Officially, my PhD is about emotion representation in the brain (or, more specifically, about the influence of context on the visual representation/processing of emotions). My main “tool” to probe these neural representations is (f)MRI, and several of my PhD’s “side projects” are about (f)MRI data analysis (like the application of machine learning on MRI data) and improving reproducibility/transparency of (f)MRI analyses. More information can be found on this website’s Software section!

Below, I’ve outlined some of the things I have worked or am working on!

Using machine learning to probe shared emotional processes

During my master’s (Research Master Psychology at the University of Amsterdam), I started working on a project (supervised by Suzanne Oosterwijk) aiming to investigate the neural overlap between self-experienced emotions and understanding emotions in others. In other words, are the neural networks engaged in processing and representing your own emotions also used when understanding the emotions of others? During this project, I learned how to use machine learning to “decode” neural representations (as measured with fMRI); for those interested, I kept track of the code for the analysis on Github.

Shared States The shared neural pattern between emotion experience and perception revealed by our cross-decoding analyses

After about 2.5 years, we finally published our study, titled “Shared states: using MVPA to test neural overlap between self-focused emotion imagery and other-focused emotion understanding” in the journal “Social, Cognitive, and Affective Neuroscience” (which is open source! Find the article here).

Multivoxel Confound Analysis (MVCA)

Multivoxel pattern analyses (MVPA), and especially machine-learning-based anayses, have been suggested to be more sensitive to neural representations than traditional univariate analyses. Others have argued, though, that part of this sensitivity might be driven by the analyses being more sensitive to confounds as well. Together with Steven Miletic, I am working on how to properly correct for these confounds in MVPA-based studies. Also, we are investigating how improperly controlling for confounds can lead to (strong) negative bias in model performance. We’ve presented a poster on this topic at ICON2017 and we hope to publish a preprint soon.

mvca The basic problem that occurs when you don’t cross-validate your confound regression …

Contextual impact on emotion perception

Apart from the side-projects mentioned above, we’re setting up a large project aimed at investigating how “context” may influence the perception of emotion in facial expressions. Specifically, we want to see if “affective knowledge” (i.e. the emotional associations developed with experience) influences how people perceive (see) the emotional expressions of others.

Scientific publications

Oosterwijk, S., Snoek, L., Rotteveel, M., Barrett, L. F., & Scholte, H. S. (2017). Shared states: using MVPA to test neural overlap between self-focused emotion imagery and other-focused emotion understanding. Social Cognitive and Affective Neuroscience. PDF.

Presentations/posters

Below, I’ve listed the academic (poster) presentations I’ve given so far:

  • A universal method of controlling for confounds in MVPA. Poster Presentation together with Steven Miletic at the International Conference for Cognitive Neuroscience (ICON), Amsterdam, The Netherlands, 5 August 2017. PDF.

  • Decoding Emotions: Using MVPA to explore the neural overlap between emotion experience and emotion understanding. Associatie van Sociaal-Psychologische Onderzoekers (ASPO) conference, Amsterdam, The Netherlands, 11 December 2015.

  • Local vs. global brain representations. Poster presentation at the NVP winter conference 2015, Egmond aan Zee, The Netherlands, 18 December 2015. PDF.

  • Decoding emotions in the brain. ABC-BIC Neuroimaging Symposium, Amsterdam, The Netherlands, 28 April 2015.

  • Exploring the neural overlap between emotion experience and understanding. Poster presentation at the Organization for Human Brain Mapping Conference 2015, Honolulu (HI.), U.S.A., 17 June 2015. PDF.

  • Patterns of Emotion Components. Presentation at the Brain and Emotion EASP pre-conference, Amsterdam, The Netherlands, 8 July 2014.

Also, recently I gave a “popular science” talk about “brain-scans in psychological research and society” (in Dutch):

  • Het (on)meetbare brein. Spui25: Proost op de Wetenschap, Amsterdam, Nederland, 12 Mei 2017. PDF.