Pieter R. Roelfsema

This paper proposes a recurrent neural network architecture that learns to trace curves and objects using multiscale, incremental grouping, showing generalization from short to longer curves and 2D-objects.

This study compares tissue reactivity to stiff silicon and flexible polyimide implants in mice, finding that flexible polyimide reduces lesions and immune response.

This paper identifies a convergent coding scheme in the primary visual cortex that enables visual scene segmentation by creating an orthogonal representational geometry for object recognition.

This paper introduces a biologically plausible neural network that uses modulatory feedback to perform scene segmentation and object-based attention, replicating neurophysiological signatures of recurrent processing.

This study develops an end-to-end framework to learn safe stimulation parameters for cortical neuroprosthetic vision, improving phosphene interpretability while maintaining image reconstruction performance.

This paper develops an automatic method to optimize electrode placement for visual prostheses by matching phosphene distributions using retinotopic predictions and individual anatomy data.

This study examines cerebellar dentate nucleus neuronal activity during a visual attention task in rhesus macaques, revealing dynamic encoding of visual stimuli and task-related behavior.

The paper uses deep generative networks to infer visual neuron tuning properties in monkeys and demonstrates their influence on visual perception through microstimulation.

The paper shows that pronouns reactivate noun-specific neurons in the human hippocampus during reading, linking memory and language at the single-cell level.

The study shows that the direct and indirect pathways of the basal ganglia have opposing effects on perceptual decisions and cortical activity in mice during visual detection and working memory tasks.

The paper shows how visual cortex neurons adjust their responses to direct inputs and contextual signals through opposing plasticity mechanisms, enhancing improbable inputs or omissions.

This paper shows that feature-disentangled GAN latents better predict neural responses to visual stimuli than other representations, and can reconstruct visual perception with high fidelity.

This paper trains a recurrent neural network with reinforcement learning to solve multi-step visual tasks, showing it can learn elemental operations and propagate enhanced activity between subroutines.

This paper presents a differentiable phosphene simulator for optimizing visual cortical prostheses using biologically plausible models of visual perception.

This study shows that the superior colliculus in mice is causally involved in figure detection based on contrast, orientation, and phase, with stronger neural responses to figure elements than background elements.

This study shows that chandelier cells in the visual cortex change their function and structure with experience, responding more to arousal-related events and showing plasticity in their connections.

This paper introduces a biologically plausible recurrent neural network called RECOLLECT that can learn to remember and forget information over time, demonstrating its ability to perform tasks similar to animal learning.

The paper demonstrates a cortical mechanism where visual search suppresses salient distractors by inverting their attentional signal into suppression, using behavioral and neuronal data from monkeys.

The paper shows that V1 neurons in monkeys and humans respond to occluded scenes using contextual information, with representations resembling those in humans.

The paper compares fMRI-based population receptive field (pRF) models with neurophysiological recordings in non-human primates to determine what neuronal signal pRFs represent.

The paper investigates how objecthood in natural images influences neuronal activity in early visual cortex, showing that object boundaries and interiors elicit more activity than background.

The paper shows that human single neurons encode sequence order through phase-locked spikes relative to theta oscillations, similar to rodent phase precession.

This study investigates Purkinje cell activity in the medial and lateral cerebellum of rhesus macaques during antisaccade tasks, revealing distinct roles in volitional eye movement suppression.

The paper investigates how spatial attention spreads to enhance object representations in V1, showing attentional selection operates at an object level even when objects are not task-relevant.

This paper introduces WorkMATe, a neural network model that learns to control working memory through reinforcement learning, solving both simple and complex tasks with flexible memory management.

This study shows that delayed V1 neuronal responses are necessary for figure-ground segregation in mice, with VIP interneurons playing a key role through disinhibitory circuits.

This paper reports difficulties in measuring acetylcholine release in awake macaques using amperometry during a cognitive task, highlighting challenges for future research.