I know some of you out there are in the midst of chaotic semesters, already starting to feel like the grasshopper who fiddled away the long days of summer without preparing enough lecture notes, and now must squeeze class preparation into days that feel too short, and continue to get shorter. But at the risk of coming off like Marie Antoinette extolling the virtues of dairy farming, I have to admit that I am really enjoying getting ready to teach my seminar course this semester.
I inherited the title — “Fundamentals of Cognitive Neuropsychology” — and, since I teach as an adjunct and have no authority to change it, just went ahead and arranged the syllabus to be like an introduction to cognitive neuroscience despite the title. The first set of readings we cover are about historical foundations of the field. I use two chapters from Finger’s Origins of Neuroscience, one on early cortical localization, the other on holist critics of localization. This sets the tone for a course introducing a field where there are still more questions than answers, and where many of the questions are in fact very, very old.
My favorite anecdote from this set of readings concerns a debate between Friedrich Goltz, a German holist, and David Ferrier, a British localizationist. The setting of the debate was a huge international (i.e., all of Western Europe was represented) conference on medicine, which included a reception at the Crystal Palace, and was attended by over a hundred thousand people, including royalty and various dignitaries. In a morning session, Goltz described dogs with multiple lesions who did not appear to lose any specific functions, whereas Ferrier described monkeys with smaller lesions, who did have specific impairments. Since Goltz had brought one of his dogs with him(!) and Ferrier worked right around the corner at King’s College, it was agreed that the various luminaries and interested parties would reconvene in the afternoon to examine the animals. The demonstration went well for Ferrier, whose monkeys demonstrated a gross double dissociation between deafness and hemiplegia. He was further able to show that Goltz, despite making large lesions in his dog’s cortex, did not destroy key portions of motor cortex.
From Finger’s description, this seems to have been a pivotal event in establishing consensus about localization of function in cortex. It is an odd anecdote because on the one hand it presents a very early model of the ethics that drive open science: both scientists (Goltz likely at great expense and inconvenience) brought their data before a large audience for general inspection and analysis. At the same time, the rarified, luxurious setting — Ferrier was treated to a champagne toast for his victory — and the image of a group of mustachioed gentlemen and their wealthy benefactors stroking their whiskers and declaring that a scientific consensus had been reached make this hard to reconcile with how we do things today.
I don’t mean to minimize the role of prestige and hierarchy in contemporary scientific discourse — indeed, discussion here and here of how “glamour mags” function in the field suggest that this situation has not so much evolved as oscillated over time. Nor would it be wise to pretend that hierarchical advantage and prestige don’t accumulate inequitably to the essentially the same demographic groups as in the past — specifically, white men.
At the same time, it is hard to imagine any kind of scientific consensus being established by a single dramatic demonstration involving a single dog and two monkeys. We don’t settle disputes by scientific duel anymore although the Neurobiology of Language Conference has experimented boldly with this, producing interesting results. But we are still pacing much of the same territory.
What’s striking in reviewing the early debates about localization vs. holism largely is that the more interesting arguments are actually about what kinds of things we are willing to countenance as “mental functions.” When we talk about phrenology, for instance, it is very straightforward that Gall and his colleagues were wrong about the relationship between the skull and the brain — briefly, anatomists eventually figured out that your brain doesn’t make your skull lumpy. On the other hand, although it seems obvious that they were way off base about the kinds of mental functions they were trying to localize, this is harder to explain.
Gall’s initial work included associations between anatomical features and things like “destructiveness” or “veneration.” Spurzheim studied women who had been convicted of infanticide and found malformations in a region he associated with “philoprogenitiveness,” or the ability to love children. The exercise I ask students to do is to try and explain why these seem like poor candidates for localization. It takes time to dig beneath the intuition that “That’s not, like, A Thing: primogeniwhatever,” to appreciate that without some kind of explanatory framework for what mental activity is like — e.g., experimental psychology — the hunt for structure-function relationships is bound to produce a wunderkammer of idiosyncratic observations.
