Topographic ERP Analyses: A Step-by-Step Tutorial Review

Abstract

Using analysis of surface-recorded event-related potentials (ERPs) and reference-free spatial (topographic) information, response strength, latency, and topography between or within experimental conditions can be measured.
Measures somatosensory evoked potentials (SEPs) that were responses to stimulation of each hand.
Conceptual and mathematical descriptions.

Introduction

EEG and ERPs can store information more than just time course of brain responses or "components"
They can identify strength of response, latency, and source of response
Now there are 64-256 channels and amplifiers that digitize data from 500 Hz upwards, lots of data can be collected

How to Analyze Data

Choice: priori select time periods or components of interest
- ERP: "a set of waves, pick peaks and troughs, and measure amplitude and latency of deflections"
- However, some time slots that might be important can be overlooked.

Organization

To analyze limitations and pitfalls of canonical waveform analysis
Provide concrete examples
Detail the procedures for each step of electrical neuroimaging.

About the Data

Continuous EEG from 6 participants at 512 Hz (128 channels)
Peri-stimulus epochs of EEG averaged for stimulation of left and right hand and for each participant.
Tactile stimuli were square wave pulses
- two spatial positions and two vibration frequencies

ERP Limitations and Pitfalls

Voltage ERP waveforms have a limitation, and that is that they are reference-dependent.
Reference will remain to be a choice and a bias
Group-averaged ERPs was used as a reference to stimulate both left and right hands
If we use T7 to emulate left mastoid/earlobe vs T8 to do the right side, the ERPs are shown to be different in shape, latency, and variance.

ERP Limitations and Pitfalls

Equi-opposite effect for both C5 and C6 electrodes (left and right hand)
Only left hemisphere responses to tactile stimuli of the hand
All biases are considered correct. However, results can only apply for that particular reference and doesn't work for any other ones.

Why aren't canonical analyses of voltage waveforms as effective?

In an analysis of variance (ANOVA), modulations in topography cannot be differentiated from modulations in amplitude if the data isn't nicely scaled. There's a list of 3 distinct scaling methods:

Identify instantaneous maximum and minimum for each condition. Subtract the minimum and the difference between min and max.
Pre-defined value- ANOVAs data from voltage waveform will always be reference dependent

Methods to avoid ERP pitfalls

ERP voltage pitfalls happen with attention, language, and emotions as well.
EEG/ERP should identify a reference-independent measure. Make projected axis and equator indicate the 0 microVolts plane. Make sure shape of topography is constant even if color of given position changes. The confirugation of electric field on head should be reference independent.

Current Source Density (CSD)

alternative measurement to voltage waveforms.
eliminates reference-dependent problem that are inherent to voltage waveforms
However, they do not provide information as though for some of the neurophysiologic mechanism that cause modulation between experimental conditions.
Not readily calculated at the border of electrode montage
More sensitive to level of noise
Still faced the dilemma of time periods fo analyze

Why We Use the Average Reference

A common average reference is usually used for EEG/ERP.
Inverse solution method: recalculate data to a common average reference.
Net source activity at each brain sums to zero.
Need sufficient sampling to calculate that average.
Artifact-contaminated channels.

Global Field Power

GFP constitutes a single, reference-independent measure of response strength.
GFP is the root mean square of all average referenced electrode values
How strong a potential is being recorded. does not tell how potential is distributed.
Ex: four conditions are different.

Topographic ERG

By tsunwong625

Topographic ERG

407

tsunwong625