University of California, San Diego, philosopher of the mind Patricia S. Churchland told me with unabashed skepticism: “Mental modules are complete nonsense. There are no modules that are encapsulated and just send information into a central processor. There are areas of specialization, yes, and networks maybe, but these are not always dedicated to a particular task.” Instead of mental module metaphors, let us use neural networks.
The brain is not random kludge, of course, so the search for neural networks associated with psychological concepts is a worthy one, as long as we do not succumb to the siren song of phrenology.
Editor's Note: This story was originally printed with the title "A New Phrenology?"
Subscribe
Buy This Issue
Already a Digital subscriber? Sign-in Now
If your institution has site license access, enter here.
Already a Digital subscriber? Sign-in Now
If your institution has site license access, enter here.
See what we're tweeting about
11 Comments
Add CommentGee, dis mus' be da shortest
Reply | Report Abuse | Link to thisShermer colum on record. Wassup dude?
The idea of modularity seems to rest on another assumption: that everyone's brain is hooked together in the same way. Maybe it has large general-purpose areas that hook up uniquely based on experience beginning at birth (or even before!).
Reply | Report Abuse | Link to thisThe evidence for the modularity of mind doesn't stem from fMRI results, but rather well documents phenomena wherein certain functions of the brain (eg object recognition) are completely damaged (eg by aphasia). While modularity may be misused to make claims about, say, the way swing voters think about politics, this doesn't in any way discredit the theory that there exist some discrete functions of the mind.
Reply | Report Abuse | Link to thisFurther, fMRIs don't really reveal much information about the brain at a functional level. The truth is, we don't really understand what it means for different areas to "light up" in terms of psychological states like beliefs, desires, etc. Seeing physically disparate areas of the brain become active doesn't entitle us to conclude that the underlying functional process isn't discrete. That may be the case, but we just don't know enough about what fMRIs show (if anything) at right level. It is very likely that the brain is modular, just not in the way Shermer divides it up.
There still must be something to modules. When a brain is damaged in the Broca's area, speech is impeded. When the visual area is damaged, vision is affected, and so on. It is not the whole story, but it is far from "complete nonsense."
Reply | Report Abuse | Link to thisI think there's some confusion of terminology: the 'modularity' that the author is arguing against is the notion of encapsulated brain structures that have a single function.
Reply | Report Abuse | Link to thisIt's obvious that different brain areas specialize for certain tasks. The anti-modularity folks are not trying to argue against that. They're saying (1) There's no brain area that does only one thing; and (2) Specialization is a process driven by what data is routed where, not a structure defined in the genes; and (3) a lot of very specific tasks (e.g. language and theory of mind) result from a lot of areas with more general functions collaborating -- not from domain-specific modules.
The degree of sensitivity, and the line pair resolution of fMRI is not adequate for making all but the most general of assumptions. Brocca had better spacial resolution, and probably had better hints at cellular function than that which MRI provides.
Reply | Report Abuse | Link to thisTalk about getting ahead of the data.
Anybody in the mood for some philosophy?
Reply | Report Abuse | Link to thisI thought a more interesting statement was “[i]Science traffics in metaphors because our brains evolved to grasp intuitively a world far simpler than the counterintuitive [b]world that science has only recently revealed[/b][/i].†Ah! The truth shall set us free.
Let’s see, J.J. Thomson at some point revealed that the Greeks and Dalton had it too simple – what’s really going on is “plum puddingâ€: electrons embedded in a gel of positive charge. Oh no, said Rutherford, you’re overusing the metaphors. Let me reveal the true nature of atom to you: the electrons revolve around the nucleus like the planets revolve around the Sun. But wait, here comes Niels Bohr and reveals it to us again: the planets would collapse, there are actually electron shells and electrons jump between them emitting photons. Did we finally get it? Nope. Here comes De Broglie’s revelation: electrons have wavelike nature and exist at all places in the allowed orbits at the same time. So, have we finally revealed the world or should we wait for the genius who will unite Relativity and QM to reveal it to us one more time? Should I also mention the idealized truths of Euclidian geometry which now we’re told fly out of the window as soon as you introduce mass?
Pardon my theatrical rhetoric, but maybe a more sober approach would be to acknowledge that there is no world to be [i]revealed[/i]. For one, it’s kind of arrogant to presume that you finally revealed it. Secondly, do you even have the criteria that tell you when you finally get there? Maybe, instead of revealing the reality, we need to say that we [i]conceptualize[/i] the truths with available to us notions that we acquire through learning language. How do we learn? We generalize what we see and use analogies – metaphors. We then mentally manipulate these generalizations and metaphors to [i]construct[/i] more sophisticated models that yield more precise predictions in dealing with the environment. But achieving higher predictions to solve certain problems is all there is to it. To say “oh, and by the way, there’s this truth out there and we just got closer to it†is unnecessary, unparsimonious, and it flies in the face of Occam’s razor, something science uses as a reliable heuristic in its endeavor.
My point is, “[i]science traffics in metaphors[/i]†NOT “[i]because our brains evolved to grasp intuitively a world simpler[/i]†but because metaphors is the only thing we’ve got. What replaces the "modules" is also a metaphor which will sound inadequate as we get more sophisticated. Although Feynman encourages us to “just do the mathâ€, mathematics can get you only so far in understanding the world, and that’s presuming that math itself is not a type of “dry†metaphor that we conceptualize to achieve better predictions, and I would argue that’s an implausible presumption. So what I propose is we first need to stop painting a confusing picture that will perpetually give rise to articles like this one, the type that suggest that the reality out there is more complicated than our primitive metaphors. We should acknowledge that if we want to interact with the world in more productive ways, instead of abandoning metaphors, we need to continuously develop new ones that make sense out of experimental predictions. Instead of “just doing the mathâ€, I wish Feynman had challenged philosophers of science to come up with ways of conceptualizing Quantum Mechanics by developing a new set of perspectives and visualizations. Same goes for the brain. Don’t tell me that modules in the brain is just a metaphor, that that’s not how things are. Please tell me how I can visualize a new model of “how things are†by offering a more sophisticated set of metaphors. That, I think, would be much more helpful in understanding science.
