Intelligence in the Brain: Conversation with Neuroscientist Rogier KievitShare
Neuroscientist Rogier Kievit discusses the neuroscience of intelligence.
Rogier Kievit is currently a PhD student at the department of methodology at the University of Amsterdam. He studies psychological constructs from two explanatory levels, namely brain activity and behavioral measurements. His PhD project is aimed at developing conceptual and statistical tools to examine the relationship between these two types of measurements. He draws on a variety of perspectives-- from structural equation modeling, philosophy of mind, theoretical psychology and cognitive neuroscience.
I really dig his research and was happy to chat with him about a bunch of important topics relating to intelligence, the brain, and philosophy of mind.
1. What is intelligence?
Various lengthy definitions have been proposed, and I don’t think I can do much better than any of those. But my quick and dirty interpretation would be something like “A set of abilities required to acquire, process, synthesize and to put to effective use of knowledge and abilities”. However, I think insights into the nature of intelligence will come from the gradual accumulation of knowledge rather than single, unifying, ‘perfect’ definition.
2. What do IQ tests measure?
IQ tests give an estimate of a wide range of positively correlated abilities that together summarize an individual’s ability to synthesize and purposefully use knowledge and reason. Although IQ-tests certainly don’t capture everything about cognitive abilities, or even about human intelligence, they can be useful for certain purposes. However, we should always be wary of putting too much faith in strict boundaries or cut-off criteria: Even within a person there is measurement error and considerable fluctuation over time, so it is dangerous to try to capture all of someone’s cognitive ability in a single number.
3. How important is mental speed for intellectual functioning?
Assuming mental speed is interpreted as the time it takes to process and distribute information, I think mental speed is probably a necessary, but not a sufficient condition for (high) intelligence. The ability to balance and synthesize a lot of information is dependent on different kinds of information being available to simultaneously access and process. However, it is of course only one element of intelligence, and singular explanations haven’t fared well historically. Every proposed simple explanation or substrate or intelligence I know of has led to funny, but insightful, counterexamples. For instance, in several mental tasks, including some concerning working memory and mental speed, we loose, badly, to chimpanzees on both speed and accuracy. And in terms of the cortex-to-body ratio, humans are behind several species of rodents and some species of fish. The lesson to take from this is that there are, simply, no simple answers to understanding intelligence. Or, conversely, that we are consistently underestimating the intelligence of rodents and fish...
4. Where is intelligence in the brain?
I don’t think intelligence is somewhere in the brain, in the same way that ‘athletic ability’ is not in some particular place in your body, and the top speed of a car isn’t somewhere in your car. Firstly, intelligence reflects differences between people, so it can’t really be ‘in’ a person. Secondly, although intelligent behavior is ultimately dependent on the function of the brain, and perhaps more so on certain aspects of brain functioning than on others, I think attempting to locate intelligence somewhere is asking the wrong questions. We can learn a lot from studying the brain, but we can’t just peer in and hope for a clear, simple answer.
5. What do you think of current approaches to understanding the neural basis of g?
Traditionally, studies have looked for correlations between individual neurological measures and g. This was useful initially, but also inherently limited. Increasingly, the field is moving towards combining different kinds of data, and I think this is a good thing. Recent attempts have started to integrate both functional and structural data, to get a grip on the relationship between brain function and structure with regards to intelligence and intelligent behavior. Most importantly, there is an increasing realization that we need to focus on the dynamics of intelligence. For instance, the field of network analysis has been gaining popularity. These studies look at the dynamics and changes of brain activity, how these relate to intelligent behavior, and how these vary across people. For example, recent large-scale collaborations are moving towards studying groups of people as they develop and change over time, both behaviorally and neurologically. Such studies allow researchers to study the development of intelligence and the brain during childhood and adolescence, and during aging later in life. This is especially important, as it allows us to get a grip on the dynamical changes over time. We know already that this is not a simple, linear process, but the interaction of a range of abilities. In the past, intelligence research has learned a lot by studying changes over long periods of time, as this showed that every generation scores considerably higher on IQ-tests, the so-called Flynn-effect. The hope is that similar gains in understanding will be possible by being able to track and model the development of intelligence and the brain during childhood and old age.
6. What do you think of the Parieto-Frontal Integration Theory (P-FIT)?
Although the scope of the effort is both admirable and useful in collating a large body of research, it is ultimately not satisfying. Firstly, the P-FIT theory isn’t very specific, and the brain regions discussed include most of the cortex. As the authors themselves state, this same network has also been found to underlie cognitive functions related to perception, short-term memory storage, and language. Then again, this may simply be the way it is, but then we should perhaps ask more specific questions. For instance, my main problem with the P-FIT model is that it combines evidence from both the inter-individual study of intelligence (how do people differ from one another in terms of intelligence and/or brain structure?) and the intra-individual study of intelligence (what happens in people’s brains when they perform complex tasks?). Although these are both very interesting questions, they are also very different questions. Taking them together as ‘intelligence and the brain’ is perhaps a partial explanation of why the network they identified is so broad: It summarizes a range of answers to two complex, but distinct, questions.
7. Describe your latest research.
Generally, my research is concerned with using statistical tools to better understand the relationship between psychological behavior and the brain [paper]. For instance, one of the things we are looking at currently is integrating psychometric models to study what happens when people try to solve complex problems, both behaviorally and in terms of brain function. To really be able to understand this process, we must carefully take into account both differences between people and differences between different tasks, in terms of difficulty. Only then can we study the similarities and differences in strategies when solving complex reasoning tasks. We use certain psychometric (Rasch) models to take into account these differences, and so hope to better understand the similarities and differences in how people solve complex reasoning problems.