Here we can see that the science really has evolved over the last century or so. Cognitive neuroscientists do shake out into “holist” or “localizationist” camps, but only roughly. No one can look at maps of the primate visual system and claim that vision is accomplished by the action of an undifferentiated mass of tissue. On the other hand, defining the specific “function” of even the first step in that processing stream with any precision is viciously complicated, in ways that Ferrier and colleagues could not have imagined. The problem doesn’t seem to be that we lack the tools to identify brain regions that have distinctive functional properties. (Perhaps because I’m not an anatomist, I tend to think the anatomists have really got their act together with respect to primary visual cortex.) It’s more likely that we haven’t quite figured out what vision is,, how it is broken down into component functions that we can talk sensibly about finding in the brain.
Vision is an interesting example because it’s something that even early holists would concede was likely to have some degree of cortical localization. Lashley, Jackson, and other famous holists were willing to accept that “lower” functions like sensation and motor control could be localized (especially after the dog vs. monkey debate). What they doubted was that “higher” functions like attention and reasoning could similarly be pinned down to specific regions on Dr. Brodmann’s map, which just celebrated its centenary.
When we start trying to talk about localizing functions like Language, Attention or Reasoning, the importance of defining just what it is one is trying to localize is critical to getting any kind of neuroscientific inquiry off the ground becomes clear. What abilities should be impaired in a patient who has lost her Language faculty as a result of a brain lesion? More importantly, what abilities should we expect to be spared? These turn out not to be straightforward questions, and the answers from the aphasiology literature are better summarized by someone with Oliver Sacks’s novelistic appreciation of the uncanny than by the kinds of dry declarative sentences we might hope to fill textbooks with.
You don’t find real “holists” in contemporary cognitive neuroscience, what you find is people who suspect that mental processes, as defined psychologically, tend to be subserved by distributed networks of regions whose functions are relatively general and context-dependent. The contemporary localizationist “camp,” then, differs largely as a matter of degree: everyone agrees that there is some specialization of function in cortex, localizationist arguments are arguments that those functions are specific and well-aligned with some psychological description.
So we have to be careful, when drawing a historical lineage from someone on my side of the argument back to Lashley or Jackson, or describing someone like Nancy Kanwisher as an intellectual descendant of Ferrier or Broca, not to ignore how these positions have evolved. Early proponents of localization were often vague about both structure and function in ways that are easy to paper over when use historical-anatomical terms like “Broca’s Area” to describe the results of modern experiments.
Broca’s famous patient’s loss of speech was relatively non-specific, both in how gross impairments in the ability to produce speech were taken as evidence that he had lost his “Language” ability, and in the fact that he was also globally apraxic — meaning he had serious motor control issues that extended beyond the articulation of speech sounds. In the neuropsychological literature of the following century, “Broca’s aphasia” came to refer to a much more specific set of symptoms having to do with grammatical processing. The question is whether grammatical processing is an isolable function of the mind that has a corresponding structure in the brain — that is, whether grammatical processing is “like, A Thing” — is still a vexed one. Alternatively, grammatical processing could be conceived of as a specific example of a more general function — say, keeping track of sequences, or executing serial actions — and that this is actually the appropriate characterization of what the brain regions associated with grammatical processing are actually “for.”
“Broca’s area,” on the other hand, is used to refer to a dog’s breakfast of subregions of the inferior frontal gyrus. A recent meta-analysis by Lindenberg, Fangerau and Seitz found that studies claiming to have found activity in Broca’s area using a variety of imaging methods had identified an anatomically heterogenous jumble of regions throughout the inferior frontal cortex, with no coherent ascription of functions to distinct subregions. That is, although particular scientists have particular structure-function relationships in mind when they use the term, the field at large applies it arbitrarily to refer to broad swath of tissue that probably does not have an isolable mental function.
This, I think, is a great way to introduce students to cognitive neuroscience. It is like throwing them into the deep end of the pool. Ask a dozen bright undergraduates what they know about structure-function relationships in the brain and one of them will tell you that “Broca’s area” is “for language.” By sensitizing students to how complex and historically fraught both the anatomical and functional definitions actually are, they can develop a sense of the difficulty of asking good questions about the brain (much less getting satisfactory answers). Another practical advantage to starting this way is that it has very salutary effects on class size. I normally start with around twenty students, which is far too large for the kind of open-ended group discussion that makes a good seminar. After the first session I am usually down to under ten.
Image credit: Rorden, C. & Karnath, H.-O. (2004) Using human brain lesions to infer function: a relic from a past era in the fMRI age?,Nature Reviews Neuroscience, 5, 812-819.