51 papers
RECAP is a bio-inspired image classification method that couples untrained reservoir dynamics with a self-organizing Hebbian prototype readout to achieve robust, backpropagation-free learning capable of generalizing to corrupted inputs without prior exposure.
This study demonstrates that the choice of musical feature encoding fundamentally shapes network representations of music, revealing a critical trade-off where compressed single-feature models offer high uncertainty but low learning error, while richer multi-feature models preserve fine distinctions at the cost of increased state space complexity and higher model error, thereby determining how plausibly these networks reflect realistic listener expectations.
This paper argues that the interaction between vulnerable individuals with mental health conditions and AI chatbots creates dangerous feedback loops—driven by human cognitive biases and AI sycophancy—that can destabilize beliefs and foster dependence, necessitating urgent coordinated action across clinical, developmental, and regulatory domains to mitigate emerging public health risks.
This paper introduces MIG-Vis, a method combining variational autoencoders with mutual information-guided diffusion models to directly visualize and validate that neural populations in the macaque inferior temporal cortex are organized into structured, semantically meaningful latent groups encoding specific visual features like object pose and category transformations.
This study demonstrates that a cross-species transfer learning approach, utilizing an attention-based BiLSTM model pretrained on mouse Patch-seq data and fine-tuned on human data, successfully improves the prediction of conserved GABAergic interneuron subclasses from electrophysiological recordings to transcriptomic identities.
This paper demonstrates that Restricted Boltzmann Machines can effectively model large-scale neural activity from approximately 1,500 to 2,000 simultaneously recorded neurons, capturing complex higher-order statistics and revealing anatomically structured interaction networks that align with visual behavior and global dynamics.
This paper presents a statistical-mechanical theory linking the linear decoding efficiency of continuous variables to the geometric properties of neural manifolds, revealing increasing decoding capacity for object position and size along the monkey visual stream.
The paper introduces JEDI, a hierarchical model that jointly learns shared embeddings over recurrent neural network weights to robustly infer generalizable, task-specific neural dynamics from noisy, high-dimensional experimental recordings, successfully recovering underlying mechanistic structures and providing insights into motor control.
This paper introduces behavior-decomposed linear dynamical systems (b-dLDS), a novel modeling approach that disentangles behavior-related neural dynamics from internal computations in large-scale brain recordings, demonstrating superior performance over existing supervised models and successfully scaling to tens of thousands of neurons in zebrafish hindbrain data.
This paper introduces CODEC, a method that utilizes sparse autoencoders to decompose neural network computations into sparse, causal motifs of hidden-neuron contributions, thereby revealing how nonlinear processes evolve across layers and enabling greater interpretability and control of both artificial and biological neural systems.
The paper introduces CITS, a nonparametric statistical framework that overcomes the limitations of existing causal inference tools to accurately identify interpretable causal networks in high-resolution neural time series, as validated by both theoretical proofs and applications to large-scale mouse brain recordings.
This study reveals that the human hippocampus encodes positions of self, others, and gaze within geometrically related, mostly orthogonal neural subspaces that can be flexibly aligned via linear transformations, thereby enabling reliable individuation and generalization across different agents and viewpoints.
INTENSE is an open-source, information-theoretic framework that robustly detects and disentangles neuronal selectivity from calcium imaging data by accounting for temporal autocorrelation, indicator kinetics, and behavioral covariance to link large-scale neural activity with complex behaviors.
This paper demonstrates that a parameter-free biophysical model, derived from Shannon's information theory and the principle of minimal energy, accurately predicts how synaptic information efficiency (bits per Joule) decreases as conductance deviates from its natural value, confirming that synapses operate at signal-to-noise ratios that maximize energy efficiency.
This review synthesizes two decades of theoretical and experimental research to argue that grid cells function as a biologically plausible, high-fidelity, non-linearly decodable code for path integration, offering valuable lessons for developing normative theories of neural computation.
This paper introduces a tractable and flexible normative framework for neural coding by optimizing representational similarity rather than direct neural activity, demonstrating that a broad class of such problems is convex and using this property to derive new results on model identifiability, the uniqueness of neural tunings, and the optimal structure of ON/OFF channels in retinal versus cortical codes.
This study presents a computational pipeline using multichannel EMG signals and machine learning classifiers to predict motor intentions with high spatial and temporal resolution across 25 targets, achieving up to 80% accuracy and demonstrating the potential for anticipatory control in adaptive rehabilitation systems.
This paper proposes that memory consolidation serves a computational role beyond mere stabilization, utilizing "predictive forgetting" to compress stored representations into a form that optimizes generalization by selectively retaining information that predicts future outcomes, a process necessitated by high-capacity encoding constraints and validated through simulations across diverse neural and transformer models.
This paper proposes a systematic approach to Chalmers' "easy problems" of consciousness by demonstrating how an executable cognitive system, grounded in Kantian conceptual knowledge, can computationally derive abilities such as discrimination, attention, and reportability through its learning, emotional, and information-manipulation mechanisms.
CytoNet is a foundation model trained on a million unlabeled histological patches from ten human brains that encodes complex cellular patterns into meaningful features, enabling scalable analysis of cortical microarchitecture and linking cellular structure to functional organization through various downstream applications.