Friday, April 18, 2008

Wise-guy on the Superiority of PET

And why all other methods in cognitive neuroscience (with the parenthetical exception of intracranial EEG) are rubbish. From Tuesday's talk at the 15th annual meeting of the Cognitive Neuroscience Society.

In his presentation, Professor Richard J. Wise outdid The Neurocritic on the snark front, issuing a constant stream of sarcastic bon mots to slaughter the sacred (and not-so-sacred) cows of cognitive neuropsychology and functional neuroimaging. He was the last of four speakers in a session on The anterior temporal lobes and semantic memory: Putting everything together (details and abstract below).

The style of his talk consisted of paper citations, followed by slams of the authors, their techniques, and their results. Without further ado, I give you the wit and wisdom of the Wise-guy.

(1) Catani & ffytche, 2005; Schmahmann et al., 2007.
First on the hit list is modern tractography (tracing white matter pathways) and the use of diffusion tensor imaging. "There's nothing new there" ... "it's all been done 50-100 years ago." Then it's on to straw man Norman Geshwind (prominent American behavioral neurologist) and his revival of disconnection syndromes,
condition(s) in which information transfer between parts of the brain is interrupted or blocked.
Most relevant here is the syndrome of conduction aphasia, first described by Karl Wernicke as a disconnection between Broca's area in the left inferior frontal cortex and the eponymous Wernicke's area in the left temporal-parietal region (see figure at right, adapted from Fig. 2 of Catani & ffytche, 2005). Damage to the arcuate fasciculus was thought to be the cause of conduction aphasia, but it's actually a little more complicated than that (Catani et al., 2005).

This served a launching pad for Wise to criticize all of classical and modern aphasiology, with its overemphasis on perisylvian language areas. 1

"Broca's area and Wernicke's area have expanded like Balkan states," taking over wide swaths of cortex.

The realities of the human brain's vascular anatomy cause a selection bias in the types of patients seen after a stroke -- lesions to Broca's and Wernicke's area are overrepresented -- because infarcts in the left middle cerebral artery often result in aphasia. Wise was critical of lesion studies in general, calling them nonspecific and akin to phrenology. We should "throw out most of the literature from stroke aphasia." However, it must be said that the gentleman doth protest too much, since he is the head of a group that studies, of all things, recovery from aphasia following stroke:

Brain Disorders Section: STROKE RECOVERY

No matter, onward we go. If the anterior temporal lobes (ATL) are so important for semantic memory (see also footnote #1), how about patients with pharmacologically intractable epilepsy, who have anterior temporal lobectomies (removal of the ATL) to alleviate their uncontrollable seizures? Why don't they have substantial impairments in semantic memory? The Wise answer is that these individuals are unlike those with semantic dementia, because there is reorganization of function due to chronic epilepsy. The ATL on the affected side isn't functioning normally anyway, so the story goes, so other brain regions have compensated for the faulty tissue. This argument only works for those who have had seizures since childhood, I might retort, and the literature on the topic has yielded mixed results.

OK then, how about the fact that ATL patients only have unilateral resections, whereas semantic dementia patients typically have bilateral ATL degeneration (Lambon Ralph & Patterson, 2008)? The Wise view is that "unilateral ATL damage will not cause 'semantic amnesia,' just as unilateral ATL damage will not cause 'episodic amnesia'"'2 -- which to me seems an unwise comparison, because unilateral ATL damage can, in fact, cause an impairment in episodic memory. This was just one of the no-citation, unsubstantiated factoids tossed about in his talk.

[However, I'll skip his next citation, that of Tian et al. (2001) on Functional Specialization in Rhesus Monkey Auditory Cortex in the interest of time, if anyone has made it this far...]

(2) Vigneau et al., 2006.
Professor Wise just "loves" this paper, a meta-analysis on activation peaks in language tasks. The abstract begins with the statement,
The advent of functional neuroimaging has allowed tremendous advances in our understanding of brain–language relationships, in addition to generating substantial empirical data on this subject in the form of thousands of activation peak coordinates reported in a decade of language studies.

Fig. 1 (Vigneau et al., 2006). Overview of the meta-analysis. Sagittal projection map of the 730 activation peaks. Each activation peak is color-coded according to its contrast category: phonology (blue), semantic (red), and syntax (green).

