Distinguishing deception from its confounds by improving the validity of fMRI-based neural prediction
Sangil Lee, Runxuan Niu, Lusha Zhu, Andrew S. Kayser, Ming Hsu · 2024
Read Distinguishing deception from its confounds by improving the validity of fMRI-based neural prediction by Sangil Lee, Runxuan Niu, Lusha Zhu, Andrew S. Kayser, Ming Hsu (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11648856.txt.
Description
Sangil Lee and four co-authors show that brain-based lie detection mostly detects something else, then fix it. Their 2024 study puts participants in signaling games where deception pays, and trains machine learning on brain scans to predict the lying. The first result validates the skeptics: much of the predictor's power came from side processes, since it also predicted behavior in a control task with no deception at all.
The repair adds a second goal to training. Alongside predicting deception, the predictor must perform at chance on the control task. That strips out the confounding signals while keeping the deception-related ones. Validity, measured as the gap between the two tasks, improved. The authors offer the dual-goal tuning trick beyond lie detection. It is a general way to make neural predictors mean what their labels claim, in a field whose promises have long outrun its proofs.
Catalog record
| CREATOR | Sangil Lee, Runxuan Niu, Lusha Zhu, Andrew S. Kayser, Ming Hsu |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |