It has long been shown that during perceptual decision tasks, the activity of sensory neurons correlates with a subject’s percept, even when the physical stimulus is identical. This correlation can be used to predict a subject’s choice at the end of a trial from the activity of these sensory neurons during the trial. How well one can predict the choice depends on the strength of the neuron-behavior correlation, and is commonly quantified as a neuron’s “choice-probability”. The existence of choice-probability in sensory neurons is often seen, at least implicitly, as evidence that these sensory neurons play a causal role in the decision. A widely held view is that these neuron-behavior correlations, i.e. choice probabilities, reflect the causal effect of feed forward noise in these sensory neurons on the decision. However, choice probabilities could also result from different brain-states associated with the perceptual choice (top-down effects). In the first part of my talk, I will present evidence that challenges the feed forward causal explanation. We trained monkeys to perform a binocular disparity discrimination task. While they performed the task, we recorded the activity of disparity-selective neurons in their V2. We used white-noise analysis to measure tuning-functions of V2 neurons associated with choice and simultaneously measure how the variation in the stimulus affects the monkeys’ perceptual decisions. In causal models stronger effects of the stimulus upon decisions, mediated by sensory neurons, are associated with stronger choice-related activity. However, we find that over the timecourse of the trial, these measures change in opposite directions—at odds with causal models. An analysis of effect of reward size supports the same conclusion. Finally, choice was associated with changes in neuronal gain that are incompatible with causal models. All three results are readily explained if choice is associated with changes in neuronal gain caused by top-down phenomena that closely resemble attention. We conclude that top-down processes contribute to neuron-behavior correlations. In the second part of my talk, I will present data suggesting that the organization of top-down signals may require a cortical map for the features relevant for the perceptual task. This would explain why studies using discrimination tasks based on binocular disparity in area V1 have not found these correlations, as V1 contains no map for binocular disparity. A prediction of this scheme is that activity of V1 neurons correlates with decisions in an orientation discrimination task. To test this prediction, we trained two macaque monkeys in a coarse orientation discriminination task of orientation band-pass filtered dynamic noise. The two orientations were always 90 degrees apart, and task difficulty was controlled by varying the orientation bandwidth of the filter. While the trained animals performed this task, we recorded from orientation selective V1 neurons. For both monkeys we observed neuron-behavior correlations (mean choice probability 0.55). The mean choice probability in each monkey exceeded chance level (p<0.05). These results confirm the prediction from our proposed scheme. Choice probabilities are only observed for neurons that are organized in maps for the task-relevant feature. We speculate on a novel function for the cortical columnar architecture: efficient wiring of top-down signals.