How's a network not any more of a metaphor than a module?
Reply | Report Abuse | Link to thisWith no argument for why the organization of the brain is not modular (other than Patricia Churchland says so - I wonder why the author didn't ask Jerry Fodor for his opinion...) this article does more of a disservice to lay people wanting to learn about brain and cognitive sciences. I hope the author and the magazine do a better job of presenting this debate in psychological science in the future by replacing the opinions of scientists and philosophers with data from science and arguments from philosophy and then letting the reader figure out what to think. (I never knew it was the job of SciAm to take sides on far from resolved issues in science.)
Shermer quotes Poldrack saying "...taking a bit of science and going way beyond it." But who's really going way beyond it? I think Shermer and Poldrack are. It is amusing to see my name associated with brain modularity. I support the concept of modularity in the brain as much as Sam Harris and Richard Dawkins support religious beliefs or the Christian evangelic right supports stem cell research, abortion, and gay marriage. There is nothing in the New York Times op-ed "This Is Your Brain on Politics", that advocates modularity. It is surprising to see how Shermer embraces Poldrack's gross distortions of the assumptions behind the New York Times op-ed. As a publisher of Skeptic, shouldn't Shermer have a more skeptical attitude?
Reply | Report Abuse | Link to thisIt is unlikely that there are modules [i]a la[/i] Fodor in the brain. Indeed, not even Fodor probably believes it anymore. However, there is a large peer-reviewed scientific literature that associates mental states with brain structures. It is possible to use such literature to interpret brain activity as 'probabilistic marker of mental states.' For instance, there are more than 1,000 peer-reviewed published papers that associate the amygdala with anxiety. There are also some papers that associate the amygdala with happiness. How many? Less than 100 (I did this survey about six months ago, this is why I am giving approximate numbers.) It is irrational to assume that activity in the amygdala can be associated with anxiety or happiness with equal probability.
Shermer's article fails to say that 'reversing the causal inference' is a common practice in neuroimaging. The beautiful study on the neuroanatomy of belief recently published by Sam Harris and colleagues takes the activity in the insula while reading statements that subjects did not believe as suggesting that disbelief invokes disgust. There is nothing in Sam Harris study that tells us that subjects were truly disgusted. A recent brain imaging study by John Cacioppo on social isolation interprets activity in the striatum as representing reward processing. There is nothing in Cacioppo's design that can demonstrate that the activity in the striatum truly represents processing of reward. Ironically, Cacioppo also co-signed Poldrack-led letter to the Times quoted by Shermer. I could probably find examples of 'reversing the causal inference' for all the signatories of that letter. Indeed, I would argue that even Poldrack himself exploits some level of reverse inference in his own studies. In a recent paper led by Adam Aron (he also signed the letter to the New York Times) on response inhibition, Aron and colleagues interpret activity in pre-SMA (a frontal lobe area) as reflecting a 'conflict/detection resolution' role. While this interpretation is very reasonable, it would be much less reasonable if the brain activity was located in primary visual cortex. Aron and Poldrack would have not interpreted activity in primary visual cortex during a response inhibition paradigm as reflecting a 'conflict/detection resolution' role of the primary visual cortex. Anatomical location of brain activity matters, because there is not a single study that can reveal the absolute truth. Incidentally, this notion applies to all science, not just brain imaging. Science is about reproducibility of results. The results from individual experiments must always be compared to a larger corpus of data. As Francis Crick said, " Any theory that can account for all of the facts is wrong, because some of the facts are always wrong. " It is highly unlikely that more than 1,000 scientific papers are wrong (amygdala and anxiety), but it is not inconceivable that less than 100 papers are wrong.
Our New York Times op-ed applied this rational probabilistic logic to brain responses in voters watching candidates. By doing so, it also provided a splendid example of how one can do civic education by using scientific constructs and rational thinking for issues that matter to people. Sadly, science has still a marginal role in our public discourse and this is in part due to an 'ivory tower' attitude of many scientists that are afraid of mixing the 'pure science' of the lab with real life issues. A product of this attitude is a society in which basic concepts that emerged from science - for instance, evolution - are challenged with the irrational position that the evolutionary framework can't explain all data.
--
Edited by Marco Iacoboni at 06/03/2008 3:13 AM
Schermer and Churchland go too far in denying brain modularity when it is self-evident as several others have pointed out. They don't go away when relabeled "areas of specialization" either. Certain functions are physically modularized within specific brain nuclei and interconnected by discrete pathways to form a network that is true but do not throw out the baby with the bathwater please, brain modularity is an important concept in understanding its functioning.
Reply | Report Abuse | Link to thisSchermer and Churchland go too far in denying brain modularity when it is self-evident as several others have pointed out. They don't go away when relabeled "areas of specialization" either. Certain functions are physically modularized within specific brain nuclei and interconnected by discrete pathways to form a network that is true but do not throw out the baby with the bathwater please, brain modularity is an important concept in understanding its functioning.
Reply | Report Abuse | Link to this