Next on the hit list is fMRI. "Thousands of activation points? This fosters confusion, not understanding. It looks as if someone fired a shotgun at the left cerebral cortex, with the barrel fired in limited areas - once again revealing a selection bias against the ATL." 3
In subsequent figures of the paper, the authors are guilty of "converting shotgun pellets to cannonballs."

I must admit, the man has a very good point here...

(3) Hickok & Poeppel, 2007.
And Dr. Wise-guy reserved his most sarcastic comment for the authors of this paper. "Every two years they produce a new model. They have a job for life." However, I would think that the willingness to revise one's model in the face of new and conflicting data is an essential aspect of the scientific method, no? Pose a testable (and falsifiable) hypothesis, collect data, confirm or disprove your hypothesis. Isn't that the way it works? We should congratulate Hickok and Poeppel for advancing theory and knowledge of the functional neuroanatomy of speech processing. Reaching the thin air of the scientific stratosphere does not grant one the immunity to ignore data that don't fit a long-standing rigid theory...

Wise did, however, issue a backhanded compliment to the duo by conceding that they "read and think widely and deeply." Hickok and Poeppel are also the authors of the Talking Brains blog.

In Part 2, The Neurocritic takes on the claim that PET (postitron emission tomography) provides a "more naturalistic" experimental environment than MRI, among other justifications for using an outdated method. Stay tuned!


1 This was, after all, a symposium on the importance of the anterior temporal lobes (not perisylvian cortex) in semantic memory (e.g., memory for word meanings and conceptual knowledge),
factual information and general knowledge about the world, generally thought to be independent of context and personal relevance.
2 Here, Wise is referring to an impairment in episodic memory, the typical meaning for the terms 'amnesia' or 'anterograde amnesia'.

3 This is because susceptibility artifact, or the signal dropout/distortion that occurs with fMRI, limits the images that can be obtained from this region. More on this point later, in Part 2.


Catani M, ffytche DH. (2005). The rises and falls of disconnection syndromes. Brain 128:2224-39.

Catani M, Jones DK, ffytche DH. (2005). Perisylvian language networks of the human brain. Ann Neurol 57:8-16.

Hickok G, Poeppel D. (2007). The cortical organization of speech processing. Nat Rev Neurosci. 8:393-402

Lambon Ralph MA, Patterson K. (2008). Generalization and Differentiation in Semantic Memory: Insights from Semantic Dementia. Ann NY Acad Sci. 1124:61-76.

Schmahmann JD, Pandya DN, Wang R, Dai G, D'Arceuil HE, de Crespigny AJ, Wedeen VJ. (2007). Association fibre pathways of the brain: parallel observations from diffusion spectrum imaging and autoradiography. Brain 130:630-53.

Vigneau M, Beaucousin V, Hervé PY, Duffau H, Crivello F, Houdé O, Mazoyer B, Tzourio-Mazoyer N. (2006). Meta-analyzing left hemisphere language areas: phonology, semantics, and sentence processing. Neuroimage 30:1414-32.

Symposium Session 7, Grand Ballroom B
Tuesday, April 15, 10:00 am - 12:00 pm

The anterior temporal lobes and semantic memory: Putting everything together

Chair: Timothy T. Rogers
Speakers: Matthew Lambon Ralph, Eric Halgren, Timothy T. Rogers, Richard Wise


Abstract: Functional imaging studies have emphasized the roles of posterior temporal and inferior frontal cortex in speech comprehension and production. These studies have predominantly used magnetic resonance imaging (fMRI), a technique that returns a relatively low signal-to-noise ratio in anterior temporal cortex. Positron emission tomography (PET) does not have this disadvantage. Across a number of studies, using a range of different study designs that involved normal speech comprehension, reading comprehension, propositional speech production and narrative writing, PET has demonstrated strong converging activity in anterior temporal cortex. Language comprehension, irrespective of modality, results in symmetrical activity between the hemispheres, whereas language production is associated with greater activity on the left. The comprehension of speech is associated with strong functional connectivity between the two anterior temporal lobes, a connectivity that is not present when subjects listen to acoustically matched but unintelligible spectrally-rotated speech. More recent fMRI studies are now also demonstrating language-related signal in anterior temporal cortex. Interpretation of these functional imaging studies is made in the light of what is known of impaired and preserved language and memory functions in patients with semantic dementia. Together, the data provides very powerful evidence that anterior temporal cortex is implicated in semantic memory.

Subscribe to Post Comments [Atom]


Post a Comment

<< Home

eXTReMe Tracker