Hamish Meffin A

This study shows that a biologically grounded spiking neural network model of V1 learning exhibits sparse coding, balance, and decorrelation as emergent properties.

This paper presents a spiking network model of Layer 2/3 in V1 that computes sign-specific prediction errors through compartmentalized neurons without synaptic weight updates.

This paper develops a spiking neural network model of V1 layer 4 to study predictive coding, showing it can reproduce in vivo features and maintain E-I balance under varying input conditions.

This study examines how changes in stimulus statistics affect V1 neuron receptive fields, finding increased nonlinear processing and additional filters under natural scenes.

This paper proposes a biologically plausible model for learning complex cells in V1 using BCM rule-based plasticity, showing it can replicate experimental diversity in receptive field properties.

The paper was withdrawn due to a duplicate posting and directs readers to the current version on bioRxiv.

This study investigates how selectivity and invariance in primary visual cortex cells are balanced, revealing that phase invariance is partial and invariance to orientation and spatial frequency is more extensive than previously thought.