Tuesday, January 6, 2009

A Question of Value: What Does Cognitive Neuroscience Really Contribute?

I was checking out some of the blogs listed on Sci's blogroll today and found myself over at PsyBlog reading this post on whether or not the field of cognitive neuroscience is really contributing anything to the study of "mind." It is a fascinating discussion. 

Cognitive Psychology is not my field, neither is neuroscience. That said, I have a decent level of knowledge of CogPsy and I think I know enough about current neuroimaging techniques to post about it on my own blog, though I neglected to comment on the original post. 

The main question being asked is this:

"There's no doubt the mind's cognitive processes are a function of the brain's physiological activity but these two things are nevertheless (currently) separate questions. Cognitive neuroscience's strength is in physiological processes, and as imaging technology improves, so will the importance of its findings in this area. But, again, why should a psychologist care that much which part of the brain lights up in a scanner, if the mind's functioning is still so far removed from our understanding of its physiology?"

Debate ensues in the comments section about this "split view" of mind versus brain followed by an entirely pointless discussion of whether the "hardware versus software" analogy is appropriate. 

An anonymous commenter provides the following brilliant analogy for using fMRI as a means of understanding cognitive processes:

"Here's the problem as I see it.... using current cognitive neuroscience techniques to understand brain-behavior relationships is currently like trying to understand the functioning of a car's motor by measuring the heat patterns on the hood of the car. We can tell when it's working harder (and which areas show the biggest changes), but that's a far cry from truly understanding motor function."

Yes! Abso-fuckin-lutely! Now don't get me wrong - I love sexy fMRI images just as much as the next gal, but answering the question "WHERE" is indeed a far cry from answering the question of "HOW" - which is what Cognitive Psychology traditionally pursues. 

The problem that is leading to this debate in psychology is the fact that those sexy fMRI images are dazzling the sources that provide funding and diverting much-needed dollars to these programs when they could be supporting more "valuable" lines of research. Cog-Neuro has become the new "sexy" field of psychology, drawing more and more students and money each year. 

I agree that a LOT of money is being wasted on these programs right now. We cannot learn a whole lot looking at a computer screen saying "What part of the brain lights up when we ask someone to memorize a list of words" compared to "What processes underly HOW a person memorizes that list of words?" The latter question cannot YET be answered from an fMRI image. 

However, I wholeheartedly support this branch of neuroscience, especially the funding that goes to support new advances in the technology. The problem is when money is being sent off to researchers so they can play their new toy - scanning for the sake of scanning. 

The fMRI is totally fucking cool, and if I had one I would play with it ALL THE TIME. My walls would be covered in poster-sized reprints of fMRI images with titles like "This Is Your Brain While Masturbating" and "This Is Your Brain While Listening To Nine Inch Nails." It would be awesome. 

But it would also be a tremendous waste of time. 

The fMRI is a valuable tool in all branches of psychology, but it needs to be remembered that for all intents and purposes, it can only tell us WHERE something is happening. Combining that with lesion studies, single-cell recordings, EEGs, and TMS (transcranial magnetic stimulation), and the brain mapping that neurosurgeons do can provide us with a wealth of information about the brain. But in truth, it tells us very little about cognition. 

We just don't have the technology yet that we need to study the human brain while preserving ethics. And sometimes, even the technology we DO have, that IS ethical, is still difficult to find participants for, as in the case of TMS. Imagine a researcher tells you, "We're going to apply this magnet to your head, and it's going to turn off a part of your brain for a little while. Is that cool?" Most people don't, in fact, think that's cool, even if it is temporary. 

The argument for Cog-neuro has been that once we have a whole bunch of data, we'll have some answers. Sure, that's great. But answers to what? And do those answers justify the billions of dollars spent on these studies? I'm guessing probably not, but I could be wrong. 

