Visiting Researcher Presentation
Dr Stefan Bode
Decision Neuroscience Laboratory, Melbourne School of Psychological Sciences, The University of Melbourne
Title: Decoding decisions and post-decision errors from multivariate patterns of fMRI and EEG activity
Time: Monday 25 February, noon
Location: Keats Room, Psychology Building (former Aviation Building)
Brief Bio
Dr
Bode completed his PhD in the Department of Attention & Awareness
at
the Max Planck Institute for Human Cognitive and Brain Sciences in
Leipzig and the Department of Neurology, Otto-von-Guericke University Magdeburg,
in Germany under the supervision of Prof Haynes. Since 2009, he has held post-doctoral
positions at the University of Melbourne, first with Prof Philip Smith and then
with Prof Gary Egan. He now heads the Decision Neuroscience Lab at Melbourne
University
He has had extensive
fMRI, TMS and EEG experience at Leipzig and Melbourne, and his current work
extends the capabilities of these methodologies (while making it sound ever so
simple.. ). He has published in high impact cognitive neuroscience journals in
areas including mental rotation, visual attention, decision processes and task
preparation.
Abstract
How does the human brain accomplish decision-making? Decisions – whether rule-based, perceptual, internally driven, or economic – often have categorical outcomes. I will illustrate how multivariate pattern classification approaches (MVPA) for functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) data can shed light on neural mechanisms of decision-making by directly predicting choice outcomes from patterns of brain activation. First, I will provide an example of how decision rules can be decoded from prefrontal and parietal cortex using fMRI. I will then show how this method can be used to predict perceptual decisions under high and low discriminability, even when stimuli do not provide discriminative information. Under extremely low discriminability, perceptual guessing decisions resemble so-called “free decisions”, and their underlying activation patterns in parietal brain regions can become highly similar. Furthermore, novel MVPA methods based on event-related potentials (ERPs) allow linking of zero-discriminability perceptual decisions and pre-stimulus brain activity that could reflect the starting point setting in an evidence accumulation model framework. I will then extend this approach to the analysis of post-decision neural processes and show that ERP-based MVPA can also predict decision errors from brain signals after a decision is made but prior to response execution. Finally, I will give examples of MVPA applications to economic decision-making and show how consumer choices and financial decisions can be predicted from brain activity using fMRI. In conclusion, multivariate pattern classification provides a powerful tool for decoding the content of mental processes from brain activity and can be used for the analysis of both fMRI and EEG data. This, in turn, provides a basis for an advanced investigation and a better understanding of the neural processes underlying human decision-making.
How does the human brain accomplish decision-making? Decisions – whether rule-based, perceptual, internally driven, or economic – often have categorical outcomes. I will illustrate how multivariate pattern classification approaches (MVPA) for functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) data can shed light on neural mechanisms of decision-making by directly predicting choice outcomes from patterns of brain activation. First, I will provide an example of how decision rules can be decoded from prefrontal and parietal cortex using fMRI. I will then show how this method can be used to predict perceptual decisions under high and low discriminability, even when stimuli do not provide discriminative information. Under extremely low discriminability, perceptual guessing decisions resemble so-called “free decisions”, and their underlying activation patterns in parietal brain regions can become highly similar. Furthermore, novel MVPA methods based on event-related potentials (ERPs) allow linking of zero-discriminability perceptual decisions and pre-stimulus brain activity that could reflect the starting point setting in an evidence accumulation model framework. I will then extend this approach to the analysis of post-decision neural processes and show that ERP-based MVPA can also predict decision errors from brain signals after a decision is made but prior to response execution. Finally, I will give examples of MVPA applications to economic decision-making and show how consumer choices and financial decisions can be predicted from brain activity using fMRI. In conclusion, multivariate pattern classification provides a powerful tool for decoding the content of mental processes from brain activity and can be used for the analysis of both fMRI and EEG data. This, in turn, provides a basis for an advanced investigation and a better understanding of the neural processes underlying human decision-making.
Dr Bode’s visit is hosted by
A/Prof Frini Karayanidis – please contact frini.karayanidis@newcastle.edu.au if you’d like
to meet with Stefan individually, join us for lunch after the talk or for dinner
on Monday night.
