EEG: principles ERPs & time-domain

Question 1

Who invented the Electroencephalogram (EEG) and what “Brain rythm” did that person observe?

Answer 1

Hans Berger invented the EEG in 1929 and observed the Alpha Rythm. As he connected the Rythm to some mental activity that ceased to exist, some may regard this as an early attempt of “Mind Reading”

Question 2

What are the three most basic steps for acquiring an EEG-recording?

Answer 2

1. Place an Electrode on the scalp and another reference electrode somewhere else on the body.
2. Amplify the signal to record voltage differences between these electrodes
3. Measure how voltage changes over time

Of course modern EEG-Setups are more complex, these are just the basic steps!

Question 3

Where are EEG-Signals measured from and what can they be used for ?

Answer 3

Usually EEG-Signals are a measurement of electrical activity from scalp-electrodes, that are mounted on a cap. Signals in the μV range are amplified and used in a clinical or research context.

Question 4

What type of neurons are especially relevant to EEG-recordings and how are these cells arranged?

Answer 4

Pyramidial neurons of the cell layers 5 and 6 are arranged in parallel and perpendicular to the surface of the cortex.

In this configuration, the cell bodies and dendrites of these neurons point “up” towards the surface of the cortex, the axons point “down”

Question 5

What is the cellular neural basis of an EEG-recording?

Answer 5

Dendrites of (pyramidial) neurons recieve input from other neurons. This induces a postsynaptic potential (PSP) in the apical dendrites.

This again creates negative charges in the dendrite region (net negativity), wich in turn leads to positive charges around the soma (net positivity) ultimately resulting in a dipol.

The EEG captures the modulation of extracellular currents due to this PSP and the subsequent dipol.

Question 6

Concluding: What is the “General assumption” of what the EEG measures?

Answer 6

The EEG measures the summed up activity of thousands of parallel oriented, neighboring and synchronously active pyramidal cells.

Question 7

Name three factors the EEG-signal depends on.

Answer 7

1. Orientation of dipole: activity in gyrus vs. sulcus
2. Location of synaptic activity: excitatory inputs near soma vs. near the dendrite -> positive vs. negative deflection
3. Neural synchrony: amplitude of the EEG is modulated by the amount of synchrony in a local population of neurons (cf. epileptic seizure — high amplitudes due to the extreme synchrony)

Question 8

Explain how the EEG-signal is dependent of the orientation of the dipole.

Answer 8

The Cortex is folded in Gyri and Sulci. As the neurons are perpindicularly oriented to the surface of the cortex, they have varying orientations to oneanothers.

This affects the “readability” of these dipols in the EEG (as some might cancel aout and as the EEG is especially sensitive to perpendicularly oriented dipols evoked by pyramidal neurons in the gyri.

Dipols created in the deeper parts of the cortex may have less impact on the EEG-reading than superficial dipols.

Question 9

How does the location of synaptic activity affect the EEG signal?

Answer 9

There are generally two cases:

1. If inputs arive near the soma (e.g. from thalamus neurons), the dipole is reversed. Positive charges are created near the dendrites and negative charges are created near the soma leading to a positive deflection of the EEG reading.
2. A input from other neurons of the cortex arives at the dendrites, creating negative charges at the dendrites and positive charges at the soma. If this is the case a negative deflection of the EEG is read (due to the orientation of the dipole)

Question 10

Why does neural synchrony affect the EEG-signal recorded?

Answer 10

If the activity of the single neighbouring neurons is irregular and uncorrelated, then the sum will be a rather flat line, than variant. (Cancelling each other out akin to Destructive Interference in physics)

Highly synchronized and correlated activity would lead to larger deflections of the EEG.

-> The Amplitude of the EEG is modulated by the amount of synchrony in a local population of neurons

Question 11

According to the idea, that the EEG-signal is dependent on synchrony of local populations of neurons, what should an EEG-recording of a epileptic seizure look like?

Answer 11

Large, slow and synchrounous waves in many adjacent elctrodes, as in epileptic seizures synchronoy firing of large neural populations occur.

Question 12

Name 4 Pros of EEG

Answer 12

1. High temporal resolution
2. Non-invasive
3. Relatively affordable and small
4. Light weith and even portable with modern systems

Question 13

Name 2 Cons of EEG

Answer 13

1. Low spatial resolution
2. Lenghty setup

Question 14

According to what system are electrodes placed when using an EEG? How does the system work and what are its advantages?

