Flexible categorization in perceptual decision making

Abstract Perceptual decisions rely on accumulating sensory evidence. This computation has been studied using either drift diffusion models or neurobiological network exhibiting winner-take-all attractor dynamics. Although both can account for a large amount of data, it remains unclear whether their dynamics are qualitatively equivalent. Here we show that in the model, but not an increase stimulus fluctuations duration promotes transitions between decision states. The number leads to crossover weighting mostly early evidence (primacy) late (recency), prediction validate with psychophysical data. Between these two limiting cases, found novel flexible categorization regime, which reverse initially-incorrect categorizations. reversal asymmetry results non-monotonic psychometric curve, distinctive feature model. Our findings point correcting reversals as important perceptual making.