Gabriel Kreiman

The paper identifies a localized, invariant neural representation of parts of speech in the human brain using invasive recordings from epilepsy patients.

The paper identifies order-selective cells in the human brain that represent temporal order during memory encoding and retrieval, predicting memory accuracy.

This study uses ECoG data to show that phase locking between brain regions varies with object categories and influences perceptual behavior, with global connections in the frontal lobe and local connections in the occipital lobe being more behaviorally relevant.

The paper shows that single-neuron activity in occipital and medial temporal lobe regions encodes newly learned visual information, reshaping perception after rapid learning.

The paper predicts neuronal activity in one visual cortical area from another in single trials, with or without visual stimulation, in mice and macaques.

This paper introduces a theory where policy optimization emerges from reward-modulated noise in a model of representation learning.

This paper demonstrates that neural networks can exhibit reward-seeking behavior through local predictive updates and internal noise, enabling autonomous and adaptable interaction with environments.

The paper shows that ventral visual cortex responses in monkeys remain retinotopic during natural vision, tracking individual fixations rather than becoming gaze-invariant.

The paper reports on temporally periodic neuronal firing in the human entorhinal cortex during continuous audiovisual experiences, showing invariance across different playback speeds.

A preregistered multimodal neuroimaging study across 256 human participants directly tests Integrated Information Theory against Global Neuronal Workspace Theory, confirming some predictions while revealing major empirical challenges for both.

This study shows that phase precession in human hippocampal theta oscillations correlates with memory encoding and retrieval success during naturalistic experiences.

This study investigates dynamic neural processes during complex human memory using a card-matching game and intracranial recordings, revealing gamma-band activity related to novelty, familiarity, and associative recall.

The paper decodes human identity from neuronal activity using computer vision and deep learning, finding that tens of neurons in brain regions beyond MTL can robustly represent concepts.

This study compares human and rhesus macaque visual object recognition in different contexts, finding similar behavioral patterns but highlighting gaps in explaining these with neural data or models.

This paper demonstrates a hybrid system combining optogenetics and deep reinforcement learning to enhance nematode navigation and food search performance.

This paper introduces a model for continual learning in a recurrent neural network that combines Hebbian learning, synaptic decay, and a novel memory consolidation mechanism to avoid catastrophic forgetting.

This study identifies neurons in the human medial temporal lobe that signal cognitive boundaries between episodic events and shows these signals correlate with memory accuracy and temporal order recall.

This study investigates neural processes during conflict resolution in cognitive control tasks and finds task-specific neural signatures that do not generalize across tasks.

This study introduces a deep neural network model with a fatigue mechanism to capture adaptation dynamics in neurophysiology and perception, showing it accounts for neural and behavioral adaptation properties.

The paper demonstrates zero-shot decoding of novel visual categories using neural responses from macaque inferior temporal cortex mapped onto a deep neural network.

This paper evolves synthetic stimuli using deep generative networks and genetic algorithms to maximize neuronal firing in macaque visual cortex, finding that these stimuli outperform natural images.