1. Language Part 2
Saygin, AP

2. We already mentioned that...
Patients like “Tan” do not reliably have Broca’s area lesions.
Agrammatic patients do not reliably have Broca’s area lesions.
Modern neuroimaging techniques allow us to noninvasively and accurately identify the lesions of patients.
Commonly used structural imaging methods are CT (Computerized Axial Tomography) and MRI (Magnetic Resonance Imaging) – even more advanced technologies are now available.
We discussed VLSM.
How do the classical theories hold up based on newer neuroimaging data?

4. The “new” localization
MRI or CT scans of large sets of patients transferred onto standardized templates and perform lesion-symptom mapping analyses on this common space.
Lesion overlap analyses
Voxel-Based Lesion Symptom Mapping (VLSM)

5. Lesion Overlap: Broca’s Aphasia

6. Lesion Overlap: Wernicke’s Aphasia

7. Lesion Overlap: Conduction Aphasia

8. Lesion Overlap: Global Aphasia

9. Lesion Overlap: Anomic Aphasia

10. VLSM: Quick Review No need to know lesion regions of interest
No need to categorise patients
Use the ‘voxel-based statistics’ idea
And facilitate link between lesion mapping and neuroimaging

11. VLSM: The Procedure A group of patients and a behaviour of interest
Patients’ lesions on a common space (e.g., MNI “Colin”)
Statistics over voxels using behaviour(s) of interest
Plot statistics over space
+ Behaviour

12. Speech Fluency and Auditory Comprehension
Western Aphasia Battery
Fluency: Insula, parietal white matter; also IFG
Auditory comprehension:
MTG, also STG
Bates et al., 2003, Nature Neuroscience

13. Stimulation
Positive effects – elicitation of vocalizations (not actual words; e.g., “Oh.”)
Negative effects
Total arrest of speech, or an inability to vocalize spontaneously (shaded zones)
Inability to name despite retained ability to speak (Broca’s and Wernicke’s areas)
Misnaming – using words related in sound (e.g., “camel” for “comb”), using synonyms (e.g., “cutters” for “scissors”)(Broca’s and Wernicke’s area)

14. Effects of stimulation of the anterior and posterior regions are remarkably similar
Stimulation effects beyond Broca’s and Wernicke’s areas
Stimulation affects more than just talking – deficits in voluntary motor control of facial musculature as well as short-term memory and reading
Broca’s and Wernicke’s areas do more than just language (more on this soon)

15. Simple speech production
FMRI
Produce /pa/ and /gi/
Listen /pa/ and /gi/
Simple speech production and listening does not involve Broca’s area but more dorsal motor, premotor cortex, supplementary motor area, insula (and temporal cortex because you hear yourself).
Listening and producing have overlapping representation in motor/premotor areas.
Broca’s area - when there is a more complex stimulus or task
Wilson, Saygin, Sereno, Iacoboni, 2004, Nature Neuroscience

16. Rethinking the classical theory
2) Are lesion-symptom mappings as predicted?
While certain tendencies are observable, the classical lesion predictions are not as predictive as once thought.
Recent localization attempts with finer measures (like “auditory comprehension” rather than “Wernicke’s aphasia”) have been more successful.

17. Nonlinguistic impairments in aphasia
Aphasic patients have language impairments. What else is impaired in aphasia?
Nothing else: Language is selectively impaired.
Lots of other stuff: Patients are “lame in thinking” (Jackson, 1878), “asymbolia” (Finkelnburg, 1870).
Apraxia and similar disorders in action production and comprehension are also seen in aphasic patients.
Impairments in non-linguistic domains have been demonstrated: Deficits in picture-object matching (De Renzi et al., 1968b), gesture-object matching (De Renzi et al a), sound-picture matching (Schnider et al., 1994; Spinnler & Vignolo, 1966; Varney, 1980)
Are these impairments systematically related to the aphasia?
What are their behavioral and neural correlates? 

18. Example: Speech and environmental sounds
Do aphasic patients have a selective impairment in speech comprehension or do they show concomitant deficits in non-verbal auditory processing?
What are lesion locations that lead to deficits in verbal and/or nonverbal sound processing?
Saygin, Dick, Wilson, Dronkers, & Bates, 2003, Brain

19. Results
Aphasic patients are also impaired in understanding nonlinguistic sounds.
In fact, speech comprehension and sound comprehension impairments went hand in hand.
Thus, aphasic patients’ deficits are not strictly restricted to the language domain.
Saygin, Dick, Wilson, Dronkers, & Bates, 2003, Brain

20. VLSM Results
Lesion sites were similar for the two domains – Patients with posterior lesions performed badly with both types of sounds.
In fact Wernicke’s area (superior temporal) lesions were statistically more predictive of environmental sound deficits!
Saygin, Dick, Wilson, Dronkers, & Bates, 2003, Brain

21. Interesting case: Nonverbal Auditory Agnosia
Patient M – an anomic aphasic with a large lesion covering temporal and parietal lobe (including Wernicke’s area) performed 100% on speech sounds and badly on environmental sounds. 
Saygin, Leech & Dick, 2010, Neuropsychologia

22. Interesting case: Nonverbal Auditory Agnosia
Patient M – an anomic aphasic with a large lesion covering temporal and parietal lobe (including Wernicke’s area) performed 100% on speech sounds and badly on environmental sounds. 
Also an example of how to identify significant dissociations!
Saygin, Leech & Dick, 2010, Neuropsychologia