The bottom line is this: Cog-neuro IS valuable as long as everyone remembers that the amount of speculation going into the analysis remains unchanged whether you use a computer model or an fMRI. There is a lot of very cool data coming out of this discipline, but it's not particularly useful in a manner that justifies its level of funding and visibility. Some people are hailing the fMRI as a divine gift to brain science, and it's just simply not the case. Let's not ignore the fantastic techniques and methods that traditional neuroscience has given us, because without using those in combination, the fMRI is practically useless. 

Now I know I have a lot of neuroscience blog buddies out there who might disagree with my assessment depending on their specialty. I invite you to do so, as I am more than open to learning something new. And if any of you have answers to the following questions, I would really love to hear them because I have not been able to figure them out:

- How can an fMRI image explain how the brain processes information from short-term and working memory into long-term memory, and why some things are lost while others are kept forever? (I know we have theories, I just want to know how this particular tool can help)

- How can fMRI images explain how mental imagery works? We know from fMRI scans that the visual cortex is activated when imagining a scene just like when you're actually viewing the scene, but what can it tell us about how and why?

- How can the fMRI aid in our understanding of the serial position effect or the learning curve? 

I have many more, but I think I'll stop with these. I really need to start posting blogs when I think of them rather than hours later, because this was much better organized and complete in my head around 3pm today than how it's turned out. :(

12 comments:

Anonymous said...

The problem that is leading to this debate in psychology is the fact that those sexy fMRI images are dazzling the sources that provide funding and diverting much-needed dollars to these programs when they could be supporting more "valuable" lines of research.

It's also filling pages of high-impact-factor journals with garbage at the expense of decent science:

http://scienceblogs.com/cortex/2009/01/voodoo_fmri.php

JLK said...

Thanks for the link, CPP. Though now my blogroll is becoming excessively long because of my new adventures through the blogosphere!

When I posted this, I didn't realize what was going on out there because social neuroscience isn't my thing. The weird coincidence is I just started reading "Social Intelligence" by Daniel Goleman last night, and until then I hadn't heard of it before. Suddenly today I'm finding out that it's part of this "scandal."

Strange timing, seriously.

Anonymous said...

since i focus a lot more on neurochemistry/neuropharm than on functional assays, i am not an expert on functional assays. however, the whole lack of ability to describe a cause/effect relationship, among many other things, causes me to take any fMRI study i run across with a handful of salt.

daedalus2u said...

To me, the most important take-home message from fMRI is how important NO is in the brain.

What the fMRI BOLD technique actually measures is acute changes in O2Hb level caused by acute vasodilation on sub-second time scales. That vasodilation is caused by neurogenic NO exceeding the threshold for sGC, generating cGMP and causing vasodilation. That NO is the mechanism by which the brain meta-programs itself.

That shows that basal NO is critically important in the regulation of brain activity. The NO generated to cause neuronal activation adds to the basal NO that is already there and when the sum exceeds the activation threshold is when stuff gets activated. The basal NO level is an important and necessary component of the NO that causes that activation.

This is how stress changes the meta-programming of the brain, by changing the range of the NO signals.

JLK said...

Daedalus, I have nothing but respect for your love affair with NO.

That said, I have no idea what the fuck you're talking about. I'm a social science gal, D.

"That vasodilation is caused by neurogenic NO exceeding the threshold for sGC, generating cGMP and causing vasodilation." - It may as well be written in hieroglyphics for me.

Anonymous said...

I would simply caution that in a brain, where the processing algorithms are manifest in the physiological structure, 'where' is a very large part of 'how'.

JLK said...

@ James Dean:

When I question the "how" I am thinking in terms of information that is beneficial to cognitive studies, not necessarily the physiological "how."

When we are able to figure out ways to improve learning in children through fMRI studies, we will then be making progress in cognitive neuroscience. But so far, nothing (that I am aware of) that these studies have shown us have any practical applications whatsoever to the real-world problems that cognitive psychology attempts to address.

Donna B. said...

I see both James Dean's and JLK's point, but I have to side with Dean.

The 'how' is part of the second (third? fourth?) set of questions. Can you get anywhere without answering 'where' first

Even when where, what, how, and when have been answered, translating all that knowledge into practical applications may take some time and perhaps some non-intuitive directions.

