Hedderik van Rijn

 
Experimental Psychology / Statistical Methods and Psychometrics
University of Groningen

prof. dr. Hedderik van Rijn
Heymansbuilding, room 2.76
Grote Kruisstraat 2/1
NL-9712 TS Groningen

hedderik@van-rijn.org
+31-50-363 6290

Temporal Cognition

All behavior takes place in time. My general research goal is to understand how cognitive processing is influenced by temporal constraints, how temporal patterns extracted from human behavior can be used in applied settings to improve personalization of artificial systems, and how the brain keeps track of time.

Although I use many different methods, from functional magnetic resonance imaging (fMRI), electroencephalography (EEG), pupil dilation, and behavioral studies to behavioral genetics using drosophila, the goal of all these approaches to inform formal models of temporal cognition. Depending on the purpose, these formal models can take the form of higher level mathematical (e.g., drift diffusion or linear-ballistic accumulator models), symbolic process (ACT-R based) models, the combination of both, or neurobiological modeling of low-level biological processes underlying time perception.

In addition to my work on temporal cognition, I have a keen interest in developing new methods for studying human cognition. For example, together with my colleagues I have developed new methods for model-based neuroscience, methods to assess whether behavior is driven by competing strategies in cognitive tasks, algorithms to deconvolve the pupillary response, and I have developed new apparatus and methods to do behavioral, cognitive experiments with drosophila.


Time as Cognitive Control

Although often not ackowledged, in many experimental tasks time plays an important role, if only because participants need to balance speed and accuracy. But temporal aspects also influence the nature of human performance, for example when one type of information is available earlier than another type (e.g., picture-word interference studies), when one needs to switch tasks in the context of sequential multi-tasking (e.g., glances in rear-view mirror), or when the speed of an incoming speech-signal determines how much time is available for language processing. In all these cases, temporal constraints can be seen as an additional source of cognitive control.

If one aims for precise descriptions of behavior, research cannot suffice with qualitative descriptions of these tasks as an exact duration is needed when one wants to predict the interference between to sources of information. Therefore, my work focusses on building detailed computational process models that take into account all relevant aspects of the task at hand - not just the decision stage.

Time-based Personalization

Good theories should allow us to predict what behavior can be observed in situations not yet scrutinized. Most theories in Experimental Psychology are based on distributions of response times, typically summarizing the data over many participants and trials. However, modern computers make it possible to process the response times generated by a participant as they come in, and thus allow us to update our knowledge of the underlying processes used by that individual on-the-fly.

This principle is used in the fact-learning method "SlimStampen" which is based on the memory theories of the ACT-R cognitive architecture. If, during a learning session, a to-be-learned item can be recalled fast and effortlessly, this item is probably stored well enough in memory. Therefore, learning time can better be spent on other items, and one can delay revisiting this item. On the other hand, if the correct answer was only provided after a long delay, this item should be revisted soon, because it apparently is not well stored yet.

The "SlimStampen" has been used at three different Universities in the Netherlands, at special-needs education centers in Portugal, Ireland and The Netherlands, and has been tested at many secondary education schools. As of academic year 2014/2015, the "SlimStampen" system will be offered to all students using the online learning system of Noordhoff Publishers, a major Dutch publishing house of secondary education materials.

Time in our Brain

Given the importance of time for accurate cognitive performance, it is striking to realize that it is still not known where the "clock" can be found in the brain that drives interval timing, the name for time perception at the (hundreds of) milliseconds to minutes range.

The best known information processing models of interval timing assume that an internal pacemaker generates pulses that can be integrated in an accumulator. By reading out how many pulses have accumulated, the system can have a sense of how much time has passed. If the system has to recreate the same interval, it can restart the accumulation process and wait until the same amount of pulses is accumulated.

Although this theory is very attractive because of its simplicity and face-value validity, decades of research have not been able to pinpoint the pacemaker. Even more, although EEG studies claimed to have located the "accumulator", work from my lab has demonstrated that the observed phenomena are at best an epiphenomena of time.

My work has therefore focussed on alternative accounts on interval timing, in which the source of time might be associated to other cognitive functions, such as working memory updates. To test some of these ideas, I collaborate with Drosophila experts at the research school of Behavioral and Cognitive Neurosciences and the University Medical Center Groningen to unravel the genetic bases of interval timing using fruitflies as a model animal.

Nevertheless, even though research has not settled on the neural substrate of interval timing, part of my work also focusses on the functional descriptions of interval timing and on the real-life applications and consequences of deviations between subjective and objective time.

 

Lab Members

Current PostDocs

  • Florian Sense, Adaptive Learning
  • Martin Riemer, Interval Timing in Realistic Tasks
  • Wouter Kruijne, Computational Models of Interval Timing
  • Sarah Maass, Timing in Clinical Populations

PhD students

  • Josh Salet, Timing in the Real-World (with Wouter Kruijne)
  • Maarten van der Velde, Bayesian Optimization of SlimStampen (with Jelmer Borst, Florian Sense)
  • Robbert van der Mijn, Temporal Fingerprinting (with Niels Taatgen)
  • Hermine Berberyan, Stage Detection in EEG/MEG (with Jelmer Borst, Niels Taatgen)

Programmers / Support

  • Tom Doesburg, Research Programmer
  • Thomas Wilschut, SlimStampen Support

Alumni PhD students

Utilization

For over 5 years, I collaborate with Noordhoff Publishers, bringing SlimStampen/RuggedLearning, the adaptive learning algorithm, developed in my lab to over 400.000 secondary education students. SlimStampen/RuggedLearning and the associated logo are registered trademarks in Europe, USA, China and various other countries.



Noordhoff Publishers

Grants

Over the years, my work has been supported by grants from:


NWO Netherlands Organization for Scientific Research

Horizon 2020: EU Framework Programme for Research and Innovation

EU Leonardo da Vinci Life Long Learning Programme

Innovative Action Programme Groningen

Gratama Stichting

Kennisnet, Dutch Ministry of Education

New Agendas for the Study of Time, sponsored by John Templeton Foundation

European Cooperation in Science and Technology, Intergovernmental EU Framework

European Office of Aerospace Research & Development (EOARD) of the Air Force Office of Scientific Research (AFOSR)

Granting agencies that have specifically sponsored conferences that I have (co-)organized are:


Deutsche Forschungsgemeinschaft

Office of Naval Research, International Office

European Office of Aerospace Research & Development