Skip to ContentSkip to Navigation
Research Bernoulli Institute Cognitive Modeling Research

Project: Understanding Problem Solving in the Brain: Mapping the Flow of Information with Model-Based Analyses and Hidden semi-Markov Models

(NWO Veni Grant 451-15-040)

People: Jelmer Borst

Most of our daily activities – navigating busy traffic, solving an algebra problem, or filling the dishwasher – require some form of problem solving. The key cognitive functions that are involved – working memory, declarative memory, and cognitive control – have been mapped onto the fronto-parietal brain network. However, it is unclear how these crucial functions are implemented at an algorithmic level, and how the flow of information is directed within the network.

In this project, I will develop a detailed computational model of problem solving in the fronto-parietal network. To this end, I will apply and extend two innovative analysis techniques that I co-developed: model-based fMRI analysis (e.g., Borst & Anderson, 2013) and Hidden semi-Markov Model (HSMM) analysis (Anderson, Zhang, Borst, & Walsh, 2016; Borst & Anderson, 2015). Both techniques will be used to analyze data of the same experiment measured with three different neuroimaging methods (fMRI, EEG, and MEG).

The experiment is designed to uncover general problem-solving mechanisms. It consists of two complex tasks: solving algebra problems and mentally reordering information. Model-based neuroimagin analysis and HSMM analysis are more powerful than traditional neuroimaging analysis methods, and are especially suited for uncovering the general cognitive mechanisms used by both tasks, such as working memory. Model-based analyis allows for an unsurpassed neural localization, while HSMM analyis can discover processing stages within the problem-solving tasks, including their cognitive functionality, timing, and location.

Using three different neuroimaging methods for a single experiment provides a unique opportunity to combine the strengths of these methods. fMRI has high spatial resolution, while EEG and MEG have excellent temporal resolution. The data of these three methods, in combination with the new analysis techniques, will allow me to build a computational model at the level of neural interaction, explaining how problem solving works in the brain. This knowledge can be used for improving education (e.g., algebra) and human-computer interaction, both with and without online brain measurements.

See for more information:

Anderson, J. R., Zhang, Q., Borst, J. P., & Walsh, M. M. (2016). The Discovery of Processing Stages: Extension of Sternberg’s Method. Psychological Review.

Borst, J. P., & Anderson, J. R. (2013). Using Model-Based functional MRI to locate Working Memory Updates and Declarative Memory Retrievals in the Fronto-Parietal Network. Proceedings of the National Academy of Sciences USA, 110(5), 1628–1633. http://doi.org/10.1073/pnas.1221572110

Borst, J. P., & Anderson, J. R. (2015). The discovery of processing stages: Analyzing EEG data with hidden semi-Markov models. NeuroImage, 108, 60–73.

Last modified:16 February 2021 2.01 p.m.