Article: Fischer, J., J.G. Mikhael, et al. “Functional Neuroanatomy of Intuitive Physical Inference.” PNAS 113, no. 34 (2016): E5072–E5081. DOI: /10.1073/pnas.1610344113

Assignment: Read the assigned article and think big about the problem it addresses. Think of this article as essentially a first baby step in a broader line of work. What questions does this article answer, and what do you want to know next?

For the next assignment you will have a significant written experimental design assignment (15% of your course grade) designing your own experiment (or two) to ask one of the next questions about intuitive physics in the brain, using any of the methods you have learned about in the course.

For now you are asked only to read this article and answer questions about it. But as you read it, start thinking about the broader space of questions this article opens up, and what you might propose for the larger experimental design assignment.

Provide short answers to the questions below. Be clear and concise—points will be taken off for unnecessary words.

1. What is the main question addressed in this article?

How does the brain do physical inference? Does the brain have a physics engine?

1. What is the main answer the article provides about that question?

The article identifies regions that respond more to physical inference tasks. However these regions are not exclusive to physics.

1. Briefly contrast the logic of Experiment 1 and Experiment 2. In what way are the two designs complementary?

Experiment 1 keeps the stimuli the same, but tests the difference between a physical and visual task. Experiment 2 is complementary in that it keeps the task the same, but varies the stimulus in a minimal way. The combination of the two experiments isolate that the physical inference is causing the selective activation, not properties of the task or the input.

1. Scrutinize the data shown for regions P3R and P1R in Figure 3B. What do these data suggest to you?

Both regions are more active for physical tasks. P3R is a larger region so there are wider confidence intervals.

1. Are the frontal and parietal regions identified in this study highly specialized for intuitive physical reasoning per se?

No they are utilized for other tasks such as action planning. It is part of the multiple demand network which is general purpose.

1. List three big questions you might want to answer next in this line of work, and for each one list a method that might be able to answer (or start to answer) that question.

Data driven analysis would be interesting here. There could be principal components for physical inference. We could annotate the input with physical properties. For example see if the brain reacts differently to heavy objects versus small objects.

What’s tested here is mental tasks, but motor skills tasks is more akin to the physical inference we do more often. These tasks can also be normalized for difficulty.

Of the other tasks in the MD, which are more correlated with physical inference. Does each task activate at different regions. Principal components would be useful here.