lesion studies

Question 1

What is meant by patient studies?

Answer 1

The process of understanding the neural bases of cognition by studying “atypical” individuals

• Lesions (i.e. brain damage)
• Abnormal brain input/output (e.g. sensory or motor loss)

Question 2

What is the goal of lesion studies

Answer 2

To understand the brain’s normal function

Question 3

What things can be regarded as a patient study?

Answer 3

• Trauma
• Disease
• Surgery

Question 4

What do we look at to measure brain function in lesion studies?

Answer 4

• Cognition/behaviour
• measured in experimental tasks

Question 5

How can we know the brain is/was atypical

Answer 5

• brain scan
• or post-mortem

Question 6

Describe the famous Broca study (1861)

Answer 6

Broca (1861): “Tan”

• Only speak one syllable
• Post-mortem analysis confirmed location of neurosyphilitic lesion
• in inferior frontal gyrus of left hemisphere

What was initially called “Broca’s aphasia” to describe these symptoms has since been redefined - why?

• Follow-up CT scan over 100 years later revealed damage to other areas in left hemisphere (Signoret et al, 1984)

Question 7

Describe famous study: Patient HM

Answer 7

Scoville & Milner (1957)

• Bilateral medial temporal lobe resection (removal) Including hippocampus
• for relief of epilepsy

Symptoms

• Extensive anterograde amnesia
• Intact short-term and procedural memory
• Later neuroimaging work showed further damage (Corkin et al, 1997)

Question 8

HM - what areas were differently found damaged from the study in 1957 to 1997?

Answer 8

Scoville and Milner 1957; Corkin et al., 1997

Question 9

Describe famous study: Patient SM

Answer 9

Adolphs et al (1994)

• Bilateral lesion to the amygdala through rare genetic condition (Urbach-Wiethe disease)
• Selective impairment in experiencing fear
• Instigated the modern discipline of affective neuroscience

Question 10

Describe famous study: Patient DF

Answer 10

Goodale & Milner (1992)

• Bilateral lateral occipital cortex lesion due to carbon monoxide poisoning
• Visual form agnosia
• Intact visuomotor interaction with object form
• “Two Visual Streams”

Question 11

What are the caveats of patient studies?

Answer 11

•Assumption of modularity

• Assumption of modularity
• Distributed networks
• Secondary effects of lesion on other areas
• “Snowflakes”

Question 12

What is the assumption of modularity

Answer 12

The assumption that individual cognitive processes can be mapped onto individual brain areas (Fodor, 1983)

Unlikely for many brain areas due to plastic effects there e.g.,

• prefrontal cortex neurons adapt based on task demands (Miller, 2000)
• sensory areas respond to different stimuli with training and with sensory loss

even if modularity can be assumed, this is a problem in lesion studies because the functional modules are. notrespsected by the usual causes of brain damage e.g., stroke

Question 13

Are lesion studies the only methods that assume modularity?

Answer 13

No, fMRI method does too

Question 14

Distributed networks

Answer 14

• Functions can be implemented in distributed networks
• The lesion may disrupt function by disrupting network, not because function is localized to this region

If damage to a region results in loss of function doesn’t mean this area was necessary for that.

Question 15

Secondary effects

Answer 15

• Acute localised brain injury can result in profound cerebral abnormality in function, which subsides over time
• Testing patients too soon after acquired injury can lead to incorrect assessment

Over time, other non-damaged brain areas adapt or degenerate – longterm neuroplastic changes

• Testing patients too long after acquired injury can lead to incorrect inferences about “normal” brain function•

Question 16

Snowflakes

Answer 16

No two brains are the same

• There is an enormous number of ways in which it can be damaged, and complexity increases if a cognitive process is carried out by several areas
• Might result in subtly different symptoms
• Knowledge of the damage, as well as the symptoms, is, therefore, important

Buxbaum (2006) on neglect:

“Nearly every possible fractionation of the disorder has been reported, raising the possibility that each patient may be as unique as a snowflake**”

Question 17

How to effectively use patient studies?

Answer 17

• Combine multiple patients
• Double dissociation studies
• Appropriate cognitive tests
• Combination with other methods

Question 18

how to effectively use patient studies: combine multiple patients

Answer 18

Map the area(s) damaged in patients with a common impairment

• e.g. Hayward, Naeser & Zatz, 1977
• commonly damaged areas in 3 patients with Broca’s aphasia

Consider control patients

• Those who are matched as much as possible
• Not just in terms of sex, age, intelligence, relevant socio-economical factors etc.
• In terms of injury/etiology

Question 19

how to effectively use patient studies: Combine multiple patients – Voxel Based Analysis of Lesions (VAL)

Answer 19

Requires two groups of patients,

• both of which have lesions,
• but some of which have deficit and some of which do not have deficit

Map damaged area(s) in patients with and without the impairment. Statistical comparison (χ^2) between both groups. What’s gwarning in them that not in those?

example study - Rorden & Karnath, 2004

Question 20

How to effectively use patient studies: Combine multiple patients – Voxel Based Lesion Symptom Mapping (VLSM)

Answer 20

• Requires sample of patients with lesions and range of performance on a behavioral measure.
• For each voxel we then compare performance between those with and those without lesion
• reveal association between lesion and performance on task

Question 21

How many n do you need when applying multiple patient approach

Answer 21

You will need large number of patients & controls!

