Papers

Hierarchical predictive coding proposes a compelling hypothesis of brain computation, suggesting that the cortex builds layered predictions to minimize surprise. Yet most models rely on error-coding neurons or generative modeling of unclear biological plausibility. Here, we examine a biologically plausible framework in which the functional goals of predictive coding emerge from local contrastive learning and simple activity cancellation. Building on recent machine learning advances, we present a recurrent variant of the Forward-Forward (FF) algorithm with an inverted objective that increases activity for negative data. This setup yields predictive representations across layers, capturing hallmark features of cortical computation such as top-down modulation and surprise signaling. Our results suggest that key principles of predictive coding can emerge from simple, local learning rules, offering a new bridge between neuroscience and machine learning.

Photosensitivity is a neurological condition in which the brain generates epileptiform activity in response to visual stimuli. The standardized clinical diagnosis procedure, named Intermitent Photic Stimulation (IPS), involves exposing patients to a white flashing light at different frequencies to provoke this reaction. However, clinical neurophysiologists report that this protocol is insufficient, leading to underdiagnosis. VR-IPS is a flexible visual stimulation tool that extends conventional flashing-light stimulation by incorporating configurable color lights, static and oscillatory motion patterns, and photoprovocative videos. The system comprises a stimulation module that can run on Virtual Reality headsets or on standard monitors, together with a control module that allows clinicians to create, select, and configure visual stimulus, while controlling the whole stimulation procedure.

The pruning of network connections is key to brain function but, despite its importance, there exist few biologically-plausible pruning rules with demonstrated good performance. In this work we evaluate noise-prune, a recently introduced unsupervised local pruning rule for recurrent networks that uses noisy fluctuations to determine the importance of connections. Noise-prune has previously only been empirically tested on random networks without a specific computational function. We show that noise-prune preserves task-performance in task-trained recurrent neural networks, greatly outperforming a strategy that only uses the magnitude of connections and performing on par with or exceeding a non-local strategy that uses second-order information. Rather than deterministically removing connections that fall below a certain threshold importance, noise-prune samples connections to preserve based on their importance and strengthens retained connections to preserve average synaptic strength. We show that this sampling and rescaling is essential to good performance, but that the optimal empirical degree of rescaling is lower than that predicted by the original theoretical argument. Our work thus validates noise-prune as a biologically-plausible pruning rule for functional recurrent network architectures and characterizes its optimal parameter settings.

Insomnia disorder is defined at the syndrome level, yet similar complaints can arise from different abnormalities in sleep regulation, state transition, state stabilization, spatial recruitment, architectural sequencing, and state perception. We propose a multiaxial dynamical framework whose principal contribution is organizational: a candidate profile is specified by the dynamical operation that fails, the sleep stage or boundary at which it fails, and its causal status. Objective sleep duration, age, circadian phase, comorbidity, medication exposure, and night-to-night variability are modifier/covariate dimensions rather than additional mechanistic classes. A local Landau-Ginzburg formalism, adopted from prior cortical and sleep-dynamics work, supplies a phenomenological language for nested hypotheses. Its relaxational form applies only to boundary-local dynamics under an approximate gradient description; non-gradient escape requires an action or quasipotential treatment, and whole-night REM-NREM sequencing requires reactive or oscillatory dynamics. Routine polysomnography usually identifies effective combinations rather than curvature, escape action, bias, noise, and relaxation separately. The strongest boundary-level evidence concerns sleep onset, where published results are consistent with bifurcation-like or bistable dynamics but do not yet exclude a driven smooth transition produced by the homeostatic-circadian ramp. The remaining operation classes are hypothesis-generating extensions. The taxonomy is judged by pragmatic utility, including improved communication, stratification, and prediction; the scalar-field implementation and specific dynamical profiles are separately falsifiable. The framework is a phenomenological organizing model rather than a new diagnosis, validated biomarker, or treatment-selection system.

