EEG and ERP

Question 1

Correct or incorrect: EEG signals resemble the sum of synchronized action potentials of a group of parallel neurons.

Answer 1

All correct BUT the AP part: it picks up the post synaptical potentials that arrive at the dendrites. (before the AP)
The sum of the PSPs of the parallel neurons in the population is then compared to the reference electrode, resulting in the signal.

Question 2

Correct or incorrect: ERP signals are around 20 microVolts

Answer 2

Incorrect, ERP signals are around 2 microvolts whereas global eeg activity is around 20 microvolts.

Question 3

Correct or incorrect: time frequency analysis looks at the voltage (amplitude) over time in different frequency bands, to see if the frequency distribution in a signal changes over time.

Answer 3

Correct

Question 4

Which 3 types of dipoles are there? Where can these types of dipoles be found in the brain? What affect do they have on the measured signal at the scalp?

Answer 4

1. radial: perpendicular to surface. Gyri. Gives a focal pos. or negative peak at the surface.
2. tangential, parallel to surface. Sulci. Measures negative at sink (eg left side) and positive at source (eg. right side), with 0 in between.
3. oblique

Question 5

A negative EEG signal at the scalp can be from what type of activation?
A positive EEG signal at the scalp can be from what type of activation?

Answer 5

1. Negative extracellular charge of an EPSP at the closer part of the dipole.
   OR negative extracellular charge of an IPSP at the distal part of the dipole.
2. Positive extracellular charge of an IPSP at the closer part of the dipole. OR positive extracellular charge of an EPSP at the distal part of the dipole.

Question 6

How do the extracellular charges of a population of synchronized neurons eventually reach an electrode at the scalp? Mention the two types of conduction.

Answer 6

Positive ions in a conductive volume repel positive ions and attract negative ions which causes a wave of charges. At some point there will be a physical, insulating barrier (eg myelin, skull) that ions cannot go through. This creates a capacitor: 2 pools of charges separated by the layer (one positive, one negative pool). Here a charge difference builds up. Within the next volume, volume conduction causes a wave again until the next capacitor (brain ,dura, skull, gel, capacitor). At last, engineered capacitor, the charge is measured and turned digital.

Question 7

Why is it important to wait a few minutes after applying gel? ‘settling’

Answer 7

The ions in the gel and at the electrode are given time to reach a steady state of mixed ions.

Question 8

What is the inverse problem in EEG?

Answer 8

From voltage measured at scalp, figuring out which signal components came from which source (localization)

Question 9

In which cases can a signal from deep cortical structures still be measured by EEG?

Answer 9

In cases the sum of the PSPs is really large, so that the electrical signal is very strong.

Question 10

What is a typical range of EEG frequencies that can be studied and why is it hard to study the frequencies outside this range?

Answer 10

0.1 - 30 hz
< 0.1 hz is tough because you need a very high sampling frequency to capture it accurately
> 30 hz is heavily attuenated by the skull

Question 11

Sources of internal and external noise in EEG?

Answer 11

• external/electrical: power supply in buildings, wires in devices/lights
• internal: cadiographic (heartbeat), myogenic (muscle contraction, blinking), breathing

Question 12

Solutions for noise?

Answer 12

• shielding of the rooms or electrode cables
• active electrodes: amplifier as close to electrode as possible, so that the signal is amplified before possibility of influence by noise
• internal noise can be dealt with by postprocessing

Question 13

Which sort of processes can and cant be studied using ERP?

Answer 13

Can: - study unconscious processes
- study infants
- study during which timeframe a process is affected by a manipulation
Cant: - study slow processes
- spontaneious processes that cant be timelocked
- events that occur infrequently and cannot be averaged over many trials

Question 14

Components: exogenous

Answer 14

• < 150ms
• always happen automatically after stim
• dependent on physical features stimulus and mode (audio visual)
• eg N1 (in eg extrastriate), P1, Brainstam Auditory Evoked Responses (1 to VII), C1 (in V1) (dependent on location visual field), N170 (faces/expertise of complex stimulus)

Question 15

Components: endogenous

Answer 15

• > 150 ms after stimulus
• task dependent features influence it
• less sensitiive to physical features or modality
• eg p2, p3, n400

Question 16

p3/p300/late positive potential family: what are the p3b and p3a peaks about (300-800ms)?

Answer 16

• p3B: task defined probability: for task relevant features only, when task requires categorizing stimuli, and one category is more frequent than the other, this peak shows. Gives you info on how long it takes to categorize a stimulus. Central and parietal. Oddball paradigm.
• p3A: novelty. Even for task irrelevant features, if current stimulus is a different category than the preceding ones. Frontally.

Question 17

Response related components:

Answer 17

• readiness potential: large negativity building up over motor cortex during self paced movements
• fast responses to something become intermingled with other components, but can estimate using lateralized RP (bc only shows up in contralateral hemisphere motor areas)

Question 18

What are major drawbacks of the mathematical models used to localize the signal from ERP? What is a solution?

