Bernhard Hommel

Interview with Prof. Bernhard Hommel

Department of Psychology, Cognitive Psychology Unit
Leiden University, The Netherlands
Personal website

What do you think cognitive systems are?

I do not really have a fool-proof answer but I would think a proxy would be that this is any system that is adaptive and context sensitive so everything that does not obviously function in a fixed hardwired stimulus-response fashion.

What is your area of research within Cognitive Systems?

I do many different things but I think that the majority would fall into what you can call the relation between perception and action. Even though people usually think that information processing goes from perception to action I am also interested in the opposite way how action affects perception. I am also interested in the acquisition and control of voluntary action and the integration of distributed information in the human cortex

Why did you become a researcher?

Well in general, as many people, I am just very curious and this is a good way to live that out. In my particular case I was also insufficiently talented to become a stage actor so I thought being a researcher calls for a combination of different interesting skills including acting-related ones.

How did you get into Cognitive System research?

I am a social psychologist by training originally and I still think that social psychologists are addressing much more interesting and cool phenomena. But my problem was that they provide very little explanation and are often happy with surprisingly little information about the mechanisms involved.

Where did you study and what subjects did you study?

I studied in Bielefeld which is in the north-west of Germany. Not a particularly interesting city but the university has a great library and a very strong research orientation. It is a very young university so that was good. I was officially studying psychology as one study and literature sciences as another. I also looked around into some sociology, history, and philosophy and that stuff.  

Can you describe briefly how you are doing what you do?

I am an experimental psychologist. Mainly I am doing laboratory experiments, so artificial little tasks in the cellar as it were with humans. It is mostly with young adults but also sometimes with other populations like infants, children, old people, drug users and so forth.

What are the techniques used in your research?

I am coming from behavioural research so we are manipulating variables that you can assume to have an impact on human performance. We look whether they do show this impact on behaviour, like reaction times or accuracy in behaviour. But recently we are also using more brain imaging techniques like EEG, fMRI or MEG because that gives you some more insight into the different processes involved in performance.

Can you tell me why they are important?

I mean behaviour is already pretty interesting and you can find surprisingly many things out by looking at behaviour during interesting experiments which unfortunately is increasingly underestimated in the times of brain imaging. But at some point behaviour does not tell you enough to get more into the different processes. So for example if you are thinking of simply presenting a simple object on a screen and asking someone to press a key there are many processes involved. You have to register the physical stimulus, you have to recall from your memory what that is, you have to attend to it, you have to select an action, how to respond to it, you have to carry it out and so forth and just using imaging techniques helps you a bit more to kind of segment these different processes and look at how they work.

What are the major implications of your work?

I think the most general answer is to build general theories, where possible, of how human cognition works. We are asking questions like, trying to build theories of how you select an object in attention or how you can ever carry out a voluntary action. This does not only have philosophical implications like free will but also mechanical implications regarding, say, how you actually do control your motor system in a way that is suited to reach a particular goal. The basic idea is to carry out a lot of silly little tiny experiments that if you put together gives you a general mechanistic theory of how that actually works. Once you have done this you can use it for several things, for example diagnosing processing problems and pathological deviations from normal ways people are functioning.

For example you can try to improving processes or behaviours by selective training. You can try repairing them if you find them to deviate by using behavioural and neuropsychological methods. And you can try to compensate. For example we are doing a number of aging studies where you see that performances are getting worse. Now there is very little you can do in terms of really repairing the process, but if you understand how it works you can try and compensate for it, for example with drugs, providing dopamine that is going downhill with age or by behavioural training or things like that.

What skills do you think are most important to a Cognitive Systems researcher?

(Good) Cognitive systems researchers are very broad. You need certainly abstract thinking and particularly good experimental skills as the more creative you are experimentally the more you can find out about the whole thing. I think a cognitive systems researcher in particular needs a very broad scope. You need to really relate many different findings from various research areas and disciplines and if you have a narrow scope you it is just not working.

What do you think is most satisfying about Cognitive Systems research?

I think it is related to that broadness. It touches so many problems which tend to over challenge one but at the same time this is what makes it really interesting. Challenging is in both ways like a kind of a negative and a positive thing. It bridges different disciplines that are also making it very complicated. For instance more and more you can only research in interdisciplinary groups which means that you are not an expert in all aspects of the stuff you are investigating. To stand the feeling to be stupid and trust colleagues who can fill in the expertise you do not have. That is a very interesting endeavor, to figure out how to these things work also interpersonally.

What do you consider is the most challenging about being a Cognitive Systems researcher?

It is also related to that but I think it is really always amazing how many skills you actually need, how many different skills. Just with the programming skills researchers sometimes need (which was true at least when my career started) people made as much as three times as much money that I made as a PhD student. There are other things as well such as the presentational skills that you need in order to be a good scientist, the personal skills you need to be a good part of these networks, and you need writing skills and analytical skills. So the challenge is to combine all these things and this explains why so many people fail I think.

What do you think are the main challenges for the future?

That is very difficult to tackle, and this is not a very clear answer I am afraid, but I think it is understanding the dynamics of changes over time, from very slow to very fast. For example, how does the firing rate of a neuron relate to changes in intelligence within a second or over lifetime, and how do these different levels of dynamics interact? This is very difficult because we do not even know how to describe dynamics often, how to model the processes we consider, and to relate them to models. The other general ongoing thing, it is not a new thing, is the relation between mind and brain. I mean now people think they are getting somewhat further by locating something somewhere but that does actually raises more problems than it answers because what does that mean for a mind to have parts if it located somewhere. So that is I think still very challenging.

More technically I think that these are times of where our way to do the research must change so we must organise that differently. It is no longer the times of individual geniuses but the times of social networks in which excellent research emerges. To facilitate that we must change the way we do, fund, and reward research and we do not always have very good institutions, political ways to organise that.

There are several discussions or debates associated with Cognitive Systems research.

Could you mention issues relating to your work?

One of the more general issues is for instance related to the distributed representation problem. So you know the brain represents different features of objects in very different areas like colour and shape maps and the same is true for action. So which means that many different areas, or the codes in many different brain areas need to be organised and somehow talk to each other in order to achieve cognition as it were or do some interesting cognitive work. It is really unclear how that works.

Can you outline the arguments of the opposing sides of the debate?

There are two main theoretical camps and these people really hate each other and develop strange and fiercely debates sometimes, if they talk at all. One is based on convergence logic: Let’s say you perceive an apple. Then there are neurons that understand that there is something green and you have neurons that understand that there is something round. And then you have a second type of neurons that code for the conjunctions, that are sensitive to, if it is green and round then I fire. And from that let's say the apple neuron is built.

There is another approach that is based on neural synchronization: the firing rhythm of neurons coding for green may be synchronized with the firing rhythms of neurons coding for round. This synchronisation may be a means to talk, a means of communication, so if neurons are phase synchronised this means that they are working on the same thing. I think it is obvious that both possibilities have so many strengths that it is very reasonable that the brain uses both mechanisms, but most people believe it must be only one and the other is wrong. Well they just go on in arguments about that.

Thank you!

Text only
An euCognition Network Action
COGNITIVE SYSTEMS OUTREACH
Department of Computer Science
University of Bath
Send us an email:
webmistress
euCognition - www.euCognition.org
HomeFor PressFor KidsFor AcademicsAbout Us
Meet the Scientists
Explore the Debates
Study at University
Robot Gallery
Home > Meet the Scientists > Prof. Bernhard Hommel