The life sciences and health research have started to benefit from artificial intelligence, which raises ethical concerns that are real but, we argue, not special. Any science should be governed by values that rest on how the human brain is built and socialised rather than anything distinct to artificial intelligence. Importantly, the human brain has a different, much less costly computational architecture than these machines. This is achieved through the orchestration of a global neuronal workspace, and through reward best described not as a quantity to be maximised but as a continuous cycle of wanting, liking and satiety. As such, this creates the deep tension running through the ethics of the human person, between the universality of ethical judgement and the diversity of morals. The brain networks of the global workspace and emotion are universally shared, but the diversity of content is shaped by epigenetic appropriation of the particulars of the physical, social and cultural world, which makes every person unique. Still, if we were to build machines on these principles rather than the present unaffordable reward maximisers, the question of their governance would change from restraint to upbringing. We set out the institutions such a future would require, together with the questions that remain open.

An information theory-based framework is proposed in attempt to explain insistence on sameness in autism as an instance of a general behavior pattern in which an individual tries to reduce surprise and uncertainty. It offers a new definition of autism as an impairment in which cognitive functions are restricted to discrimination, memorization and prediction of tangible properties of the environment. An analogy between insistence on sameness and constrained minimization of the entropy metric is observed and examined for a set of assumptions that describe cognitive limitations of a person with autism. The metric is given by the formula $D_H(R, M) = H(R|M) + H(M|R)$, where $R$ represents sequences of random stimuli, $M$ is a memory that stores and retrieves them, and where $H(.|.)$ denotes their conditional entropies interpreted as surprise and uncertainty, respectively. It is first inferred that to minimize the metric an individual can learn about $R$ (and store that knowledge in $M$) or can restrict $R$ to the already known $M$. Then, it is concluded that insistence on sameness is a manifestation of the latter. Moreover, it is shown that the proposed framework: (1) Helps to quantify the concepts of surprise, uncertainty, sensory overload and deprivation, anxiety, comfort zone, disappointment, disorientation, pedantry, rigidness, observance or aberrant precision. (2) Leads to a list of guidelines for learning therapies and daily care routines, and allows them to be defined as optimization algorithms and implemented as programs for robotic live-in caregivers. (3) Can be validated with the help of a Turing test-like approach that requires no experiments involving individuals with autism. The framework-if positively validated-will provide formal foundations and design guidelines for therapies aimed at improving self-reliance of individuals with autism in basic activities of daily living.

This dissertation presents the first formal theory of metacognitive skill learning. Metacognition, the capacity to monitor and control one's own cognitive processes, has been widely studied, yet the field still lacks a theoretical framework explaining how metacognitive abilities are learned. This gap limits progress in both theory and application across fields such as cognitive science, education, therapeutic practice, and artificial intelligence. The account developed here builds on classic models of skill acquisition from perceptual-motor and cognitive domains. It proposes that metacognitive skill develops primarily through proceduralization, whereby explicit, effortful metacognitive processes are transformed through practice into implicit, automatic routines. The explanatory power of this theory is shown by its ability to unify diverse phenomena, including attentional training, emotion regulation, the metacognitive threshold, and detached mindfulness. The result is an integrated, mechanistic account of metacognitive skill that organizes existing findings, generates testable predictions, informs the design of AI, and supports the strengthening of metacognitive skill in everyday life.

Biological systems must regulate competing needs under limited perceptual bandwidth, where sharpening one estimate costs the capacity to sharpen the others. Any fixed-budget system therefore has to decide where to allocate its perceptual precision. We study this in a foraging agent that must keep several bodily needs satisfied to survive, modelled with active inference. At each step it reads its own body-state beliefs, identifies the most-needed channel, and reallocates a fixed budget of interoceptive precision toward it, so that the same precision-shaped likelihood feeds both belief update and planning. In AffectWorld, a four-channel foraging gridworld, this selective allocation more than doubles learning-phase survival at matched budget against a uniform-precision agent ($0.414$ vs $0.199$ across 11 layouts, $n{=}32$ seeds each, paired cluster-bootstrap $p \leq 10^{-4}$). Two further results sharpen the mechanism. The benefit runs through planning as well as perception, since denying the shaped likelihood to the planner alone removes about half of it. It is also need-aligned, since aiming precision at the least-needed channel does worse than spreading it evenly. The attended channel additionally learns its own dynamics about twice as fast, and stays ahead even at matched observation count, a behavioural trace of the same precision routing, visible in learning speed, not survival.

