More Like this

1. The inherent capacity of neurons to learn order relations and support abstract reasoning

   https://doi.org/10.1038/s41467-026-76102-5  

   Yang, Yukun; Maass, Wolfgang (August 2026, Nature Communications)

   Abstract Brains extract relations between objects and concepts and integrate them into cognitive maps for decision-making. But it remains unclear how they achieve that. Here we present a rigorous theory showing that single neurons can already learn to extract ranks of items in a linear order with a simple local rule for synaptic plasticity. The resulting model explains human brain data on the emergence of cognitive maps from linear orders, accounts for the terminal item effect in transitive inference, and enables rapid reconfiguration of internal representations when new evidence appears. We also present a theoretical explanation for the surprising fact that 2D projections of neural representations of linear orders in the brain are curved rather than linear. Since the model requires only local synaptic plasticity in shallow networks, it is suited for relational learning and fast inference on low-energy edge devices. We demonstrate this on the neuromorphic chip Loihi 2. 

   more » « less
2. Local prediction-learning in high-dimensional spaces enables neural networks to plan

   https://doi.org/10.1038/s41467-024-46586-0  

   Stöckl, Christoph; Yang, Yukun; Maass, Wolfgang (December 2024, Nature Communications)

   Abstract Planning and problem solving are cornerstones of higher brain function. But we do not know how the brain does that. We show that learning of a suitable cognitive map of the problem space suffices. Furthermore, this can be reduced to learning to predict the next observation through local synaptic plasticity. Importantly, the resulting cognitive map encodes relations between actions and observations, and its emergent high-dimensional geometry provides a sense of direction for reaching distant goals. This quasi-Euclidean sense of direction provides a simple heuristic for online planning that works almost as well as the best offline planning algorithms from AI. If the problem space is a physical space, this method automatically extracts structural regularities from the sequence of observations that it receives so that it can generalize to unseen parts. This speeds up learning of navigation in 2D mazes and the locomotion with complex actuator systems, such as legged bodies. The cognitive map learner that we propose does not require a teacher, similar to self-attention networks (Transformers). But in contrast to Transformers, it does not require backpropagation of errors or very large datasets for learning. Hence it provides a blue-print for future energy-efficient neuromorphic hardware that acquires advanced cognitive capabilities through autonomous on-chip learning. 

   more » « less
3. Sleep disruption as a potential mechanism contributing to post‐zygotic reproductive isolation in hybrids

   https://doi.org/10.1111/brv.70083  

   Johnsson, Robin D; Semenov, Georgy A; Rice, Amber M; Taylor, Scott A; Roth, Timothy C (February 2026, Biological Reviews)

   ABSTRACT Hybridization occurs when different species mate and produce offspring. Although hybridization can have negative consequences for cognitive performance, the mechanisms underlying those effects are still poorly understood. A fundamental physiological process found in all animals studied to date that could be disrupted in hybrids is sleep. Given that mechanisms that occur within the brain during sleep may help maintain optimal cognitive performance, here we outline the potential impacts of hybridization on sleep and cognition. We suggest that sleep loss caused by hybridization could lead to negative impacts for neural and molecular mechanisms (e.g. neurogenesis, synaptic plasticity, and brain gene expression) associated with cognition, which may help explain some of the cognitive deficiency recently observed in hybrid birds. However, we acknowledge that these mechanisms may instead be directly impacted by hybridization, which in turn could also disrupt sleep with similar negative consequences for cognition. Limitations in sleep processes apparent in hybrids might influence hybrid fitness and therefore act as a post‐zygotic isolating barrier. 

   more » « less
4. Computation with Sequences of Assemblies in a Model of the Brain

   Dabagia, Max; Papadimitriou, Christos; Vempala, Santosh S (March 2024, Algorithmic Learning Theory 2024)

   Even as machine learning exceeds human-level performance on many applications, the generality, robustness, and rapidity of the brain’s learning capabilities remain unmatched. How cognition arises from neural activity is the central open question in neuroscience, inextricable from the study of intelligence itself. A simple formal model of neural activity was proposed in Papadimitriou (2020) and has been subsequently shown, through both mathematical proofs and simulations, to be capable of implementing certain simple cognitive operations via the creation and manipulation of assemblies of neurons. However, many intelligent behaviors rely on the ability to recognize, store, and manipulate temporal sequences of stimuli (planning, language, navigation, to list a few). Here we show that, in the same model, time can be captured naturally as precedence through synaptic weights and plasticity, and, as a result, a range of computations on sequences of assemblies can be carried out. In particular, repeated presentation of a sequence of stimuli leads to the memorization of the sequence through corresponding neural assemblies: upon future presentation of any stimulus in the sequence, the corresponding assembly and its subsequent ones will be activated, one after the other, until the end of the sequence. If the stimulus sequence is presented to two brain areas simultaneously, a scaffolded representation is created, resulting in more efficient memorization and recall, in agreement with cognitive experiments. Finally, we show that any finite state machine can be learned in a similar way, through the presentation of appropriate patterns of sequences. Through an extension of this mechanism, the model can be shown to be capable of universal computation. We support our analysis with a number of experiments to probe the limits of learning in this model in key ways. Taken together, these results provide a concrete hypothesis for the basis of the brain’s remarkable abilities to compute and learn, with sequences playing a vital role. 

   more » « less
5. Neural sampling from cognitive maps enables goal-directed imagination and planning

   https://doi.org/10.1038/s42256-026-01254-4  

   Lin, Hui; Yang, Yukun; Zhao, Rong; Pezzulo, Giovanni; Maass, Wolfgang (July 2026, Nature Machine Intelligence)

   Abstract Artificial intelligence systems are becoming more intelligent, but at a very high cost in terms of energy consumption and training requirements. By contrast, our brains only require 20 W of energy, they learn online and they can instantly adjust to changing contingencies. This begs the question what data structures, algorithms and learning methods enable brains to achieve that, and whether these can be ported into artificial devices. We are addressing this question for a core feature of intelligence: the capacity to plan and solve problems, including new problems that involve states that were never encountered before. Here we examine three tools that brains are likely to use for achieving that: cognitive maps, stochastic computing and compositional coding. We integrate these tools into a transparent neural network model, and demonstrate its power for flexible planning and problem-solving. Importantly, this approach is suitable for implementation by in-memory computing and other energy-efficient neuromorphic hardware. In particular, it only requires self-supervised local synaptic plasticity that is suited for on-chip learning. Hence, a core feature of brain intelligence—the capacity to generate solutions to problems that were never encountered before—does not require deep neural networks or large language models, and can be implemented in energy-efficient edge devices. 

   more » « less