Answer 18

Algorithms based on constraints, may not be valid. Without assumptions, no unique solution inverse problem. When location pinpointed, no margin of error. -> may be incorrect. Conclusions formulated stronger than they should be.
Solution: use fmri or other high spatial res methods to converge results

Question 19

Schizophrenia example (Luck): RTs are usually slowed in schizophrenia. ERPs were used to test if this slowing is due to perceptual, decisional or response selection processes. They used difference waves to obtain P3 waves and to obtain LRP waves. How did they do that? Why are difference waves useful? What do these waves represent?

Answer 19

1. P3: subtract signal of frequent trials from infrequent trials in an oddball paradigm. Represents time to categorize.
2. LRP: subtract ipsilateral from contralateral over motor cortex, for frequent trials (rel to response hand). Represent time to select response.

Difference waves are a form of subtraction method to give you the wave that is solely due to the process of interest and not general processes.

Question 20

Schizophrenia example (Luck): RTs are usually slowed in schizophrenia. ERPs were used to test if this slowing is due to perceptual, decisional or response selection processes. Results and conclusion?

Answer 20

Patients showed indestinguishable P3 waves but later (60ms) and smaller LRP waves
Conclusion: response selection impaired but not perception/categorization.

Question 21

What are some standard ways of quantifiying ERP size, why are these not always good, and what are better ways?

Answer 21

• Peak amplitude. But sensitive to noise and variability in latency. Better use average amplitude. Though cannot use this if there are N and P waves.

Question 22

What are good ways to reduce type 1 error in ERP analysis?

Answer 22

• choose timerange before testing based on hypothesis rather than choosing only time with biggest difference
• choose electrodes before testing based on hypothesis rather than choosing only sites with biggest differences

Question 23

Gazzaniga visuospatial attention models: how did they distinguish between exogenous and endogenous attention in the study?

Answer 23

endogenous: instructions that a cue will appear, of which the target is very likely but not surely to be in that direction
exogenous: no instructions, just a task irrelevant flash of light that has no predictive value

Question 24

Gazzaniga visuospatial attention models: which ERP result supports the notion that endogenous visuospatial attention causes early selection rather than late selection? Which neuroanatomical fact of the visual system explains this?

Answer 24

after a valid cue, on the side of the target, the P1, N1, N2 waves (around 100ms-200ms) is larger than if invalid cue. This is likely in early visual cortex. Note: colors/objects attention effects are processed later. Visuospatial attention has early effects bc visual field is already discriminated at the lvl of the retina.

Question 25

Gazzaniga visuospatial attention models: which ERP study showed that endogenous and exogenous attention are somewhat similar, but not entirely? Hint: inhibition of return

Answer 25

In the case of a short interval between exogenous flash and target, P1, N1 early waves are larger, similar to endogenous attention results. In case of a longer interval, >300ms, due to inhibition of return, the effect is reversed. This is so we are not constantly distracted by meaningless stimuli and we can quickly return back to the space which is relevant.

Question 26

why cant APs be measured by EEG?

Answer 26

too fast, too local, abolished by tiny time differences

Question 27

voltage at electrode of EEG is a weighted sum of all dipoles. By what is it weighted?q

Answer 27

distance to electrode, conductivity of tissue, source orientation

Question 28

10 20 system of placement. what is downside of measuring eeg relative to a reference electrode?

Answer 28

fiducials are landmark points
reference electrode at m1 or m2 mastoid bone
closer electrodes to the reference have lower A

Question 29

what is fourier analysis

Answer 29

complex wave is deconstructed to simple sinus waves, from the detected contribution of simple waves a power spectrum is made with on x axis frequency and on y axis the amplitude, iti s done at 1 timepoint.
- sinusoid wave wrapped around a circle, using windinf frequency that when it matches signal frequency causes a deviation in the center of mass (integrated function). then plot center of mass x coordinate “almost fourier transform.

Question 30

what is event related (de)synchronization

Answer 30

frequency change across time in responses to a stimulus is measured with time frequency analysis

Question 31

describe in steps how the ERP signal is achieved

Answer 31

1. recording
2. pre amplify and main amplify, filtering of unnecessary frequencies.
3. sampling (analog to digital) and storing on pc.
4. timelocking signal to event
5. normalize A of epoch to 0 based on preceding signal
6. per trial type, create lvl 1 subject average
7. grand average of conditions

Question 32

advantages of ERP

Answer 32

• subject can sit (ecological)
• ## unconscious processes, infants

Question 33

example ERP: implicit racist attitudes

Answer 33

• white pps presented with white faces black faces, negative context
• P2 and N2 larger A for white faces than black faces, suggesting seen as more discrepent from negative context
• also on P3(LPP) for those high in racist attitudes

Question 34

example ERP: affective priming mechanism: if primed with a positive word, and then a negative word has to be categorized on valenced, an incongruency effect occurs where there is slowing of RT: is this a categorization effect or a response selection effect?

Answer 34

classic explanation: its a categorization effect
but: no larger or later P3 as would be expected from this
(P3 latency = categorization time)
but A of LRP shows pps first preparing a response to the prime (on that side, eg left button) -> response selection processes are affected, also confirmed by larger N2 (response conflict)

Question 35

N170 erp

Answer 35

• affected by race, gender, so those are distinguished very quickly
• extent to which target judged as human, sensitive to prejudice
• also when categorising arbitrary ingroup