6: Assigning Events to Bins, Averaging, Baseline Correction, and Assessing Data Quality
-
-
Last updated
-
-
Save as PDF
-
Learning Objectives
In this chapter, you will learn to:
-
Design experiments that avoid subtle sensory confounds by following the Hillyard Principle
-
Create bin descriptor files for assigning events to bins
-
Predict how the signal-to-noise ratio will change as you change the number of trials being averaged together
-
Assess how the number of trials and the baseline period impact the data quality as quantified with the standardized measurement error (SME)
-
Create response-locked as well as stimulus-locked averages
-
Compare the ERPs for correct trials and error trials
-
Conduct sequential analyses
-
Evaluate the impact of overlapping activity from the previous trial
-
Implement a sequential analysis using a script
This chapter takes a close look at how events are assigned to bins. This is not something that gets discussed a lot in the ERP literature, even in methodology papers, but it’s absolutely fundamental. After all, “event” is the first word in “event-related potential.” Also, ERPLAB’s tool for assigning events to bins (BINLISTER) is fairly powerful but also not very user-friendly, so we’ll want to make sure you have a good grasp on how it works.
ERPLAB’s online documentation contains a detailed
manual page on BINLISTER
, and you should read it if you want to learn all the details of how this important routine works. I won’t repeat all those details here. In fact, I even wondered if it was worthwhile writing a chapter on bin assignment for this book. But I decided that the best way to learn about something like this is to actually use it. And once I started creating the exercises, I realized that they brought up some concepts about ERP data analysis that are important even if you end up using a different software package for analyzing your data.
This chapter will focus on the visual oddball P3b experiment from the ERP CORE, including an analysis of sequential effects and an analysis of the error-related negativity produced on error trials. The oddball paradigm has probably been used more than any other ERP paradigm over the years, so it’s good to have a thorough understanding of it. The particular version of this paradigm that we implemented for the ERP CORE contains some subtleties that are useful for learning about the design of ERP experiments. And the oddball paradigm is particularly well suited for exploring some of the issues that come up in assigning events to bins. We’ll start by taking a close look at the details of the experimental paradigm. Then we’ll perform several different analyses of one participant’s data so you can see some of the different ways that events can be assigned to bins and the issues that arise in this fundamental step.
Along the way, we’ll see some important issues that arise in averaging, including overlap from previous trials. We’ll also take a closer look at the baseline correction procedure, which seems simple but can end up creating problems in some situations. We’ll also look carefully at how data quality varies according to the number of trials being averaged together and according to the specific time period being used for baseline correction.