Whenever we look at an object, we seem to perceive it immediately. However, this is not the case for two reasons. First, it takes hundreds of milliseconds for the brain to process visual information reaching the retina. Second, we have to look at an object for a certain amount of time to perceive it (and we typically look at it for hundreds of milliseconds) – during that time, visual information is continuously received on our retinas. These facts together imply that visual information is both processed and received through time. These two temporal facets of perception, which we term processing time and stimulus time, are often conflated in the literature. Moreover, processing time and stimulus time are usually not considered together in experiments. Here, we argue that, to obtain a more complete portrait of visual perception and constrain further models of vision, it is essential to consider and measure both temporal facets simultaneously. We present a new method designed to do so that is based on reverse correlation: Time^2. We show that this method allows us to precisely characterize many neural phenomena, including rhythmic perception, predictive processing and coarse-to-fine sampling.

Network control theory applied to structural connectomes typically ranks brain regions as candidate driver nodes by their structural connectivity strength, and evaluates performance through scalar control energy. We test whether this framing captures the most relevant information about how driver-node selection shapes brain network control. We introduce an alternative criterion based on the persistent topological cycles in which each node participates—a measure of mesoscale integration that captures features beyond local connectivity—and compare it to standard degree-based selection across 70 human structural connectomes at three parcellation scales. Topology- and degree-informed driver sets achieve nearly identical scalar control energy, differing by approximately 0.2%. The geometry of the controllable subspace, however, differs substantially: topology-informed sets distribute controllability across more dimensions of state space and produce better-conditioned controllability matrices. This geometric advantage is preserved when high-degree hub nodes are removed, and it carries a functional signature: because the two criteria place driver nodes in different cortical territory, each most efficiently reaches a different class of target state. The choice of node-ranking criterion therefore shapes which brain-state transitions are energetically favored even when average control cost is unchanged. The results reveal a dissociation between control cost and control geometry, and demonstrate that persistent topology captures information about brain network control that scalar energy summaries miss.

Sleep staging provides a reproducible clinical description, but it does not by itself explain why some boundaries are abrupt while others are graded, or why transition windows contain instability, synchrony, and apparent state coexistence. We develop a local Landau-Ginzburg phenomenology in which each boundary is represented by motion in an effective potential of a spatially extended, noisy, dissipative neural field. A latent cortical-ordering coordinate phi is inferred from prespecified EEG/PSG observables through a measurement model designed to avoid circularity. The canonical boundaries are treated separately. Existing data support a fold-like loss of wake stability at sleep onset; whether that fold lies on a globally bistable cusp with hysteresis remains open. N1-to-N2 and N2-to-N3 are posed as continuous-like ordering crossovers, NREM-to-REM as a candidate first-order-like desynchronizing switch, and a possible within-N3 mixed or tricritical-like regime as a speculative hypothesis. The Ginzburg term adds spatial predictions - growth of correlation length and local-to-global recruitment - that are absent from scalar sleep-onset models. We specify the evidence needed to distinguish bifurcation, coexistence, noise-driven escape, smooth crossover, and scoring-induced discontinuity. Illustrative time-dependent Ginzburg-Landau simulations reproduce the proposed signature classes. A synthetic classification experiment partially distinguished six archetypes (cross-validated accuracy 0.49 +/- 0.005; balanced baseline 0.17), with little change under a noise-regime shift. These analyses establish the internal consistency and testability of the framework, not the proposed taxonomy in human sleep. Transition-centered EEG validation is required before clinical or neuromodulation applications are pursued.