Answer 14

• The 10-20 System is used where electrodes are placed in relation to the skull anatomy, more precisely the distance between nasion and inion.
• This allows for relative and constant placement of electrodes ensuring comparability between measurements of a participant and between participants and labs.
• The nomenclature is in accordance to the lobes with Frontal, Parietal, Temporal, Occipital lobes and electrodes at the Central sulcus. Further electrodes on the midline are marked with a z, electrodes in the left hemisphere with uneven and those in the right hemisphere with even numbers. (e.g. PO3 being a left-hemispheric, parietal-occipital electrode)

Question 15

What two domains are raw EEG signals comprised of?

Answer 15

Not brain signals:
* Eye movement artifacts
* Muscle artifacts
* Heartbeat
* Slowdrift (e.g. sweating)

Brain signals
* e.g. Alpha rythms of the visual cortex

Question 16

Name three relevant steps to preprocessing EEG-data

Answer 16

1. Substraction of systemic artefacts
2. Filtering data (high pass filter removing e.g. everything above 0.1 Hz for removing slow drift and low pass filter removing data under 40 Hz for remocing EMG activity)
3. Rejection of exsessively noisy data

Question 17

Name two different types of “signals of interest” in EEG.

Answer 17

1. Spontaneous oscillations (i.e. brain rythms that occur in frequeny bands)
2. Event-related potentails (i.e. time-locked EEG responses in short duration reflecting internal and extrenal factors)

Question 18

What is an ERP?

Answer 18

A Event Related Potential is the electribal brain activity, as measured by EEG time-locked to some event.

ERPs can be evoked by external and be generated by interaly stimuli/events

Question 19

How is an ERP extracted from continous data?

Answer 19

The continous data is segmented into epochs around the marker, marking the related event. Such epochs usually span a few hundred ms before the timepoint of the event (t=0) until a few seconds after the event.

By averaging Epochs of many trials, the ERP becomes apparent.

Question 20

Name the fundamental challenge to EEG-analysis and a common technique taht is developed to investigate signals of interest accordingly.

Answer 20

In EEG, the SNR is rather bad. Investigators want to investigate only specific components of the signal related to the event, hence the noise must be reduced.

Typically, for ERPs, the common technique to investigate the signal of interest is to record a long serirs of trials and then to average accros trials

Question 21

Based on a simple model assumption explain the most commonly used techniwe to investigate signals of interest in ERPs

Answer 21

The model assumption is, that an ERP p(t) is linearly superimposed by background brain activity and noise r(t), such that the acquired EEG signal x can be conceptualized as:

x(t) = p(t) + r(t)

If p(t) is constant over each trial K and r is gaussian distributed and independent between trials, the averaging leads to 1/K sum(x(t)) = p(t) + 1/K sum (r(t))

The average noise is determined by the mean of the gaussian distribution, being 0, noise will ultimately progress towards that mean in its average.

As the combined Standard Deviation is expressed as σ / √K, the npise goes down by a factor of √K across an average of K trials.

Question 22

Explain the Nomenclature of ERPs

Answer 22

1. Nomenclature by Polarity (N/P) and a number indicating either latency (e.g. 170) or ordinal position (e.g. 3) Example: N170 or P3
2. Less common nomenclature by functionality (e.g. ERN for Error related negativity)

Question 23

What is the P1 and what is the P300?

Answer 23

The P1 is a positive deflection of the EEG after 100 ms after the presentation of a visual stimulus. The P1 source could be reconstructed as the coming from a dipole in the right visual cortex. It reflects perceptual processing and is modulated by attention

The P300 is a positive deflection at abour 300 ms after stimulus onset and can be subdivided into P3a and P3b. It reflects the occurence of a targets stimulus in an oddball paradigm.

Question 24

ERP-Components can be split into early and late components what are their respective characteristics?

Answer 24

Early
* mainly automatic sensory responses
* highly influenced by stimulus properties
* used clincally
* can be modulated by emotion, mood attention reward etc.

Late
* reflect internal, higher order processing
* not so dependent on physical stimulus properties
* influenced by task, strategy, emotional processing etc.

Question 25

What law does the EEG-power spectrum follow. What are “Brain rythms” in this respect?

Answer 25

The 1/f law, indicating that when the spectral power goes up, the frequency goes down and vice versa.

Bumps with slightly larger power than expected by the 1/f law at certain frequency bands implicate the existence of distinguishable “Brain Rythms”

Question 26

What are the five “Brain Rythms”, when do they occur an what frequency do they possess?

Answer 26

1. Delta (0.5-4Hz): Deep sleep in adults and infants, abnormalities in the waking adults
2. Theta (4-8 Hz): More pronounced in children, drowsiness or light sleep
3. Alpha (8-13 Hz): Awake with eyes closed, found mostly occipitally; mental inactivity, physical relaxation
4. Beta (13-30 Hz): Sharp spike waves, with many different evoking factors
5. Gamma (31-100Hz): high-level processing and the “binding of consciousness”, unity of perception