions are part of a spectrum that emerges naturally in trained networks under different conditions. Our results help to clarify some seemingly contradictory experimental results on the existence of sequential versus persistent activity-based short-term memory mechanisms in the brain.

Visual working memory (VWM), the brief retention of past visual information, supports a range of cognitive functions. One of the defining, and largely studied, characteristics of VWM is how resource-limited it is, raising questions about how this resource is shared or split across memoranda. Since objects are rarely equally important in the real world, we ask how people split this resource in settings where objects have different levels of importance. In a psychophysical experiment, participants remembered the location of four targets with different probabilities of being tested after a delay. We then measured their memory accuracy of one of the targets. We found that participants allocated more resource to memoranda with higher priority, but underallocated resource to high- and overallocated to low-priority targets relative to the true probability of being tested. These results are well explained by a computational model in which resource is allocated to minimize expected estimation error. We replicated this finding in a second experiment in which participants bet on their memory fidelity after making the location estimate. The results of this experiment show that people have access to and utilize the quality of their memory when making decisions. Furthermore, people again allocate resource in a way that minimizes memory errors, even in a context in which an alternative strategy was incentivized. Our study not only shows that people are allocating resource according to behavioral relevance, but suggests that they are doing so with the aim of maximizing memory accuracy.

Human perception consists of the continuous integration of sensory cues pertaining to the same object. While it has been fairly well shown that humans use an optimal strategy when integrating low-level cues proportional to their relative reliability, the integration processes underlying high-level perception are much less understood. Here we investigate cue integration in a complex high-level perceptual system, the human face processing system. We tested cue integration of facial form and motion in an identity categorization task and found that an optimal model could successfully predict subjects' identity choices. Our results suggest that optimal cue integration may be implemented across different levels of the visual processing hierarchy.

Many human or animal diseases are related to aggregation of proteins. A viable biological organism should maintain in non-equilibrium states. How protein aggregate and why biological organisms can maintain in non-equilibrium states are not well understood. As a first step to understand such complex systems problems, we consider simple model systems containing polymer chains and solvent particles. The strength of the spring to connect two neighboring monomers in a polymer chain is controlled by a parameter s with s → ∞ for rigid-bond. The strengths of bending and torsion angle dependent interactions are controlled by a parameter s A with s A  → -∞ corresponding to no bending and torsion angle dependent interactions. We find that for very small s A , polymer chains tend to aggregate spontaneously and the trend is independent of the strength of spring. For strong springs, the speed distribution of monomers in the parallel (along the direction of the spring to connect two neighboring monomers) and perpendicular directions have different effective temperatures and such systems are in non-equilibrium states.

The wheat avenin-like proteins (ALP) are considered atypical gluten constituents and have shown positive effects on dough properties revealed using a transgenic approach. However, to date the genetic architecture of ALP genes is unclear, making it impossible to be utilized in wheat breeding. In the current study, three genes of type-b ALPs were identified and mapped to chromosomes 7AS, 4AL and 7DS. The coding gene sequence of both TaALP-7A and TaALP-7D was 855 bp long, encoding two identical homologous 284 amino acid long proteins. TaALP-4A was 858 bp long, encoding a 285 amino acid protein variant. Three alleles were identified for TaALP-7A and four for TaALP-4A. TaALP-7A alleles were of two types: type-1, which includes TaALP-7A1 andTaALP-7A2, encodes mature proteins, while type-2, represented byTaALP-7A3, contains a stop codon in the coding region and thus does not encode a mature protein. Dough quality testing of 102 wheat cultivars established a highly significant association of the type-1 TaALP-7A allele with better wheat processing quality. This allelic effects were confirmed among a range of commercial wheat cultivars. Our research makes the ALP be the first of such genetic variation source that can be readily utilized in wheat breeding.

Bayesian theories of neural coding propose that sensory uncertainty is represented by a probability distribution encoded in neural population activity, but direct neural evidence supporting this hypothesis is currently lacking. Using fMRI in combination with a generative model-based analysis, we found that probability distributions reflecting sensory uncertainty could reliably be estimated from human visual cortex and, moreover, that observers appeared to use knowledge of this uncertainty in their perceptual decisions.

