Decision Making: Hitting an uncertain target
Life is full of uncertainty. Take, for example, a football player preparing to take a penalty kick. He or she can aim to kick the ball along a number of different trajectories, and the success of the kick will depend on a range of different factors, such as small differences in the execution of the kick and, most importantly, the response of the goalkeeper. The simplest way to think about the decision that the player has to make is to say that he or she has to decide between two options: aiming for the left half of the goal or aiming for the right half (Figure 1A). Most studies of the neuronal mechanisms that underlie decision-making have focused on "target selection" situations like this, where the available options are limited in number and clearly distinct from each other (Shadlen and Newsome, 2001; Romo et al., 2004; Thura and Cisek, 2014).
However, the player taking the penalty obviously has more than two options: they can, for example, aim low, which increases their chances of hitting the goal, but also gives the goalkeeper a better chance to make a save; or they can aim high to make it more difficult for the goalkeeper and themselves. Indeed, they can aim at any point in the goal. This means that the player has to decide between an infinitely large range of possible trajectories (Figure 1B). To do this, he or she has to estimate the probability of success for each possible trajectory and then select the one with the highest likelihood of success. In the language of neuroscience the player is confronted with a "target estimation" situation.
How might the brain make a decision in such a situation? Studies of primates suggest that a region of the brain called the dorsal premotor cortex plays an important role in making target selection decisions (Cisek and Kalaska, 2005), and that the primary motor cortex is responsible for executing such decisions. Now, in eLife, Lee Miller and colleagues at Northwestern University, including Brain Dekleva as first author, report the results of experiments on monkeys that shed new light on the role of these two regions in making target estimation decisions (Dekleva et al., 2016). The experiments involved simultaneously recording neural activity in the dorsal premotor cortex and the primary motor cortex as the monkeys performed a target estimation task.
Playing football would be impractical in an electrophysiological experiment so, instead, two monkeys were trained to use a device called a manipulandum to move a cursor on a screen. The experiment involved two types of trials. During the first type of trial the monkey held the cursor over a central target: then, after a random period of time, the central target disappeared and a new target appeared in one of eight locations (which were arranged in a circle around the central target). The monkey received a reward every time it moved the cursor to the correct target. Neurons in the dorsal premotor cortex (PMd neurons) and the primary motor cortex have a "preferred direction", and the Northwestern group measured the level of neural activity as a function of the angle between the preferred direction of groups of neurons and the direction selected by the monkey (Figure 1C).
During the second type of trial the location of the new target was indicated by number of small lines that were sampled from a distribution that was centered on the target location. (There were five small lines for one monkey, ten for the other). The lines supplied the monkeys with incomplete information about the target location, and Dekleva et al. were able to vary the degree of uncertainty in this information by varying the width of the distribution from which the lines were sampled. This meant that the monkeys were now performing a target estimation task.
When the degree of uncertainty in the visual information supplied to the monkeys was high they tended to move the cursor to a location that was the average of the target locations in the previous trials: this approach makes sense when relatively little information is available. The Northwestern group also observed relatively high levels of activity in the PMd neurons representing directions other than the selected direction when the uncertainty was high: this suggests that, during decision making, the dorsal premotor cortex represents all directions that might be correct, each weighted by its probability of success. Again this makes sense: when the degree of uncertainty is high, there is a higher likelihood that a direction other than the chosen one is correct, so the PMd neurons representing these other directions are more active. In contrast, in the primary motor cortex, only those neurons representing the chosen direction were active (Figure 1C).
Previous studies of target selection reported that the activity of PMd neurons representing the chosen action increased during decision making, whereas the activity of other PMd neurons decreased (Cisek and Kalaska, 2005; Thura and Cisek, 2014). This was taken as evidence that the dorsal premotor cortex plays a direct role in decision making. In contrast, by showing that the PMd neurons representing all possible actions are active throughout the planning and execution of a target estimation task, the results of the Northwestern team argue strongly against a direct role for the dorsal premotor cortex in decision making because it is the M1 neurons, but not the PMd neurons, that encode the chosen direction unambiguously.
