Anne-Marike Schiffer

Groups overcoming partisan bias

When people start a group discussion, this ideally leads to individuals learning from more knowledgeable members of the group, until the collective outsmarts individuals. However, research has also shown that group interaction, for example, on social media, can lead to formation of echo chambers and drive people’s opinions away from the ground truth, particularly on partisan issues. This poses the question under which conditions social interaction about strong partisan issues can lead to more people having better-informed opinions. Douglas Guilbeault, at the University of Pennsylvania, and colleagues conducted an experiment to establish under which conditions groups of liberal and conservative participants would perform best in estimating trends in climate change, following structured online interaction. The researchers found that the best performing group was made up of liberals and conservatives who did not know about each other’s political leanings. If groups only contained subscribers to one ideology, or if people in mixed groups knew the political leanings of other members, then partisan bias led groups to perform significantly worse. This study represents an advance into understanding the conditions that make human social learning successful, even when people initially show partisan bias. These results for climate change suggest that it will be important to see whether similar mechanisms can improve decision-making on other partisan issues.

Babies infer value from effort

We invest more effort to get hold of objects that are dear to us compared with those we value less. Conversely, by observing how much effort someone invests obtaining an object, we can learn about their preferences — and potentially about the object’s true value. Learning the value of actions is important for children, but whether they can infer value from observed effort is unclear. Shari Liu, of Harvard University, and colleagues show that ten-month-old infants can judge the value of an object from observing the effort actors spend. In three experiments, infants watched a cartoon character jumping a higher fence, climbing a steeper hill and leaping across a wider trench for one object, but not for another. The researchers then measured infants’ interest when the actor chose the object they were previously less willing to exert effort for when no effort was involved in obtaining either object. Children showed a higher level of interest in this choice, indicating that they had learned which object had a higher value. It is an open question whether infants also have concepts of non-physical effort, or risk, and can take these into account when judging the value of goals based on observed behaviour.

Prediction and validation

That the brain is a ‘predictive machine’, constantly predicting the next input, is an influential theory in cognitive science. Widely cited evidence for this idea came from research into language comprehension, including an influential 2005 study by DeLong and colleagues. But some of the conclusions of this study have been questioned. How much do we really predict? A new publication led by Mante Nieuwland and based on collaboration between 23 researchers reports concerted attempts from 9 laboratories to fully replicate the findings of the 2005 paper. The original study concluded that listeners automatically predict both upcoming nouns (for example, ship/airplane) and articles (for example, a/an). The new set of studies validated the key finding that listeners predict the nouns in a sentence, but did not provide strong evidence that articles are predicted as well. This supports the view that prediction in the domain of language processing may be less comprehensive than previously suggested. Further research will need to clarify the extent of prediction as well as its neural signatures.

Believe in the effect

Social psychology research has long asked how beliefs about social status can affect behaviour, suggesting, for example, that people who believe they are in a powerful position in a dyadic interaction perform cognitive tasks better than those who think they’re playing a lowly role. But are these effects truly driven by participants’ beliefs about their social status, or may behaviour instead be influenced by social signals sent by the experimenter? Thandiwe Gilder of Bangor University and Erin Heerey of Western University addressed this question in a series of experiments, in which they varied which social role participants were playing and the experimenters’ beliefs about participants’ role. They found that participants did respond faster and more accurately in some conditions than others. But it wasn’t the subjects’ belief about their status that mattered. Instead, participants were faster or more accurate when experimenters believed they were in the role associated with faster and more accurate responses. This research highlights two important aspects: first, it shows that subtle social beliefs can have real effects on behaviour. Second, it emphasizes how important it is for scientific progress that previous results are put to the test.

Circuits of Parkinson’s disease

Parkinson’s disease is a neurodegenerative disorder widely associated in the public mind with the striking motor symptoms, but it also leads to cognitive changes. These days it’s increasingly common to treat the motor and cognitive symptoms of Parkinson’s disease with deep-brain stimulation, which involves implanting electrodes in the brain and using weak electrical currents to affect neural signalling. But why and how does this work? In a recent study, Wolf-Julian Neumann, of the Charité Universitaetsmedizin, and colleagues combined measures of behaviour, a computational model and brain scans that track the connections between different brain areas, to better understand in which neural circuits the behavioural changes in Parkinson’s disease originate, and how deep-brain stimulation alleviates these adverse effects of the disease. Comparing the performance of controlled and cued movements under stimulation and without stimulation, the researchers found that, in particular, cognitively controlled movements rely on a brain circuitry known as the hyperdirect pathway, while motor control relies more heavily on another well-known connection, the indirect pathway. Teasing apart the contributions of different brain areas and their connections is an important endeavour, not only to better understand the origins of symptoms, but also to tailor interventions to obtain the best outcomes for patients.