I do agree that it's time to move on the next stage instead of simply repeating fMRIs one after the other because we can.

JLK said...

James Dean and Donna - if you haven't had a chance to read the PsyBlog posting that inspired this post, take a minute and check it out.

Or google "social cognitive neuroscience contraversy" and see what comes up, because there has been a sudden shitstorm in the field over an article showing that faulty statistics used in fMRI studies led to inflated, actually impossible correlations.

The researchers "forgot" that the probability of two events occuring together cannot be greater than the probability of each occuring separately.

Whether the fMRI itself is useful is not in question. The problem is that some cognitive neuroscientists are trying to stretch fMRI-only data to make conclusions about cognitive theories. The author of psyblog asks (not an exact quote) "How can an fMRI be used to compare two theories?"

I'm not looking to argue against fMRI as a tool in cog-neuro, but there are some very important questions and evaluations that need to be taken into consideration before we continue to invest billions of dollars into this line of research.

There's been way too much, "Wow! The area that I thought would light up during this task lit up on the fMRI! Therefore, my theory that children learn lots of valuable skills from playing video games must be correct!"

Huge leaps are being taken to connect the "where" to the "how" and/or "why." Which is how science works when generating new research and tools for discovery. But we have yet to see a transition to the middle-ground experiments where step-by-step analysis evaluates the "how" and "why" that were initially arrived at by that huge leap.

Anonymous said...

Hey there, CogPsy from physioprof here. I don't have a blog, sadly. I have a dissertation in the works - kind of kills all other pursuits.

I think cognitive neuroscience CAN be well-done and can make substantive contributions if done correctly. There are actually a few decent fMRI studies out there, really, there are. In fact, we've got some collaborators in Belgium that we work with on fMRI stuff. There's also been a lot of advancement in the actual fMRI analyses that makes the method more precise in both space and time, and some new techniques can eliminate some of the sketchy statistics involved (see, for example, http://www.psy.vanderbilt.edu/tonglab/publications/YamashitaEtAl_2008_Neuroimage.pdf). The problem is wading through all the crapola to find these few gems.

I come from a "break the model" training background - take a model and break it's assumptions and/or pit it against another model that makes an opposing prediction, then see what happens. The conclusions are pretty damn clear in these types of studies. As I said over on physioprof, the key is that a neuro study has to be good on it's own - WITHOUT the fMRI portion. Say I find that people's behavior is totally different in two conditions. If I can also show that the areas of the brain they use in those conditions are completely dissociated, that's pretty strong evidence that they are using an entirely different strategy/information set/etc. to do the task. But, without a compelling behavioral result, I don't think we have any business AT ALL doing fMRI. Certainly "where" is a big part of "how", but if you don't have a decent behavioral study in the first place, then you are missing the "WHAT". And that is a very serious problem. The funding agencies are only exacerbating it by pouring money into neuro and other applied research while starving basic research.


@ Daedalus - are you familiar with Randy Nelson's work with gene-knockouts and NO? I think you'd love it. What's not to love about psycho mice?

JLK said...

Thanks for stopping by, CogPsy!

You're absolutely right about some fMRI and other neuroimaging studies that are valuable for what they bring to the table. Those studies are usually (but not always) the ones that end up in textbooks, and it's fantastic to have them.

What frustrates me primarily about cog and soc neuroscience is the amount of money being diverted to them without regard to the quality and value of the science they are producing. As an undergrad, I was advised that if I had the slightest interest in neuroimaging to make that my focus because that's where the money is and the jobs are.

It holds a lot of promise for the somewhat-distant future, but I feel that right now it gets a disproportionate amount of press and financial attention. I would prefer to see those $$ diverted into advancing the technology first, and using the existing technology second. I also think, to echo your statements, that we need to train students in these fields to avoid disregarding the behavioral components of their research. I think too many of them are doing just that.

viagra online said...

cognitive neuroscience didn't work at all, we need to change all its principles.

Related Posts Plugin for WordPress, Blogger...