Renato Duarte

This study shows dendritic nonlinearities, particularly NMDA receptors, can drive Up-Down states in cortical neurons independently of network activity.

The paper develops a multi-layer model of somatosensory cortex with three interneuron types and cell-type-specific short-term plasticity, reproducing experimental observations on network dynamics and inhibitory effects.

The paper shows dendritic networks with STDP can form and reactivate sequential memories based on phoneme order, outperforming point neuron networks.

The paper shows that topographic modularity in neural circuits improves signal-to-noise ratio and task performance through recurrent inhibition.

The paper replicates and analyzes a modular spiking network with reward-based learning, highlighting its robustness and limitations under different conditions.

The paper introduces the Tripod neuron model, a simplified dendritic structure that captures complex dendritic computations like coincidence detection and temporal binding.

This paper introduces a spiking neural network architecture combining unsupervised learning and clustered connectivity, showing improved performance on reinforcement learning tasks compared to unclustered or static input networks.

The paper investigates how a spiking network model of sentence comprehension uses neuronal memory to assign semantic roles and resolve binding relations in language input without relying on persistent activity or synaptic plasticity.

This study investigates how structured connectivity in modular spiking networks enables efficient and accurate information transfer between modules during stimulus processing.

The paper investigates how heterogeneity in neuronal and synaptic properties enhances computational power in cortical microcircuits, showing that heterogeneous models outperform homogeneous ones.