Reading focus
Syllabus: How Do Brain and Body Support Mind and Behavior?
Follow why knowing that a region is active during a task is different from identifying the process occurring whenever that region is active.
Why you’re reading it. Establish a standard for interpreting the neural measurements encountered throughout this path.
A question to consider. A news report says a reward-related region activated while someone viewed a photograph, so the person must have enjoyed it. What comparison and background information are missing?
Why this reading
A brain region becomes active during a task. What does that tell us about what someone is thinking? Poldrack turns this familiar inference into an explicit question about how selectively the region responds, and shows why a compelling brain image is not enough to answer it.
Before you read
Functional magnetic resonance imaging (fMRI) measures changes in blood oxygenation associated with neural activity. Reverse inference proceeds from measured activation to a proposed mental process. Conditional probabilities depend on their direction: the chance of activation during remembering differs from the chance that remembering is occurring when activation is observed. Bayes' rule connects these quantities with the prior likelihood of the process.
Read for
- Distinguish a logically guaranteed conclusion from an inference that changes its probability.
- Follow how selectivity and the prior probability affect the language example in Table 1.
- Examine how the database's task categories and the size of a brain region influence the estimate.
Consider afterwards
- What comparison would help evaluate a claim that an activated region shows a person is afraid?
- Could an imperfectly selective brain measure still add useful evidence? Explain the conditions.
Editorial context
This argument does not prohibit inference from brain activity. It asks how much evidence an observation adds and what assumptions support that estimate. Poldrack's later decoding perspective describes more formal predictive approaches and their limits. Predicting a task or mental state does not by itself establish the mechanism producing it.