Trait decay may occur when selective pressures shift, owing to changes in environment or life style, rendering formerly adaptive traits non-functional or even maladaptive. It remains largely unknown if such decay would stem from multiple mutations with small effects or rather involve few loci with major phenotypic effects. Here, we investigate the decay of female sexual traits, and the genetic causes thereof, in a transition from haplodiploid sexual reproduction to endosymbiont-induced asexual reproduction in the parasitoid wasp Asobara japonica. We take advantage of the fact that asexual females cured of their endosymbionts produce sons instead of daughters, and that these sons can be crossed with sexual females. By combining behavioral experiments with crosses designed to introgress alleles from the asexual into the sexual genome, we found that sexual attractiveness, mating, egg fertilization and plastic adjustment of offspring sex ratio (in response to variation in local mate competition) are decayed in asexual A. japonica females. Furthermore, introgression experiments revealed that the propensity for cured asexual females to produce only sons (because of decayed sexual attractiveness, mating behavior and/or egg fertilization) is likely caused by recessive genetic effects at a single locus. Recessive effects were also found to cause decay of plastic sex-ratio adjustment under variable levels of local mate competition. Our results suggest that few recessive mutations drive decay of female sexual traits, at least in asexual species deriving from haplodiploid sexual ancestors.

Working memory is widely considered to be limited in capacity, holding a fixed, small number of items, such as Miller's 'magical number' seven or Cowan's four. It has recently been proposed that working memory might better be conceptualized as a limited resource that is distributed flexibly among all items to be maintained in memory. According to this view, the quality rather than the quantity of working memory representations determines performance. Here we consider behavioral and emerging neural evidence for this proposal.

There is strong behavioral and physiological evidence that the brain both represents probability distributions and performs probabilistic inference. Computational neuroscientists have started to shed light on how these probabilistic representations and computations might be implemented in neural circuits. One particularly appealing aspect of these theories is their generality: they can be used to model a wide range of tasks, from sensory processing to high-level cognition. To date, however, these theories have only been applied to very simple tasks. Here we discuss the challenges that will emerge as researchers start focusing their efforts on real-life computations, with a focus on probabilistic learning, structural learning and approximate inference.

The ability to search efficiently for a target in a cluttered environment is one of the most remarkable functions of the nervous system. This task is difficult under natural circumstances, as the reliability of sensory information can vary greatly across space and time and is typically a priori unknown to the observer. In contrast, visual-search experiments commonly use stimuli of equal and known reliability. In a target detection task, we randomly assigned high or low reliability to each item on a trial-by-trial basis. An optimal observer would weight the observations by their trial-to-trial reliability and combine them using a specific nonlinear integration rule. We found that humans were near-optimal, regardless of whether distractors were homogeneous or heterogeneous and whether reliability was manipulated through contrast or shape. We present a neural-network implementation of near-optimal visual search based on probabilistic population coding. The network matched human performance.

Recent psychophysical experiments indicate that humans perform near-optimal Bayesian inference in a wide variety of tasks, ranging from cue integration to decision making to motor control. This implies that neurons both represent probability distributions and combine those distributions according to a close approximation to Bayes' rule. At first sight, it would seem that the high variability in the responses of cortical neurons would make it difficult to implement such optimal statistical inference in cortical circuits. We argue that, in fact, this variability implies that populations of neurons automatically represent probability distributions over the stimulus, a type of code we call probabilistic population codes. Moreover, we demonstrate that the Poisson-like variability observed in cortex reduces a broad class of Bayesian inference to simple linear combinations of populations of neural activity. These results hold for arbitrary probability distributions over the stimulus, for tuning curves of arbitrary shape and for realistic neuronal variability.

Aim: To display the aggregation-prone ligand binding domain (LBD) of the human peroxisome proliferator-activated receptor gamma (PPARgamma) on the surface of bacteriophages to establish an easy screening assay for the identification of PPARgamma ligands., Methods: Plasmids were constructed for the expression of the PPARgamma LBD as a fusion to the N-terminus of the g3p protein of filamentous phage or the C-terminus of the capsid protein D (pD) of phage lambda. The fusion proteins were expressed in E coli and solubility characteristics were compared. Polyclonal antibodies against the LBD as well as the pD protein were prepared for Western blot analysis and phage capture assay., Results: The pD-LBD fusion protein was partially soluble, whereas the LBD-g3p fusion protein was detected only in the insoluble fraction. The pD-LBD fusion protein was efficiently incorporated in phage particles. Furthermore, the LBD was shown to be displayed on the surface of bacteriophage lambda. On average, the pD-LBD fusion protein accounted for 28% of the total pD protein in the lambda head capsid., Conclusion: The hydrophobic PPARgamma LBD was expressed as a soluble form of fusion protein in E coli and displayed on the surface of bacteriophage lambda when it was fused to the lambda pD protein. The lambda pD fusion system could be used for improving the solubility of proteins that tend to form inclusion bodies when expressed in E coli. The lambda phage particles displaying the LBD of PPARgamma may be of great value for the identification of novel PPARgamma ligands.