The value of advice from the perspective of a decision neuroscientist

Our lab investigates metacognition or ‘thinking about thinking’. This thinking can be about our one’s own thoughts, such as understanding that we perceive reality differently from how it objectively is; or thinking about other people’s thoughts, such as understanding that someone else perceives reality differently from us. This thinking, albeit being hidden somewhere deep in our cognition, can have very real consequences to everyday decisions. For example, a patient that does not acknowledge their illness may see little reason to comply to their medication or seek medical help.

When these occurrences of ‘incompliance’ are disputed in court it is often mentioned that, by law, markers of decision-making capacity may not be mixed up with the decision made. In other words, just because you chose not to follow your doctor’s advice and did not take your medication, this does not directly mean that you are unable to decide. Instead of focusing on the outcome of the decision (e.g. to take medication or not), the court requires a capacity assessment to be informed by the decision-making process (e.g. whether the advice was considered in the first place), which is much more difficult to measure.

When I learned this at the start of my PhD I started to see where and how decision neuroscience may contribute to law in the context of decision-making capacity. Decision neuroscience may contribute with a better understanding of decision-making processes, i.e. What is going on in someone’s mind when they are faced with different choice options?

Intrigued by this question I decided to focus my PhD on the process of advice-taking. This is an interesting concept from many angles. We often consider people that are too malleable to advise ‘soft’ and ‘easily manipulated’; while at the same time, often consider people that never change their minds to be considered ‘stubborn’ or ‘conservative’. This tension is also reflected in capacity assessments, which require people to be open to new viewpoints or the clinician’s advice when it is helpful; but not so much that they change their opinions with every changing situation.

Breaking down the advice-taking process into separate decision-making components may help to provide more targeted support. There may be individual and cultural differences in how we mentally figure out who is right: some people may be more inclined to overly doubt themselves even when they know that not all advice is accurate; others may systematically overestimate the accuracy of advisers at the cost of their own opinion. Whereas the latter group may benefit more from training their metacognitive accuracy, the former may benefit more from clearer communication and learning to take some advice with a pinch of salt.

A combination between neuroscience and computational modelling can help us reveal that, while different decision processes can have similar decision outcomes, they may still have completely different cognitive bases. For example, someone can take advice, not because they have ‘seen’ and adopted the adviser’s perspective’ (private compliance), but just because social compliance is a socially accepted thing to do (public compliance).

I was recently involved in the Dear World Project, a public engagement project where neuroscientists and artists were paired to explore each other’s research methods to collaborate on a piece of art reflecting the neuroscientists work. I was paired with artist Tom Berry and this paper was very useful to kickstart the collaboration between Tom and me. I tend to use a lot of technical language when talking about my work (sorry!), and it already became clear quite early on in the collaboration that it was important for Tom and me to ‘speak the same language’. At one of our first meetings, we met at a cafe and went through the main definitions that my research focuses on. A few weeks later, he sent me a quiz: he had visualised four core definitions of my research: metacognition, mentalising (metacognition about other people’s thoughts), private and public compliance, and I had to guess which definition corresponded to which drawing. These initial drawings were the start of what was later to become our artwork (A Sense of Direction): a beautiful visualisation of five different people who interact with one another around a simple choice: whether to go left or right. This time it was the audience’s turn to guess which process corresponds to what concept and write down how this process played a role in their everyday lives.

There may be a lot to gain if scientists and policymakers would reach out to one another more often, so as to profit from each other’s knowledge. With an evidence-based government and one of the most leading neuroscience laboratories right around the corner, I found London a very encouraging place to work on such a translational turn.

The full open access paper can be accessed here.

You can read more about Elisa’s contribution to the Dear World Project here.

Elisa also told us more about her path into neuroscience in our #WhoIsWCHN series, which you can view here.


