fMRI- what do we learn about cognition?

Question 1

Bonferroni-correction

Answer 1

• divide the significance level by the number of test (= voxels), and use the new significance level for each test: eg. 0.01/ 50,000 -> .000002 (corresponds to the overall risk of 1% to have one false positive.)

Question 2

reverse inference

Answer 2

• the FFA studies discussed in the last lecture highlight that there is a problem in how we arrive at an interpretation
• we ‘draw conclusions about cognitive processes from the presence of activation’- this is called reverse inference- because we already need to know that the activation means in order to draw conclusions about brain function.

Question 3

Reverse inference

Answer 3

• the main issue with drawing conclusions from cognitive studies is that when task A is happening and brain region Z is active, and in other studies when cognitive process X is happening, brain region Z is active.
• thus, in this study activity in Z __> engagement of cognitive process X
• the problem is that assumption 2 is not exclusive, brain region Z may be active for many other tasks, not just A, so it may also be related to other cognitive processes, not just X.

Question 4

addressing the issue of drawing conclusions

Answer 4

• if the brain region is activated using many cognitive processes, we learn very little from observing activation in those areas in a specific experiment
• but how do we know whether a brain region is also activated by other cognitive processes?
• the problem is we can also only infer this by observing activation in other experiments - reverse inference.

Question 5

example of reverse inference

Answer 5

• the regions in the prefrontal cortex are notoriously difficult to ‘understand’
• these regions appear to be activated by many different cognitive tasks

Question 6

Duncan (2010)

Answer 6

• proposed that these prefrontal regions are part of a multiple demand network, which computed many high-level cognitive processes
• rather than serving one particular function each, they might be recruited more strongly the more demand there is

Question 7

dorsal axis

Answer 7

• abstractness of rules

Question 8

ventral axis

Answer 8

• abstractness of memory representations and retrieval

Answer 9

• others are not as pessimistic about being able to learn from observing activation in these areas and suggest that there might be specialized, not so much about the task, but about how abstract the cognitive process is
• more anterior regions (towards the front of the brain) represent more abstract infromation
• more posterior regions represent more specific content

Question 10

reverse inference- end statement

Answer 10

• Duncan agrees that the prefrontal cortex does have some specialization. there are relatively more neurons in sub-regions recruited by different tasks, but there is no absolute specialization
• if the task is difficult enough, then neurons from prefrontal are recruited.

Question 11

task specificity

Answer 11

the second problem is that task A, which we have selected might not be ideal to measure cognitive process X.
- relatedly, if task A involves multiple cognitive processes, we do not really know how much it tells us about cognitive processes X
- in consequence, we observe specific brain region Z being activated, we still do not learn much about the cognitive process X

Question 12

Poldrack- task specificity

Answer 12

• Poldrack expressed these problems in probalistic terms. the probability that we can really learn from fMRI results that cognitive process X is involved depends on the
• the specificity of the task to measure the particular cognitive process
• the specificity of the brain region to reflect only this cognitive process

Question 13

Null results (interpretation of fMRI results)

Answer 13

• another issue is how to interpret ‘null findings’ (non significant statistical test results)
• if a contrast (eg. task A > task B) was not significant in a particular brain region, does that mean the region is not involved in the cognitive process?
• our statistical tests are designed to make it difficult for the H1, not for the H2, meaning that we can’t interpret null results.

Question 14

limitation and problems with BOLD fMRI

Answer 14

• the smallest measurement unit is a ‘voxel’, which is the 3D pixel, and the standard voxel size is a 3 x 3 x 3 mm
• this means we can learn nothing about what happens within a voxel
• however, since a single voxel contains >100,000 neurons, there will be a lot going on that we will miss.
• in consequence, our fMRI might simply not be sensitive enough to detect the differences between conditions, even if they exist.

Question 15

example of lim of fMRI results

Answer 15

• in a primary vision region, if we contrast vertical lines vs horizontal lines, we may not see any significant results
• the reason is that in each voxel, there are many neurons coding for all possible lime organizations
• this is due to the spatial resolution for fMRI. with a typical voxel size of 3 mm3, we have clusters of neurons, each firing for one orientation, all in one voxel
• if we have smaller voxels, we might be able to detect difference.

Question 16

2nd lim of fMRI

Answer 16

• temporal resolution: can we see the entire cognitive process unfolding in our brain images- because it takes 1-2 seconds to measure the entire brain once, we cannot see any changes taken place within this time period
• this is referred to as poor temporal resolution, we want to measure neural activity changes for fast processes, we need a different method (EEG, MEG)

Question 17

adding value to understand cognition

Answer 17

• the consequence is that we never conclude that brain regions are NOT involved in a cognitive process
• neuroimaging often seems to add value to our results, but that is not always the case, so always ask what you really learn from it. if you want to do more than just localizing a cognitive function, then clever experimental paradigms must be used