tracking the neural dynamics of hypothesis evaluation with ...€¦ · -3.9 -2.3 2.3 4.4 tracking...
TRANSCRIPT
-3.9 -2.3 2.3 4.4
Tracking the Neural Dynamics of Hypothesis Evaluation with Model-based fMRI Nikki Marinsek, Benjamin O. Turner, Chloe Steindam, & Michael Miller University of California, Santa Barbara
Introduction
fMRI procedure
Model & parameter estimation
Model predictions fMRI results
Summary
References
EASY WORD SETS Designed to minimize hypothesis formation & evaluation
Fifteen subjects’ brain activity was recorded with fMRI as they guessed what categories groups of words belonged to. Subjects pressed a button when they discovered the appropriate category and reported their accuracy after each trial.
Exp. 1: present first word, sub-jects list possible categories
Exp. 2: present category, subjects list category members
word frequency
HARD WORD SETS Designed to elicit repeated cycles of hypothesis formation & evaluation
This research was funded by the Institute for Collaborative Biotechnologies under grant W911NF-09-D-0001.
fMRI results
FIGURE 1: Hypothesis certain-ty increases over the course of the trial, and tends to increase faster in easy trials.
FIGURE 2: Word sets that have low predicted certainties tend to be solved slowly. The line of best fit is shown in gray.
FIGURE 3 Brain areas in
CORAL are more active when
hypotheses are CERTAIN and
appear to reflect DMN activity.
Areas in NAVY are more active when
hypotheses are UNCERTAIN and
may support hypo-thesis making.
FIGURE 4 Brain activity in the inferior frontal sulcus (white ROI in FIG 3) is high when the predicted hypothesis certainty is low.
2.3 3.4
▶︎
◀
FIGURE 5: Brain areas that are more active when the P(O|H) IS HIGH (that is, when the presented word is representative of the category).
FIGURE 6: Brain areas that are more active when the HYPOTHESIS CERTAINTY INCREASES.
1. Marinsek, N., Turner, B. O., Gazzaniga, M., & Miller, M. B. (2014). Divergent hemispheric reasoning strategies: reducing uncertainty versus resolving inconsistency. Frontiers in human neuroscience, 8.
2. Turner, B. O., Mumford, J. A., Poldrack, R. A., & Ashby, F. G. (2012). Spatiotemporal activity estimation for multivoxel pattern analysis with rapid event-related designs. NeuroImage, 62(3), 1429-1438.
We have proposed that the left hemisphere, and in particular the left vlPFC, strives to reduce uncertainty by creating explanations, and the right vlPFC strives to resolve conflict between beliefs and evidence.1
To test these predictions, we built a Bayesian model of participants’ hypothesis making behavior and used the predictions of the model to identify brain regions that are sensitive to uncertainty and belief updating.
Consistent with our predictions, the left vlPFC is more active during uncertainty, suggesting that it plays a role in inference making. We did not find any evidence that the right vlPFC supports belief updating. However, this may be due to a lack of power since very few TRs were associated with decreases in the P(H|O).
◀
◀
◀
◀