23. Interesting case: Nonverbal Auditory Agnosia
Patient M – an anomic aphasic with a large lesion covering temporal and parietal lobe (including Wernicke’s area) performed 100% on speech sounds and badly on environmental sounds.
Further testing revealed that he had a persistent deficit for identifying environmental sounds
No answer on 50% of test items.
Could not identify some of the easiest sounds like doorbell or car horn
Very good conversational language and auditory comprehension.
Wernicke’s area lesions do not always cause Wernicke’s aphasia – in this patient, there was no auditory speech comprehension problem but severe impairments in environmental sound comprehension.
So remember, there are always exceptions! 
Saygin, Leech & Dick, 2010, Neuropsychologia

24. Patient M: fMRI Evidence for Reorganization
Persistent and severe deficits restricted to nonverbal auditory comprehension can follow a unilateral lesion to the classical Wernicke’s area.
This selective deficit observed for nonverbal material appears to be a result of functional reorganization (possibly in response to extensive speech training).
Saygin, Leech & Dick, 2010, Neuropsychologia

25. Pantomime Comprehension
We talked about this in aphasia videos – there are more systematic studies of action, gesture and pantomime processing deficits in aphasia

26. Pick the matching object

27. Non-linguistic (pantomime) condition

28. Linguistic (reading) condition
He is licking the…

29. Pantomime Comprehension (VLSM Results)
Lesion results were very interesting! For pantomime comprehension, we found Broca’s area (not for linguistic equivalents tested with reading comprehension)
Large region in the IFG, premotor and motor cortex (IFG/PMC) - includes Broca’s area
Small region in the primary sensory cortex (PSC)
Head of caudate (CAU)
Saygin, Wilson, Dronkers & Bates, 2004, Neuropsychologia

30. Biological Motion Perception
Map of the t-statistic at each voxel
Large region in temporal and parietal (BA 21, 22, 37, 39, 40)
Smaller area in inferior frontal (BA 44, 45, 6)
Saygin, 2007, Brain

31. Rethinking the classical theory
Is aphasia specific to the linguistic domain?
NO. It has long been observed that patients present with a number of nonlinguistic problems such as deficits in pantomime and environmental sound processing.
It appears that at least in some cases the deficits are correlated with the aphasia.
These nonverbal deficits are sometimes associated with lesions in “classical” language areas.
Wernicke’s area lesions can cause nonverbal sound processing deficits (Saygin et al., 2003). Similarly Broca’s area lesions were associated with pantomime and biological motion perception deficits (Saygin et al., 2004).

32. “Language areas kept their day jobs”

33. Specific deficits in grammar
Specific deficits in grammar? Broca’s Aphasia and Agrammatism Claim: Broca’s aphasia is a disorder of grammar

34. The Task Two pictures of animals appear
The animals are named (e.g, “cat”, “dog”)
A sentence is played (e.g, “the cat chased the dog”)
The subject hits the button under the animal doing the action in the sentence

35. Sentence Types Active: “The dog is biting the cat” Subject Cleft:
“It’s the dog that is biting the cat”
Object Cleft:
“It’s the cat that the dog is biting”
Passive:
“The cat is bitten by the dog”

36. Deficits are not specific to lesion site or aphasia subtype

37. College students under “stress”
Examples of sensorimotor ‘stressors’ or ‘disruptors’
Spectral degradation (low pass filter, noise mask)
Temporal degradation (time compression)
Attentional stress (memory load, dual-task)
Dual-stress conditions

38. Students under stress
Low-pass filter plus speech compression shows dramatic deficit
Dick, Bates, Wulfeck, Utman, Dronkers, & Gernsbacher (2001), Psych. Review

39. Aphasic patients and dual-stress students
Students under low pass filter plus compression pattern show selective deficits qualitatively and quantitatively similar to those of aphasic patients
Dick, Bates, Wulfeck, Utman, Dronkers, & Gernsbacher (2001), Psych. Review

40. Note: The argument is not that the aphasic patient have a problem with ‘hearing’ because the stimuli are acoustically distorted for the normal controls.
View these as processing stressors

41. More on Grammar
Trace Deletion Hypothesis: Agrammatic aphasics are selectively impaired in their ability to process structures involving traces of maximal projections (Grodzinsky & Finkel, 1998).

42. More on Grammar
Trace Deletion Hypothesis: Agrammatic aphasics are selectively impaired in their ability to process structures involving traces of maximal projections (Grodzinsky & Finkel, 1998).
Grammaticality judgment task.
Difficulty confound! The “trace” sentences were much more subtle (i.e. harder) than the “other” sentences!
We used the same sentences, plus generated new ones that were “easy” or “hard” as well as “trace” and “other”.
Confirmed easy/hard confound and tested our new stimuli with studies on controls.
When did John do what?
*What did John do when?
Could they have left town? *Have they could left town?

43. More on Grammar
Wilson & Saygin, 2004, Journal of Cognitive Neuroscience

44. No Evidence for Trace Effects
Aphasia Subtype, Agrammatism, Broca’s Area Lesions
Wilson & Saygin, 2004, Journal of Cognitive Neuroscience

45. Behaviorally: Comprehension
Wilson & Saygin, 2004, Journal of Cognitive Neuroscience

46. VLSM: Posterior (Wernicke’s Area!)
Wilson & Saygin, 2004, Journal of Cognitive Neuroscience

47. Summary
In science, it is important to know about the classical theories but also to not be constrained by them – this is especially true in neurolinguistics.
Early pioneers identified syndromes and lesion-syndrome mappings that are still with us today.
Later theorists have stipulated more abstract and modular models of language in the brain and very specific linguistic roles for these areas.

48. Summary
However…
Lesion-Symptom mappings are not as clear-cut as predicted
Aphasic patients have non-linguistic deficits as well – sometimes these are correlated with linguistic deficits
Language areas kept their “day jobs”
There don’t seem to be special grammatical knowledge selectively impaired in aphasia: Syntactic structures that are hard for aphasics are hard for you too!