Neural correlates of expectation vs. expectation violation in delayed fear extinction learning

Poster No:

732 

Submission Type:

Abstract Submission 

Authors:

Isabelle Ridderbusch1, Adrian Wroblewski1, Yunbo Yang2, Hans-Ulrich Wittchen3, Andre Pittig4, Andreas Ströhle5, Jennifer Mumm6, Alfons Hamm7, Jan Richter2, Maike Hollandt7, Christoph Szeska8, Martin Lotze7, Volker Arolt9, Udo Dannlowski9, Katja Koelkebeck10, Dirk Adolph11, Jürgen Margraf11, Silvia Schneider11, Jan Cwik12, Jürgen Deckert13, Katharina Domschke14, Martin Herrmann13, Ulrike Lueken15, Ricarda Evens15, Constantin Rothkopf16, Winfried Rief1, Tilo Kircher1, Benjamin Straube1

Institutions:

1University of Marburg, Marburg, Germany, 2University of Hildesheim, Hildesheim, Germany, 3University of Dresden, Dresden, Germany, 4University of Goettingen, Goettingen, Germany, 5Charité Berlin, Berlin, Germany, 6Freie Universität Berlin, Berlin, Germany, 7University of Greifswald, Greifswald, Germany, 8University of Potsdam, Potsdam, Germany, 9University of Münster, Münster, Germany, 10University of Duisburg-Essen, Essen, Germany, 11University of Bochum, Bochum, Germany, 12University of Cologne, Cologne, Germany, 13University of Würzburg, Würzburg, Germany, 14University of Freiburg, Freiburg, Germany, 15Humboldt Universität zu Berlin, Berlin, Germany, 16University of Darmstadt, Darmstadt, Germany

First Author:

Isabelle Ridderbusch  
University of Marburg
Marburg, Germany

Co-Author(s):

Adrian Wroblewski  
University of Marburg
Marburg, Germany
Yunbo Yang  
University of Hildesheim
Hildesheim, Germany
Hans-Ulrich Wittchen  
University of Dresden
Dresden, Germany
Andre Pittig  
University of Goettingen
Goettingen, Germany
Andreas Ströhle  
Charité Berlin
Berlin, Germany
Jennifer Mumm  
Freie Universität Berlin
Berlin, Germany
Alfons Hamm  
University of Greifswald
Greifswald, Germany
Jan Richter  
University of Hildesheim
Hildesheim, Germany
Maike Hollandt  
University of Greifswald
Greifswald, Germany
Christoph Szeska  
University of Potsdam
Potsdam, Germany
Martin Lotze  
University of Greifswald
Greifswald, Germany
Volker Arolt  
University of Münster
Münster, Germany
Udo Dannlowski  
University of Münster
Münster, Germany
Katja Koelkebeck  
University of Duisburg-Essen
Essen, Germany
Dirk Adolph  
University of Bochum
Bochum, Germany
Jürgen Margraf  
University of Bochum
Bochum, Germany
Silvia Schneider  
University of Bochum
Bochum, Germany
Jan Cwik  
University of Cologne
Cologne, Germany
Jürgen Deckert  
University of Würzburg
Würzburg, Germany
Katharina Domschke  
University of Freiburg
Freiburg, Germany
Martin Herrmann  
University of Würzburg
Würzburg, Germany
Ulrike Lueken  
Humboldt Universität zu Berlin
Berlin, Germany
Ricarda Evens  
Humboldt Universität zu Berlin
Berlin, Germany
Constantin Rothkopf  
University of Darmstadt
Darmstadt, Germany
Winfried Rief  
University of Marburg
Marburg, Germany
Tilo Kircher  
University of Marburg
Marburg, Germany
Benjamin Straube  
University of Marburg
Marburg, Germany

Introduction:

Fear extinction is learning that a prior threat signal no longer indicates danger. An essential aspect of extinction learning is the violation of an expected outcome, in the context of fear specifically the expectation of an upcoming aversive event. This is experimentally operationalized via extinction training: a CS+ (conditioned stimulus, previously paired with an aversive stimulus) is repeatedly presented without an US (unconditioned aversive stimulus). A meta-analysis on neural correlates of fear extinction showed consistent brain activation in the cingulate cortex and insula in the overall CS+>CS- comparison in healthy subjects (HS) [1]. However, fMRI studies typically do not distinguish between expectation (EXP) and its violation (VIO) during CS-presentations. To close this knowledge gap, here we analyzed the beginning of CS-presentation (where the EXP of US/no US is set), separate from the timepoint at which the US would be present during conditioning (VIO) to investigate the following questions:

1) How does neural activation differ between the EXP of a potential aversive event and the VIO of this expectation? 2) Is there an interaction between CS-type and EXP vs. VIO?