Detecting structure in complex networks, especially those arising from physical systems, is a central problem across the sciences. One approach is via rich club analysis, which identifies important vertices using a centrality metric and measures whether those vertices are more tightly interconnected than expected by chance. While informative, this approach captures only pairwise interactions, missing out on higher-order ones known to shape the structure and function of many complex systems. We propose a hyper-rich club pipeline that asks whether central vertices are more tightly interconnected than expected by chance through hyperedges encoding higher-order interactions, which also enables the inclusion of important, often omitted, directional information. We work in a broad class of hypergraphs, which we call general directed hypergraphs, that includes as special cases undirected hypergraphs, head-and-tail directed hypergraphs, and totally ordered hypergraphs (a hypergraph related to directed simplicial complexes from topological data analysis). This unifies several non-equivalent notions of directed hypergraph under one definition. On these hypergraphs we define a hyper-rich club framework whose concrete construction depends on explicit choices the domain scientist fixes according to their research goals. Particular choices recover the existing rich club notions for graphs and undirected hypergraphs, and yield the first such notion for each version of directed hypergraphs. We demonstrate that the pipeline recovers meaningful structure in data by studying networks of very different origins: connectomes, temporal networks of infectious spread, networks of poems, and the XGI hypergraph database, in each case detecting structure the standard graph rich club misses.

Understanding the dynamics of suicidal ideation and behaviour in youth and the factors associated with transitions from thoughts to behaviours is critical for early identification, monitoring, and prevention. Using longitudinal self-report data from the Adolescent Brain Cognitive Development (ABCD) Study (n = 11,864) spanning ages 9 to 13 years, we developed a time-inhomogeneous discrete-time Markov chain framework to model transitions across eight states defined by suicidal ideation and behaviour and the co-reported presence or absence of non-suicidal self-injury (NSSI). The framework enables inference of year-to-year and multi-year transition probabilities, quantification of transition stability and uncertainty, and statistical comparison of transition likelihoods. We identified structured but developmentally changing transition patterns, including a generally high probability of recovery to a no-report state following reports of suicidal ideation or behaviour. Co-reported lifetime NSSI marked a distinct trajectory profile characterized by both greater risk and lower predictability: children with NSSI were more likely to transition to or persist in suicidal behaviour, were less likely to recover to a no-report state, and exhibited greater uncertainty in their transition likelihoods compared to those reporting suicidal ideation and/or behaviour alone. These findings suggest that NSSI marks not only elevated risk but also greater trajectory instability during pre- to early adolescence. By estimating probabilities of escalation, remission, persistence, and fluctuation in longitudinal cohort data, the framework provides a systematic, interpretable, and scalable approach to characterising suicidal trajectory dynamics in large, sparse mental health datasets, with potential to inform future research on monitoring, early prevention, and time-varying risk in youth mental health.

Recent studies have reported clearly identifiable dynamical changes in the high-frequency range of EEG signals recorded during specific stimuli, such as visual or auditory inputs, or in cases of brain disorders like epileptic seizures. In this study, we utilized Dynamic Mode Decomposition (DMD) to extract consistent and persistent dynamical changes in the high-frequency band from the signals of neurologically relevant EEG channels. High-frequency DMD modes were employed as features, composing a feature table. Through post-processing, a random distribution test was performed, revealing that approximately 70% of the samples exhibited consistent high-frequency dynamics within the signal of a specific channel. Furthermore, classification experiments confirmed that the PCA components of the feature table that passed the test formed a consistent pattern that distinguished the alcohol-dependent group from the control group.

