Alan A. Stocker

This study investigates how temporal structure affects evidence accumulation in human decision-making, finding that subjects dynamically adjust their strategies to compensate for memory loss during breaks.

The paper shows that efficient ensemble coding depends on the simultaneous presentation of reference and stimulus ensemble, with reduced coding and decision accuracy when they are presented asynchronously.

The paper shows that brain activity in the parietal cortex encodes evidence consistently with prior choices, leading to confirmation bias in decision-making.

The paper shows that heading estimation biases from optic flow depend on response range size and introduces a Bayesian model to explain these biases.

The paper shows that the tilt illusion results from increased neural coding precision for stimuli similar to an oriented surround, using psychophysical and fMRI data from humans.

The paper shows that sensory adaptation optimally reallocates coding resources to maximize encoding accuracy for future expected stimuli based on natural input statistics.

The paper shows that gradient descent learning in neural networks creates representations sensitive to common image features, mirroring efficient coding principles observed in animal sensory systems.

The paper shows that a Bayesian model constrained by efficient coding explains both human visual speed perception and neural tuning in area MT, with priors from behavior matching those from neural data.

This study shows that individuals with ASD have reduced sensory encoding capacity and limited adaptability to changing visual stimuli compared to controls.

The paper clarifies a decoding model proposed by Ding et al. and shows that using high-level information to constrain low-level decoding does not improve perceptual performance.