Many studies have reported correlations between the activity of sensory neurons and animals'judgments in discrimination tasks (Choice Probability, or CP). Although this demonstrates a link between the activity of individual sensory neurons and perception, it also raises new questions. In particular, it is not understood why neurons in some areas show decision-related activity whereas equally selective and informative neurons in other areas do not . A common pattern emerged from the study of the existing literature : cortical neurons that displayed decision-related activity were organized in clusters or maps for the stimulus variable that was reported. Combined with our earlier work showing that CP is derived from top-down signals, this suggests a new functional role for the columnar organization of the cerebral cortex the map like organization allows top-down signals to be efficiently routed to neurons with similar properties. These require that feedback connections influence the activity of a group of neurons that share a certain feature selectivity without affecting other neurons. If neurons with similar selectivities are grouped together, this greatly simplifies the wiring problem: a single feedback connection could modulate a group of similar neurons very straightforwardly. This could dramatically reduce the number of fibers the brain needs to alter the firing of sensory neurons in a context-dependent way and may explain why the brain evolved map-like organization in much of cortex. This principle has implications not only for choice-related activity, but also for the organization of attentional signals. This scheme makes the striking prediction that activity of V1 neurons should correlate with decisions in an orientation-discrimination task. This is striking because no CP has been shown in V1 for any task. To test this prediction, we trained two macaque monkeys in a coarse orientation discrimination task using band-pass-filtered dynamic noise. The two orientations were always 90 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 (n = 82, n = 31 for Monkey 1, n = 51 for Monkey 2). For both monkeys, we observed significant correlation of the V1 activity with the monkeys'perceptual judgments. In one of these animals, we had previously measured choice probabilities in a disparity discrimination task in V1, which had been at chance. The choice probabilities in this monkey for the orientation discrimination task were significantly larger than those for the disparity discrimination task. These results are provide strong support for the idea that choice probabilities depend upon a columnar organizations.
Showing the most recent 10 out of 36 publications
