Figure 4. Functional integration of uncertainty.
Cool and geeky article that takes an information theory approach to explicate the neuroscience of uncertainty in making decisions (and how the brain integrates the various processes). This is a PLoS ONE open source article, so the whole thing is freely available online and as a PDF.
The Neural Substrate and Functional Integration of Uncertainty in Decision Making: An Information Theory Approach
1 Neuroimaging Laboratory, Department of Neurosciences, Center for Applied Medical Research, University of Navarra, Pamplona, Spain, 2 IESE Business School, University of Navarra, Barcelona, Spain
Abstract
Decision making can be regarded as the outcome of cognitive processes leading to the selection of a course of action among several alternatives. Borrowing a central measurement from information theory, Shannon entropy, we quantified the uncertainties produced by decisions of participants within an economic decision task under different configurations of reward probability and time. These descriptors were used to obtain blood oxygen level-dependent (BOLD) signal correlates of uncertainty and two clusters codifying the Shannon entropy of task configurations were identified: a large cluster including parts of the right middle cingulate cortex (MCC) and left and right pre-supplementary motor areas (pre-SMA) and a small cluster at the left anterior thalamus. Subsequent functional connectivity analyses using the psycho-physiological interactions model identified areas involved in the functional integration of uncertainty. Results indicate that clusters mostly located at frontal and temporal cortices experienced an increased connectivity with the right MCC and left and right pre-SMA as the uncertainty was higher. Furthermore, pre-SMA was also functionally connected to a rich set of areas, most of them associative areas located at occipital and parietal lobes. This study provides a map of the human brain segregation and integration (i.e., neural substrate and functional connectivity respectively) of the uncertainty associated to an economic decision making paradigm.
Here is a little taste from the introduction to whet your interest.
Full Citation:Introduction
Consider an economic decision paradigm with two options. The first option (A) is constant and consists of winning
euros after
month with
% of probability, while the second option (B) can consist, for instance, of winning the same amount of money after
months with
% of probability. Some people would prefer the first -closer in time but riskier- option and some others would prefer the second -delayed in time but safer- option. When varying the probability and the time of option B, one could find a task configuration where both options are evaluated as highly similar in terms of attractiveness. This kind of situation gives rise to a decision conflict. Different task configurations might produce heterogeneous decision patterns covering from a predominance of option A to a predominance of option B. Briefly, task configurations that produce a predominant answer (either option A or B) can be characterized by a low uncertainty, while task configurations with a balanced number of A and B outcomes can be characterized by a high uncertainty. The variability of the outcomes comes from within- and inter-subject variabilities. The former happens when decisions of a subject for certain configuration are not self-consistent and the latter happens when different subjects provide opposed decisions.
How can the level of conflict in a decision be evaluated? This question has received increasing attention in the last decade. Prediction paradigms, where participants have to anticipate an outcome have been the norm. In such paradigms, the level of ambiguity of the experiment is controlled, manipulating either the information the subject used to correctly make the prediction [1], [2] or the probability of success [2]–[6]. Consequently in these studies the ambiguity level was proposed a priori during the design stage. However, it has been shown that sometimes participants behavior does not necessarily correspond to that inferred from the probability of success [7]. In two of the earliest studies, participants had to advance the color or the suit of a card [3] or whether the next card was bigger or lower than the previous one [4]. This permitted the comparison between low and high difficulty guessing. Prefrontal areas, but also the anterior cingulate, were more related to trials with high difficulty. In other studies [7], [8] participants predicted the appearance of stimuli. Prefrontal, parietal and thalamic areas were active during such task. Volz et al. [1], [2], [5] presented pairs of alien comic figures and subjects had to infer which figure would win in a fictional fight. In one of the experiments there was an unknown probability of winning for each pair of figures that had to be learned as the experiment advanced. In the other experiment there were a set of rules that marked which figure won each time. The level of uncertainty of the experiment was manipulated by varying the degree of knowledge of the winning rules provided to the participants. Although there were minor differences in brain activation between the two paradigms, a fronto-median cluster correlated with the degree of uncertainty independently of the paradigm used. Huettel et al. also found a frontomedian activation when processing uncertainty in a paradigm where visual cues helped to predict the following answer [6]. In a more recent article [9], male subjects were required to discriminate attractiveness between pairs of women faces. Each picture had been rated previously by another group of participants, allowing to estimate and control the level of decision conflict. While all these studies associate pre-frontal and or fronto-median areas to the processing of conflict, a role in uncertainty management has also been assigned to the cerebellum [10].
As shown above, the concepts of certainty/uncertainty have been commonly used in decision making studies and most of their quantifications have been represented by either theoretical probability distributions or by empirical relative frequencies. Interestingly, an uncertainty descriptor that can be quantified from any probability distribution is the central measurement of information theory.
Goñi J, Aznárez-Sanado M, Arrondo G, Fernández-Seara M, Loayza FR, et al. (2011, March 9). The Neural Substrate and Functional Integration of Uncertainty in Decision Making: An Information Theory Approach. PLoS ONE, 6(3): e17408. doi:10.1371/journal.pone.0017408
Tags:
No comments:
Post a Comment