Friday, December 3, 2010

Did Wernicke really postulate only two language centers?

Every once and a while I look back at Wernicke's original 1874 monograph and every time I do, I learn something new. It is not much of a stretch (and might even be true) to say that our modern accounts of the functional anatomy of language are relatively minor tweaks to Wernicke's model -- despite what Friedemann Pulvermuller claims to the contrary ;-)

So today I looked again and noticed that in contrast to current belief, including my own, Wernicke did not just postulate two language centers. In fact he postulated a continuous network that comprised the "first convolution" together with the insular cortex as "a speech center". By "first convolution" Wernicke means the gyrus that encircles the Sylvian fossa, i.e., the superior temporal, supramarginal, and inferior frontal gyrus (it does make a nice continuous arc).

But this was a network organized into a functional continuum, with the superior temporal region serving sensory (auditory) functions, and the inferior frontal region serving motor functions. Now we all think that Wernicke considered these two zones to be connected via a white matter fiber bundle, the arcuate fasciculus, but this is not true (the AF was postulated later). My earlier readings of Wernicke suggested to me that he thought the connection was via a white matter tract that coursed behind the insula. But it seems that this is wrong too. Rather, Wernicke proposes that the entire first convolution zone is interconnected via the insular cortex. Here are the relevant quotes:

The existence of fibrae propriae [a term from Meynert referring, I believe, to connection fibers generally]..., between the insular cortex and the convolutions of the convexity has also been demonstrated. Since to my knowledge these have not previously been described and since they constitute a major proof of the unitary character of the entire first primitive convolution and the insular cortex, the reader will permit me to speak further of them. p. 46

He goes on for several paragraphs describing fibers that seem to connect the first convolution with the insula. At one point he even gives advice on how to see them for yourself... is best first to apply the scalpel about halfway up the inner surface of the operculum... p. 47

I suppose that is kind of like us now saying that it is best first to apply spatial smoothing with a gaussian kernel... Anyway, here he states his conclusions on the matter quite clearly:

The consideration of the anatomical circumstances just described, of the numerous supporting post-mortem studies, and finally of the variety in the clinical picture of aphasia thus leads compellingly to the following interpretation of the situation. The entire region of the first convolution, which circles around the fossa Sylvii serves in conjunction with the insular cortex as a speech center. The first frontal convolution, which is a motor area, is the center of representations of movement; the first temporal convolution, a sensory area, is the center for sound images. The fibrae propriae which come together in the insular cortex form the mediating psychic reflex arcs. p. 47

So it isn't just Broca's area, Wernicke's area, and a white matter bundle. Rather it is a continuous but functionally graded region inter-connected by a -- dare I say -- computational hub, the insula. He may not have been entirely correct about the insula as a whole, but what seems clear is that the 19th century neurologists, including the so-called "classical ones" exemplified by Wernicke, had a much more dynamic and complex view of brain systems that we give them credit for.


Wernicke C (1874/1969) The symptom complex of aphasia: A psychological study on an anatomical basis. In: Boston studies in the philosophy of science (Cohen RS, Wartofsky MW, eds), pp 34-97. Dordrecht: D. Reidel Publishing Company.

1 comment:

Long Ouyang said...

Forgive my ignorance - I do behavioral stuff so I'm not too familiar with neuroscience. How apt is this continuous network description?