Avoiding safe options

Negative experiences make us cautious to avoid similar situations in the future. But would one bad burger make you forego all American food? People apparently differ in how much negative experiences generalize to similar situations. But what happens in the brain when people learn to avoid similar items or events? And is it possible to link this tendency of over-avoidance to ways of thinking or psychiatric traits? This is the question that Agnes Norbury and colleagues of Cambridge University addressed across two studies. The researchers measured brain activity while participants played a game in which they pressed a button to avoid an electric shock when they saw specific pictures of flowers, but needed to avoid pressing the button for other flowers to avoid pain in the future. The second study used a similar task in an online setting, and collected data on participants’ mood, anxiety and obsessivecompulsive traits. The researchers found that when people reacted towards stimuli that look similar to the ones that signify danger, as though these predicted danger too, the response in visual brain areas towards the stimuli became more similar. The online study showed that people who have high anxiety scores are particularly prone to treat ‘safe’ events as though they predict negative consequences.

How words take on new meaning

In English, we use the same word to describe the flesh of animals when served as food, a person’s secret plan, and a sports event. In fact, the word ‘game’ has even more meanings. How do words take on new or alternative meanings? This is a question addressed by Christian Ramiro, at the University of California, Berkeley, and colleagues, who used computer models to test different accounts of word-meaning evolution. Three examples are a model in which a new sense of the word always relates back to the original sense, a model in which only the last developed sense of a word influences its new meaning, and a model in which meaning follows a nearest-neighbour chain. In the chaining model, different meanings can evolve into separate strands, and the newest meaning in each strand always relates back to the most closely related meaning that exists for a given word. Of the compared algorithms, this model best predicted in which order a set of about 5,000 English words took on new meanings over the past 1,000 years. The current study provides an elegant formal account for how words evolve, pointing to mechanisms that ensure that new meanings are understandable to listeners.

Stable brain networks

How different parts of the brain communicate with each other is a central question in neuroscience. So far, the work has led to seemingly contradictory results. Connected areas, so called ‘functional networks’ have been described at the group level, indicating that humans share the same networks. But differences in network properties have also been identified as markers of psychiatric traits, arguing for dissimilarities between people. Moreover, different tasks have been linked to differences in network activity, suggesting that even within one person, networks can vary over time. Caterina Gratton, of Washington University, and colleagues report a new study in which they analysed a dataset of repeated measures of the same participants performing the same tasks to tease apart these different potential influences on brain networks. The study shows that network properties are largely shared by all participants in the group, but that there is indeed individual variability. However, measuring the same individual typically yields relatively stable results and the modulation by task is also comparatively small. The study increases trust in previous conclusions, including links between individual differences in brain networks and traits. The findings also afford specific recommendations for how future network analyses need to be planned and conducted to identify effects at the different levels.

Slow response a force of habit

Reaction time (RT) — how fast a person responds — is the most widely used measure to study the processes of the mind that underlie interacting with the world. But what cognitive computations are reflected in RTs? Aaron Wong of Johns Hopkins University and colleagues report experiments showing that participants’ prior experience can greatly influence RTs. Especially strong evidence for this claim comes from one experiment in which participants moved their hands from one position to another, avoiding virtual obstacles. Participants were generally faster to respond to a task when they could follow a cued path around the obstacles than when they performed the movement unguided. However, if participants had any experience with performing the unguided version of the task, their response times remained slow when the visual cue was reintroduced in subsequent rounds. This finding shows that the time when a movement starts can be determined habitually, rather than by the time it takes to plan the movement. It reinforces the suggestion that initiation and planning of movements are separate processes in the brain. It is yet unclear whether the finding translates to other domains — but given the ubiquity of serial experimental designs, it may have far-reaching implications for cognitive science.

No confidence in wrong actions

Obsessive-compulsive disorder (OCD) is often associated with the urge to perform ritualistic actions, such as checking or washing. Strikingly, OCD sufferers usually know that these actions are not necessary and that repetition fulfils no real purpose. This raises the question whether OCD involves a more general problem in choosing and evaluating actions. Matilde Vaghi and Fabrice Luyckx, of Cambridge University, and colleagues show that OCD patients may suffer from a specific impairment in choosing the right action — despite their intact ability to evaluate success. The researchers used a computerized task in which patients and controls caught flying particles by placing a (virtual) bucket where they predicted these to land. Participants then indicated their confidence in their placement. Where particles landed always varied slightly, with occasional drastic changes in position. Compared with controls, OCD patients often adjusted actions too much in response to small errors, while their confidence in their placements showed that they had some insight that these large adjustments were unwarranted. The researchers say that these findings raise new questions about the link between actions and confidence, and that it is important to understand whether OCD entails changes in brain function relating to evaluation and choice of action.