Methods:

We used an optimized protocol of delayed extinction training [2]: For fear conditioning, one of two neutral visual stimuli (CS+) was followed (5 seconds after stimulus start) by an aversive electric US (electric stimulus, 60% reinforcement rate) while the other stimulus (CS˗) was never paired with the US. The US was applied by an MRI compatible electrodermal electrode attached to the inside of the non-dominant forearm. 24h later, uninstructed extinction over 20 trials was tested during functional magnetic resonance imaging (fMRI). Time courses of subjects' brain activity were acquired using 3-Tesla MR scanners equipped with a 12-channel head matrix receive coils. Functional images were obtained using a T2-weighted gradient-echo echo-planar imaging (EPI) sequence sensitive for the BOLD contrast (TE=30ms, TR=2s, flip angle 90°, matrix size 64×64 voxels, voxel size 3.6×3.6×4.0mm, slice thickness 4mm, inter-slice gap 0.4mm, field of view (FOV)=230mm, 33 slices, ascending phase encoding direction) [3]. A Family-wise error (FWE) correction was used. Quality controlled data-sets of n=103 HS from five sites in Germany were included.

Results:

1) Differential activation between the overall EXP and VIO during the full course of the experiment was found: activation in the MCC and anterior insula was higher during EXP, in the basal ganglia (nucleus caudate, putamen, pallidum), thalamus, middle frontal and temporal cortices activation was higher during VIO.

2) An Interaction between stimulus type (CS+ vs. CS-) and the EXP vs. VIO phase of the stimulus presentations was found in the inferior frontal gyrus, supramarginal gyrus and cerebellum, suggesting that EXP and VIO provide complementary information about neural processing of fear and safety signals. Separate post-hoc analyses for EXP and VIO respectively on CS-type associated differences revealed higher activation in the middle cingulate cortex, nucleus caudate and in the anterior insula towards CS+ than towards CS- during EXP, whereas in the paracentral lobule it was higher towards CS . During VIO, activation in the basal ganglia, the anterior cingulate cortex, the inferior frontal cortex including the anterior insula and in temporal areas and the cerebellum was higher towards CS+ than towards CS-, whereas in the medial orbitofrontal cortex and gyrus rectus it was higher towards CS-.

Conclusions:

Our findings suggest that the differentiation between EXP an VIO phase of CS-presentation provides useful additional information about extinction learning processes on the neural level. This finer strategy of analysis holds the potential to possibly better detect and understand altered processes in patients with anxiety disorders compared to HS in future analyses.

Emotion, Motivation and Social Neuroscience:

Emotional Learning 1

Learning and Memory:

Learning and Memory Other

Novel Imaging Acquisition Methods:

BOLD fMRI 2

Keywords:

ADULTS
Anxiety
Basal Ganglia
Design and Analysis
FUNCTIONAL MRI
Learning
Memory
NORMAL HUMAN
Other - Extinction Learning; Expectation and Violation

1|2Indicates the priority used for review

Provide references using author date format

[1] Fullana, M.A., Albajes, A. et al. (2018), ‘Fear extinction in the human brain: a meta-analysis of fMRI studies in healthy participants’, Neurosci. Biobehav. Rev., vol. 88, pp. 16-25.

[2] Hollandt, M., Wroblewski, A., et al. (2020), ‘Facilitating translational science in anxiety disorders by adjusting extinction training in the laboratory to exposure-based therapy procedures’, Transl. Psychiatry, vol. 10, p. 110.

[3] Ridderbusch, I. C., Wroblewski, A., et al. (2021), ‘Neural adaptation of cingulate and insular activity during delayed fear extinction: A replicable pattern across assessment sites and repeated measurements’, NeuroImage, vol. 237, pp. 118157.