Computations of confidence are modulated by mentalizing ability
E van der Plas, D Mason, LA Livingston, J Craigie, F Happe, SM Fleming
paper | code

Abstract: Do people have privileged and direct access to their own minds, or do we infer our own thoughts and feelings indirectly, as we would infer the mental states of others?  In this study we shed light on this question by examining how mentalizing ability—the set of processes involved in understanding other people’s thoughts and feelings—relates to metacognitive efficiency—the ability to reflect on one’s own performance. In a general population sample (N = 477) we showed that mentalizing ability and self-reported socio-communicative skills are positively correlated with perceptual metacognitive efficiency, even after controlling for choice accuracy. By modelling the trial-by-trial formation of confidence we showed that mentalizing ability predicted the association between response times and confidence, suggesting those with better mentalizing ability were more sensitive to inferential cues to self-performance. In a second study we showed that both mentalizing and metacognitive efficiency were lower in autistic participants (N = 40) when compared with age, gender, IQ, and education-matched non-autistic participants (N = 40). Together, our results suggest that the ability to understand other people’s minds predicts self-directed metacognition.

Isolating cultural contributions to metacognition
E van der Plas, S Zhang, K Dong, D Bang, J Li, ND Wright, SM Fleming
paper | code

Abstract: Some aspects of human metacognition, such as the ability to consciously evaluate our beliefs and decisions, are thought to be culturally acquired. However, direct evidence for this claim is lacking. As an initial step in answering this question, here we examine differences in metacognitive performance between populations matched for occupation (students), income, demographics and general intelligence, but drawn from two distinct cultural milieus (Beijing, China and London, UK). We show that Chinese participants have heightened metacognitive evaluation of perceptual decision-making task performance in comparison with UK participants. These differences manifested in boosts to post-decisional processing following error trials, despite an absence of differences in first-order performance. In a second experiment, we replicate these findings in a new task that replaced post-decision evidence with equivalent social advice. Together, our results are consistent with a proposal that metacognitive capacity is shaped via socio-cultural interactions.

A systematic review of metacognitive decision-making in autism
E van der Plas, D Mason, F Happe

Abstract: Autistic people often have an atypical profile of abilities: while excelling on some structured tasks, many struggle with making real life decisions. To test whether decision-making in autism is different from in typically developing controls, we reviewed 74 studies that compared decision-making performance between autistic and comparison participants (N=1,932 autistic and N=3,179 comparison participants) between 1998 and 2020. Our searches revealed four main decision-making paradigms that are widely used in the field of decision neuroscience: perceptual discrimination, reward learning, metacognition, and value-based decision-making tasks. Our synthesis highlights that perceptual processing and reward learning were relatively intact in autistic versus comparison participants, whereas subjective decision-making and metacognitive accuracy were often atypical. Furthermore, decision-making differences were most pronounced when the autistic participant was explicitly probed to report on an internal belief, whilst implicit markers of the same decision (e.g., error related response times) were usually not different. Our findings provide evidence in favour of a metacognitive explanation of decision-making atypicalities in autism.

Advice-taking as a bridge between decision neuroscience and mental capacity
E van der Plas, AS David, SM Fleming
paper | blogpost

Abstract: A person’s capacity to process advice is an important aspect of decision making in the real world. For example, in decisions about treatment, the way patients respond to the advice of family, friends and medical professionals may be used (intentionally or otherwise) as a marker of the “use or weigh” requirement of decision-making capacity. Here we explore neuroscientific research on decision-making to identify features of advice-taking that help conceptualize this requirement. We focus on studies of the neural and computational basis of decision-making in laboratory settings. These studies originally investigated simple perceptual decisions about ambiguous stimuli, but have more recently been extended to more complex “value-based” decisions involving the comparison of subjective preferences. Value-based decisions are a useful model system for capacity-related decision-making as they do not have an objectively ‘correct’ answer and are instead based on subjective preferences. In this context, advice-taking can be seen as a process in which new evidence for one or other option is integrated, leading to altered behaviour or choices. We use this framework to distinguish between different types of advice-taking: private compliance consists of updating one’s privately held beliefs based on new evidence, whereas in the case of public compliance, people change their behaviour at a surface level without shifting their privately-held beliefs. Importantly, both types of advice-taking may lead to similar outcomes but rely on different decision processes. We suggest that understanding how multiple mechanisms drive advice-taking holds promise for targeting decision-making support and improving our understanding of the use and weigh requirement in cases of contested capacity.