Lecture 3: EEG part 2 Flashcards
participant generated noise
- eye blinks
- electrical interference
- muscle activity
- head movement
Muscle activity
- Generates electrical currents
- Can be removed or eliminated by asking participants to relax.
Eye movements and blinks
- The eye has a strong electromagnetic field that is established by neurons in the retina.
- Should record eye movements with using eye-trackers or EEG electrodes near the eyes.
Bell’s phenomenon
- When the eye blinks, the eyeballs move upwards.
- The retina is negatively charged, and the corneas is positively charged.
- We can see these deflections on the EEG data
Electrical interference
- Unavoidable
- Reducing by correct wiring and shielding rooms.
- Interference from TVs, phones, pacemakers
- However, EEG has high discrimination and input impedance - usually rejects extreme interference
Electrodes and equipment
- Faulty equipment can be an artefact
- Movement of electrodes can cause severe artefacts
- Make sure electrodes are securely attached.
Impedance
Impedance is a term that refers to the electrical resistance of the scalp, measured in kΩ. The higher the electrical impedance of the scalp, the greater will be the impact on the recording
how to have low impedence
- Make sure participants have their hair washed and dried
- No pins or hair clips
Types of analysis
- Frequency-based analysis: not about precise timing of stimulus related activity, but around general mental states
- ERPs: study brain processes in relation to events or stimuli
Pre-processing
Transforming raw data into data suitable for interpretation and analysis.
What does pre-processing do?
- removes noise e.g. blinks, movement
Steps in pre-processing
- filter data - digital filters are used to reduce noise.
- removing bad channels - automatically detects bad electrodes
- re-referencing: reference electrodes placed in neutral place, average of two mastoids/earlobes or average of all electrodes often used.
Frequency based analysis
Examines how frequencies vary in the brain, depending on changes in internal states or environment (5 different wave bands)
What can frequency-based analysis do?
- reveal mental, affective or cognitive state of participants
- reveal abnormal brain activities such as epilepsy or sleep disorders.
How to start frequency-based analysis
- record EEG data for 2 minutes with eyes open
- record 2 mins with eyes closed
Eyes closed and open: FBA
- Closed: higher frequency alpha band - usually see alpha when eyes are closed.
- Alpha power reduced when eyes open - this is called Alpha blocking
- Alpha power increases when eyes close again.
differential amplifier
amplifier takes two inputs and displays the output as the difference between two outputs
Alpha rythm
- Upward deflections: negative polarities, represent the eyeballs moving down and eyes opening
- Downward deflections: positive polarities, eyeballs moving upwards as the eyes close (Bells)
bipolar montage
- with FBA, a further difference is taken than the differential amplifier
- further difference is taken by two active eletrodes
- this will then create a tracing, which then becomes a chain
Characteristics of Alpha example.
- Alpha rhythm is blocked with the eye opening, and re-emerges after eye closure
- this is known as reactivity
Frequency of alpha example.
- should normally be between 8-12
- can assess by looking at a second of artefact free recoding when eyes are closed
- 11 cycles per second, this is considered normal
Amplitude of alpha example.
- switch to referential montage instead of bipolar montage
- insert scale legend to estimate alpha amplitude
- measure by looking at total deflection from lowest to highest point
- this is 1/3, meaning it’s between 40-50 mvs
- this is normal
Symmetry of alpha rhythm example.
- shift scale legend to other hemisphere
- common to have higher voltage on right than left
- asymmetry fine, as long as less than 50%
Distribution of alpha rhythm.
- normally in the posterior head regions; the particularly occipital electrodes
Goal of ERPs
collect brain processes triggered by an event
- events are a time period of interest usually in an experiment.
Event codes
- Cars vs faces shown
- Embed event codes for the time of the experiment that corresponds to the event of interest
- e.g. event code for stimulus onset, and response (correct or incorrect)
Uncovering ERP from messy data
- have sufficient number of events - at least 50 points of interest
- pre-processing: filtering, re-referencing etc.
- epoching: extract data according to event codes, around 1-1.5 millisecs around event
- baseline correction: make epochs the same scale, take average signal from baseline period (time before event) and subtract from entire epoch
- remove epochs with extreme values
- average epochs within a subject: average all epochs from a given condition - e.g. all car trials
- average across subjects: average of all averages for each participant
Peaks or components
- all peaks are named
- N1 = first negative
- P1 = first positive
- can also be named by time period.
We might analyse ERP…
- peak amplitude
- peak to peak deflection
- latency of peaks
- distribution of peaks on scalp