Question 22

How to effectively use patient studies: Double Dissociation studies

Answer 22

2 patients with different lesions - produce opposite effects in each.

• very strong evidence that two processes are functionally independent
• Address the potential issue that one process might simply be a subcomponent of the other, or, “simpler” than the other

Question 23

Give me two examples of double dissociation studies

Answer 23

E.g. visual form agnosia and optic ataxia (Goodale & Milner, 2004)

• Patient DF: Impaired visual shape recognition–Intact visuomotor shape interaction
• Patient RV: Intact visual shape recognition impaired visuomotor shape interaction

E.g. Broca’s and Wernicke’s aphasias (1800s)

• Broca’s–Non-fluent aphasia: Patient understands language but can not speak
• Wernicke’s–Fluent aphasia: Patient produces (jumbled) speech but with no comprehension

Question 24

How to effectively use patient studies: Appropriate cognitive tests

Answer 24

Accurate classification of the patient’s impairment is entirely determined by the test that you use

• Subtly different behavioural/cognitive tests can reveal very different types of impairment
• Example: Apperceptive vs. associative agnosia

Question 25

Apperceptive vs. Associative agnosia

Answer 25

• Visual agnosia is an inability to recognize or appreciate the identity of objects, faces or letters
• Usually attributed to lesions to high-order modules of the visual system, which combine visual cues to represent the shape of objects, e.g. lateral occipital cortex
• Those with apperceptive agnosia can not copy.
• Those with associative agnosia can copy.

Question 26

What methods can we combine lesion studies with

Answer 26

• Neuroimaging
• Virtual Lesions
• Animal models

Question 27

Combining lesion studies with other methods: neuroimaging

Answer 27

• These days standard to combine patient work with neuroimaging to get structural images of brain (and lesions) while patient is alive (historically one had to wait for post mortem, and even this could lack accuracy)
• Findings from patient studies can be tested/developed by functional neuroimaging experiments

Question 28

Combining lesion studies with other methods: virtual lesions (e.g., TMS)

Answer 28

Pro:

• Experimental manipulation

Con:

• More surface based
• Large spread

Question 29

Appropriate experimental approaches/design and statistics

Answer 29

There are three main approaches to assessing the performance of a patient:

• Patient is administered fully standardised neuropsychological tests and compared to large sample normative data
• Patient is not compared to any controls. Intra-individual approach
• Patient is compared to a modestly sized control sample

Question 30

Appropriate experimental design and statistics: Standardised neuropsychological tests

Answer 30

E.g.

• Wechsler Adult Intelligence Scale
• Wechsler Memory Scale
• Comprehensive Aphasia Test

Can only be used in limited situations

• New constructs are constantly emerging in neuropsychology
• Collection of norms is a long and arduous process
• Normative data might be out of date or not applicable to patient

Question 31

Appropriate experimental design and statistics: Intra-individual approach

Answer 31

Inferences made based on patient’s performance on task A compared to task B. No reference to a normal population

E.g.

• Patient LN has a brain lesion.
• He is able to name living things but not non-living things
• The brain area is implicated in naming non-living things

Question 32

Appropriate experimental design and statistics: Intra-individual approach

how is this done statistically?

Answer 32

Usually tests of frequency distribution, based on categorical classification of performance on trials

• Binomial test
• Chi-squared test
• Chi-squared is a test of whether two variables are independent–Frequency data (i.e. “count” data)
• The result of this chi-squared test is significant (p=0.027). This patient’s performance in tasks A and B is dependent upon the condition (X or Y)

Question 33

Appropriate experimental design and statistics: Intra-individual approach

whats the problem with not making comparisons to normal population

Answer 33

Sometimes the things they are impaired at - arent different from controls

Laws, Gale, Leeson & Crawford (2005)

• Categorical dissociations in Alzheimer’s disease
• Some patients with “dissociations” were not abnormal when compared to controls

Question 34

Appropriate experimental design and statistics: Intra-individual approach

Answer 34

The best experimental design with a patient is…“Does patient X have an impairment in Y compared to controls?”

• This requires comparing the performance of one subject to a group of n subjects.

How is this statistically possible?

• patient scores 5 on a memory test
• compairson n score: 10, 7, 6, 8, 4, 5, 6, 11, 10, 8, 9, 7

Option 1: WRONG

• Treat the ‘case’ as a fixed value to which the control group is compared (e.g. one sample test: One sample t-test: t(11)=4.159; p=.002)
• This is wrong because the measurement of both the control group and the case by itself will have some error (i.e. variance)!

Option 2: Correct

• Consider the case a randomly drawn sample from a distribution of cases, to which the control group is compared.
• Use a modified t-test (Crawford & Howell, 1998) to test the hypothesis that patient did not come from the control population (Singlims_ES: t(11)=1.274; p=.229)