Making decisions with unknown sensory reliability

• Main Author: Sophie eDeneve
• Format: Article
• Language: English
• Published: Frontiers Media S.A. 2012-06-01
• Series: Frontiers in Neuroscience
• Subjects: Decision Making; adaptation; Prior; Bayesian; uncertainty; expectation-maximization
• Online Access: http://journal.frontiersin.org/Journal/10.3389/fnins.2012.00075/full

To make fast and accurate behavioural choices, we need to integrate the noisy sensory input, take into account prior knowledge, and adjust our decision criteria. It was shown previously that in a two alternative forced choice tasks, optimal decision making can be formalized in the framework of a sequential probability ratio test and is then equivalent to a diffusion model. However, this analogy hides a chicken and egg problem: To know how quickly we should integrate the sensory input and to set the optimal decision threshold, the reliability of the sensory observations has to be known in advance. Most of the time, we cannot know this reliability without first observing the decision outcome. We consider here a Bayesian decision model simultaneously inferring the probability of two different choice alternatives and estimating at the same time the reliability of the sensory information on which this choice is based. We show that this can be done within a single trial, based on the noisy responses of sensory spiking neurons. The resulting model is a non-linear diffusion to bound where the weight of the sensory inputs and the decision threshold are both dynamically changing over time. In difficult decision trials, sensory inputs early in the trial have a stronger impact on the decision, and the threshold collapse such that choices are made faster but with low accuracy. The reverse is true in easy trial: the sensory weight and the threshold increase over time, leading to slower decisions but at much higher accuracy. In contrast to standard diffusion model, adaptive sensory weights construct an accurate representation for the probability of each choice. This information can thus be appropriately combined with other unreliable cues, such as priors. We show that this model can account for recent findings in a motion discrimination task, and can be implemented in a neural architecture by fast Hebbian learning.