The research group on Experimental Cognitive Science is headed by Prof. Dr. Volker Franz. Here are the main topics we address:

Biological information processing

One focus of our research is on the question of how humans process visual information in order to either directly guide motor actions or perform cognitive tasks, such as object recognition.

A famous issue is the question of how we can define functional sub-processes within the cognitive system. Current theories often assume that the neuronal areas and processes that create visual awareness or that are relevant for object recognition are different from those that guide visual actions. We have shown that these processes are more similar than is currently thought—which was very unexpected and surprising for many scientists.

Other issues we have investigated are related to the information processing that is used to guide visual actions. We investigate which aspects of visual objects are most relevant to guide an action, to which degree visual feedforward and feedback processes are involved when performing the action, how adaptation and memory affect actions and the corresponding brain activity, how different visual cues are combined, and have developed a simple model of motor guidance in grasping, which forms a link to applied research in robotics.

To conduct this research, we have employed a variety of techniques. For example, we have used virtual environments that allow us to simulate and manipulate the sensory input and the motor interaction with the environment (using stereo--computer graphics and robot arms that are attached to the fingers of the participants), complex psychophysical methods, EEG and fMRI measurements of brain activity during visually guided actions, developmental studies with children, and studies with neuropsychological patients.

Statistics & Research methods

A second focus of our work is on the combination of neuroscientific issues with methodological and statistical topics. Repeatedly, we found that methodological inaccuracies can lead to serious problems in the interpretation and evaluation of neuroscientific research. In these cases, we pursued the methodological questions in separate projects attempting to resolve them in a general way.

For example, we investigated confidence sets and their graphical presentation as 'error bars'. These are important tools for graphical data analysis and there is a wide consensus that they should be used more often in all empirical sciences. However, there are many situations where the data pose specific problems such that it is difficult to determine the appropriate confidence sets. In different projects, we tried to improve on this situation.

In other projects, we developed mathematical models that allowed to better investigate neuroscientific questions. For example, we presented a mathematical model that helped to clarify under which conditions it is possible to compare effects on two different dependent variables. This model shows that under certain conditions, corrections are necessary.

A recent application of such methods are questions related to conscious/unconscious processing. Here, again, we found that inappropriate methods can create serious bias in the interpretation of empirical data.

For more details on all these topics, please visit our publication page.