8. How does your own research move the field forward and overcome some prior limitations of studying the neural basis of intelligence?
We hope to add something by emphasizing the importance of using measurement models to the study of intelligence and the brain. Although a lot of research has looked at the correlations between intelligence and certain neurological properties, such as brain volume or gray matter, we don’t really know whether these different brain correlates covary together, either within populations or within individuals over time. Is it simply the case that some people have globally ‘better’ brains than others, or are there far more complex differences between, and within, people? I think the latter, but to get a grip on such questions, we need psychometric models. We use Structural Equation Modeling in much of our research, which has the benefit that the proposed hypotheses can be represented visually, and contrasting models can be compared in terms of how well they explain the data.
So far, we have looked mostly at general intelligence, to show that using such models are useful. However, we are currently applying these models to more detailed cognitive abilities, to see if the models that work best for general intelligence are different to those that work best for, say, working memory capacity. Ultimately, we can then, hopefully, graphically represent this hierarchy of relationships.
9. What are the practical implications of your research? How would school psychologists benefit from reading your research?
Ultimately, I believe that increased understanding can, and often does, yield unexpected practical benefits, so furthering understanding in and of itself is a worthwhile goal. For instance, the detailed study of patterns of brain activity and how to best measure these patterns has recently led to the, rather striking, ability to communicate with people who were thought to be in a vegetative state. This was hardly envisioned when people were developing fMRI-techniques.
The research we do doesn’t have a direct link with school psychology, but I think there are indirect links. Research done by my colleagues aims to shed light on the developmental trajectory of intelligence, which may have practical benefits in different ways. For instance, the mutualism model of intelligence, proposed by van der Maas, shows that if you assume that different, unique, cognitive abilities positively influence each other over time, you would expect exactly the pattern of intelligence we currently observe. This means that an increase in, for instance, working memory ability at a certain time may, over time, benefit a much wider range of (school) skills. These beneficial interactions have potential practical implications: It suggests that improvement of generally applicable skills, such as working memory ability or verbal ability, may have benefits far outside the realm of the that specific skill. This could be tested empirically by studying changes over time: It would imply that improvements in a certain skill at a certain time leads to improvements in other, related, skills, even without training on those specifically. Again, the comparison to athletic ability is apt: Many athletes train specific skills, such as individual muscles, to achieve better performance, over time, on some other, more general athletic skill. Conversely, improving overall stamina allows you to go on for longer, which leads to improvements on more specific skills or abiities. Better understanding of how cognitive abilities and properties of brain function affect each other over developmental time has, I think, the most promise for practical applications in the future. But, much more research is needed before such tentative findings can be translated into clear practical advice.
10. What does your research suggest about the malleability of intelligence?
My research so far has focused on cross-sectional studies of intelligence, that is, the study of a single population at a single time, so I cannot make any claims regarding the malleability of intelligence. But, there is tantalizing evidence that the brain is, at least to a certain extent, malleable. For instance, studies have suggested that intense behavior, such as learning to juggle or to pass a taxi exam, may change the structure of the brain. It is hard not to imagine your education as a much longer, more intense version of this behavioral manipulation. This is very relevant for intelligence research. We know that intelligence changes over lifetime, and we know that the brain changes over lifetime, but we know very little concerning the co-occurrence of these changes. Are they simultaneous, or does neurological change occur prior to behavioral performance, as is the case in Alzheimer’s disease? Recent research has shown how incredibly important it can be for healthy aging to keep engaging in even relatively modest cognitive activities. Even getting a slightly better grip on these changes may be very useful. These are very complex questions, but understanding them better may possibly provide insights into how intelligence develops, and how to cope with aging.
11. Will we ever be able to make drugs to make us smarter?
That’s not really my area of expertise, but I don’t think there will be a pill that will simply add a bunch of IQ points. This is because intelligence isn’t a single, monolithic ability. However, I wouldn’t be surprised if it turned out to be possible in a more modest sense, for certain aspects of cognitive function, such as alertness, or perhaps even specific aspects of memory. For instance, everyone has experienced that you feel less smart early in the morning, or at the end of a long day. Reversing these local and transient effects may achievable with ‘medication’ such as coffee. For example, coffee, for regular drinker like me, quite literally is a way to make me slightly smarter (or slightly less dim) in the morning. However, intelligence as a whole is far too complex and multifaceted for a single-pill ‘solution’. I think education is the best, and most sensible, way to make us smarter.
12. Does your psychological research have any implications for philosophy of mind?
I think the literature on philosophy of mind has much to tell empirical cognitive neuroscience, mainly in two ways. Firstly, it makes clear which assumptions underlie our thinking in relating the brain to behaviour, and whether these assumptions are tenable and sensible. More importantly however, different theories have different explanations and descriptions concerning the relationship between the brain and behavior.
In our work, we aim to take these different explanations and think about possible predictions that follow from such theories, and may allow us to compare them against one another [paper]. It turns out that theories such as identity theory, supervenience and emergence can be very useful in thinking about how to model brain measurements and behavioral measurements together. Although philosophy traditionally deals with questions outside of the realms of empirical testability, it can be very useful to inspire new ways of thinking about behavior and the brain. Ultimately, this is one of the most complex challenges for all of science: How can it be that this blob of cells in my head somehow gives rise to my ability to experience, think, perceive and feel? To get a grip on such questions, we need all the help we can get, and philosophy of mind certainly is one of the branches that can be insightful in this respect.