One difference is that the tasks involving NE typically have rare targets (perhaps boosting unexpectedness), whereas those involving ACh have common targets. Epacadostat price It would be interesting to record phasic NE and ACh signals simultaneously (perhaps indirectly in human subjects via pupil dilation; Gilzenrat et al., 2010)—one might expect that NE would be released to the cue, as a temporal alert, but that it is the phasic rise in ACh that prepares the ground for the

(now expected) reward to be delivered. Particularly for the case of DA (Servan-Schreiber et al., 1990) and NE (Brown et al., 2005), there has been work on how an effect of these neuromodulators on the input-output gain of neurons might influence overall network dynamics that implement inferences such as decision making. One of the simplest decision making networks involves effective mutual inhibition between two competing groups of neurons (Usher and McClelland, 2001), with action initiation occurring when the activity of one group reaches a threshold (Bogacz et al., 2006; Gold and Shadlen, 2002; Lo and Wang, 2006). Boosting the gain of the neurons in such a network can make it unstable and therefore allow whichever of the two groups currently has the greater activity to reach the threshold promptly, with barely any further integration.

This therefore controls a speed-accuracy tradeoff. Brown et al. (2005) considered the problem of decision making architectures in which one network determines Rapamycin purchase the

release of NE, which then modulates another network that is more directly responsible for initiating the decision. They pointed out what is a general issue for phasic activity (U), namely that the time it takes for the neuromodulator to be delivered to its site of action (norepinephrine fibers are not myelinated) appears to be at the margins of the period in which there is a chance to have a suitable effect on the on-going computation. Unlike utility, which seems a natural candidate for neuromodulatory realizations, oxyclozanide uncertainty does not, because of the exquisite selectivity that subjects should exhibit in their sensitivity to uncertainty. Nevertheless, substantial evidence suggests the involvement particularly of acetylcholine and norepinephrine in representing and acting on uncertainty, and we have also seen that there are rich links between these neuromodulators and also with dopamine. Many of the general lessons that we learnt for utility have been reiterated, and some new ones learned, particularly concerning the overall architecture of influences. This review has considered general properties of neuromodulators through the lens of effects on decision making. The latter is a critical competence, and we have seen the rich involvement of very many aspects of neuromodulation.