L-POP: Language Processing by Overlapping Predictions

Funded by ERC Consolidator Grant (no. 617891).

This site is under construction.


This ERC project investigates the role of predictive processes for language comprehension, unsing various methods from cognitive psychology and cognitive neuroscience, including EEG, MEG, and functional MRI.


Predictive Processes in Visual Word Recognition

In this part of L-POP, we investigate the role of predictive processes during visual word recognition, including the hypothesis of word recognition based on an orthographic prediction error, as well as priming studies to investigate pre-activation of different aspects of word-level information.

See for example here or here.

encodingdecoding4web.jpg

Semantic Predictability

In this part of L-POP, we explore various aspects of how brain activity can be predicted from models of semantics.

See here for more details.


cross4web2.jpg
P3P64web.jpg

Domain-general effects of prediction violation

In this part of L-POP, we investigate to which degree predictive processes during language processing are similar to error signals produced during expectation violation in other domains.

See for example here.

Cross-level predictions in language comprehension

One of the main theoretical assumptions of L-POP is that predictive processes at different levels of linguistic representations interact. We examine this by investigating how semantic predictability influences lower level neural representations of speech input.


L-POP Team Members:

Prof. Dr. Christian Fiebach (PI)

Dr. Jona Sassenhagen (PostDoc)

Dr. Benjamin Gagl (PostDoc)

Susanne Eisenhauer, MSc (PhD Stdent)

Edvard Heikel (PhD Student)

(see here for contact details and more information)