Predictive processing theories portray the brain as a hierarchical prediction engine that minimizes prediction error, yet they lack operational definitions for the structure of a "prediction," the standardized response to a prediction error, and the mechanism that maintains consistency across successive updates. Bayesian cognitive science attempts to subsume all uncertainty under probabilistic belief updating, but it presupposes a closed hypothesis space and provides no generative account of how the objects over which probabilities are distributed become discrete, identifiable referents in the first place. This paper introduces Predictive Set Theory (PST), a formal generative framework that reconstructs cognitive architecture from first principles. PST anchors cognition in a minimal set of operations—a sensor formalized as an identity function, set-theoretic state refresh, and three fundamental forms of reference chains (reference, counter-reference, and semi-reference)—and rigorously derives core cognitive functions including state sequences, demand, comparison, efficiency, and finite-horizon probabilistic planning. Rather than modeling neural mechanisms, PST constitutes a design specification for any system that must maintain internal consistency while acting under incomplete information and irreversible risk. The framework offers novel resolutions to classical problems such as Russell's paradox, the cognitive status of Gödelian incompleteness, the grounding of negative feedback, and the comprehension of film editing. The primary purpose of this paper is to establish, through the public academic record, the originality and completeness of the Predictive Set Theory framework.

The ability to robustly maintain and update continuous variables is a hallmark of working memory. While classical continuous attractor networks suffer from severe fine-tuning fragility, standard artificial recurrent neural networks (RNNs) like GRUs and LSTMs typically fail to stably learn continuous manifolds, instead shattering the state space into discretized point attractors. To bridge this gap, we draw inspiration from divisive normalization, a canonical neural computation widely observed across cortical circuits, and propose the Recurrent Divisive Normalization Network (RDNN), a minimal and algebraically isolated model of dynamic division. Through dynamical systems analysis on canonical working memory tasks, we demonstrate that this biophysical constraint allows the network to converge to robust, high-fidelity slow manifolds. Furthermore, we analyze the gradient dynamics of divisive normalization during Backpropagation Through Time (BPTT), showing that it introduces an activity-dependent local gradient scaling. This scaling dampens parameter updates in highly active regimes, which empirically aligns with a significant self-compression of the network's effective rank, confining the recurrent dynamics to a tight, low-dimensional subspace while avoiding the optimization pathologies associated with explicit low-rank factorization. Finally, ablations demonstrate that while subtractive inhibition can maintain static memories, divisive normalization is mathematically essential to prevent manifold shattering under time-varying inputs. Our findings identify divisive normalization not merely as a biological artifact, but as a critical computational mechanism for learning high-fidelity continuous representations.

Forecasting human brain activity during naturalistic experience requires modeling how endogenous neural states evolve causally under continuous sensory drive. Existing brain encoding models instead frame this as stimulus-to-response regression without strict temporal constraints, allowing future stimuli to leak into current predictions. We introduce NeuroWorld, to our knowledge the first brain world model, which casts naturalistic brain functional dynamics prediction as stimulus-conditioned evolution in a learned latent brain-state space, separating endogenous states (measured via fMRI) from exogenous multimodal stimuli across two stages. Latent Dynamics Learning (LDL) jointly learns a transition-sufficient representation and causal dynamics through next-latent prediction, without reconstructing the observed fMRI signal. Latent Rollout Decoding (LRD) freezes LDL, autoregressively rolls latent states forward from an observed fMRI prefix, and decodes them into subject-specific whole-brain responses. Across three naturalistic movie-fMRI benchmarks spanning 30 participants, including our newly collected Singapore Multimodal Imaging & Naturalistic Dataset (SG-MIND; 20 participants, 8,519 paired stimulus-response clips, 140.7 person-hours of viewing), NeuroWorld achieves state-of-the-art multi-step rollout performance under strictly causal stimulus access, with greater robustness to long-horizon autoregressive drift, supporting reliable simulation of extended brain-state trajectories. Extensive interpretability analyses characterize the functional organization of the learned dynamics, establishing latent-space world modeling as a principled framework for causal forecasting of human brain activity.