Future experiments should compare the role of PMd neurons in both target selection and target estimation tasks. Examining situations that are intermediate between target selection and target estimation might also help to resolve the difference between the latest findings and previous work.
The latest results also shed some light on the question how probability and uncertainty are encoded in the brain. There is some evidence for an explicit representation of uncertainty in the orbitofrontal cortex of primates (O'Neill and Schultz, 2010). However, recent work in rodents indicates that other interpretations are possible (Ogawa et al., 2013), and the results of Dekleva et al. support an implicit (rather than explicit) representation (Ma et al., 2006).
By pioneering the study of an important class of real-world situations that have been neglected up to now, the work of the Northwestern group opens up a rich new field of investigation.
Bayesian inference with probabilistic population codesNature Neuroscience 9:1432–1438.https://doi.org/10.1038/nn1790
Neural basis of a perceptual decision in the parietal cortex (area LIP) of the rhesus monkeyJournal of Neurophysiology 86:1916–1936.
Article and author information
- Version of Record published: July 15, 2016 (version 1)
© 2016, Stuphorn
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
- Page views
Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Midbrain and striatal dopamine signals have been extremely well characterized over the past several decades, yet novel dopamine signals and functions in reward learning and motivation continue to emerge. A similar characterization of real-time sub-second dopamine signals in areas outside of the striatum has been limited. Recent advances in fluorescent sensor technology and fiber photometry permit the measurement of dopamine binding correlates, which can divulge basic functions of dopamine signaling in non-striatal dopamine terminal regions, like the dorsal bed nucleus of the stria terminalis (dBNST). Here, we record GRABDA signals in the dBNST during a Pavlovian lever autoshaping task. We observe greater Pavlovian cue-evoked dBNST GRABDA signals in sign-tracking (ST) compared to goal-tracking/intermediate (GT/INT) rats and the magnitude of cue-evoked dBNST GRABDA signals decreases immediately following reinforcer-specific satiety. When we deliver unexpected rewards or omit expected rewards, we find that dBNST dopamine signals encode bidirectional reward prediction errors in GT/INT rats, but only positive prediction errors in ST rats. Since sign- and goal-tracking approach strategies are associated with distinct drug relapse vulnerabilities, we examined the effects of experimenter-administered fentanyl on dBNST dopamine associative encoding. Systemic fentanyl injections do not disrupt cue discrimination but generally potentiate dBNST dopamine signals. These results reveal multiple dBNST dopamine correlates of learning and motivation that depend on the Pavlovian approach strategy employed.
The cell bodies of postganglionic sympathetic neurons innervating the heart primarily reside in the stellate ganglion (SG), alongside neurons innervating other organs and tissues. Whether cardiac-innervating stellate ganglionic neurons (SGNs) exhibit diversity and distinction from those innervating other tissues is not known. To identify and resolve the transcriptomic profiles of SGNs innervating the heart, we leveraged retrograde tracing techniques using adeno-associated virus (AAV) expressing fluorescent proteins (GFP or Td-tomato) with single cell RNA sequencing. We investigated electrophysiologic, morphologic, and physiologic roles for subsets of cardiac-specific neurons and found that three of five adrenergic SGN subtypes innervate the heart. These three subtypes stratify into two subpopulations; high (NA1a) and low (NA1b and NA1c) neuropeptide-Y (NPY) -expressing cells, exhibit distinct morphological, neurochemical, and electrophysiologic characteristics. In physiologic studies in transgenic mouse models modulating NPY signaling, we identified differential control of cardiac responses by these two subpopulations to high and low stress states. These findings provide novel insights into the unique properties of neurons responsible for cardiac sympathetic regulation, with implications for novel strategies to target specific neuronal subtypes for sympathetic blockade in cardiac disease.