Of particular interest however was a response that preceded the actual cue and seemed to signal the expected information. This response arose at the time of the monkeys’ selection (marked “Target” in Figure 3B) and was slightly stronger if the trial BMN673 included an informative rather than an uninformative cue (red versus blue traces). This early response seems to signal a superordinate property of “informativeness” (or reliability) that is independent of a specific message,

and to correspond to the monkeys’ behavioral preference for the informative cue. Unfortunately however, because the information in this task was about a primary reward, the results do not conclusively rule out alternative explanations based on this reward. It is well known that monkeys modulate their anticipatory licking based on stimulus-reward associations and will stop licking when observing a low-reward cue ( Fiorillo et al., 2003). In addition as I mentioned above, subjects

direct attention based on stimulus-reward associations, and may have gazed for longer periods at the high-reward versus the low-reward cue (e.g., the green cross versus green wave in Figure 3A; Hogarth et al., 2010). It remains therefore possible that by selecting the informative cue the monkeys did not specifically seek information but simply sought to minimize their effort (by avoiding having to lick AZD2281 concentration for or look at a low-reward pattern) or perhaps to bring about the motivationally salient, high-reward pattern ( Beierholm and Dayan, 2010). At this time therefore it remains an open question whether

the brain has a bona fide reliability representation. Rather than searching for an “intrinsic” preference for information, studies of eye movements in natural behaviors have adopted a more pragmatic approach and attempt to estimate the material value that an eye movement may bring (Hayhoe and Ballard, 2005; Tatler et al., 2011). The studies make use of so-called Markov decision chains—mathematical methods that allow one to formulate a task description as a sequence of steps and estimate the cumulative future reward that can be expected by traversing these steps. By including an estimate of the uncertainty that arises at each step, one can crotamiton further calculate the costs of this uncertainty and the benefits of reducing it by obtaining information (Dayan and Daw, 2008; Rothkopf and Ballard, 2010; Sprague and Ballard, 2005; Tatler et al., 2011). For instance, in the tea-making task, one can calculate how uncertain one is about one’s position and distance from the faucet, and what the benefit would be of reducing that uncertainty through a shift of gaze. These studies have shown how, when applied to complex tasks (such as an agent walking through an environment while avoiding obstacles and picking up litter) these methods can be applied to identify the uncertainty and informational requirements of intermediate steps (Rothkopf and Ballard, 2010; Sprague and Ballard, 2005).