10.2: Design of the Flankers Experiment
-
- Last updated
- Save as PDF
This section will provide a brief overview of the experimental design and main results from the ERP CORE flankers experiment. This experiment was designed to isolate two different ERP components, the lateralized readiness potential (LRP) and the error-related negativity (ERN).
The LRP is a negative voltage over the hemisphere contralateral to a hand that is being prepared for a response. For example if I were to tell you “raise your left pinkie,” a negative voltage would be present over the motor cortex of the right hemisphere, preceding the actual pinkie response by 100-200 ms. The LRP is related to the preparation rather than the execution of the response. For example, if I were to tell you “raise your right pinkie when I say GO,” the negative voltage over the right hemisphere would appear right away, even before I said “GO.” Like the N2pc component, the LRP is isolated with a contralateral-minus-ipsilateral difference wave (relative to the side of the response hand). For an excellent review of the LRP, see Smulders and Miller (2012).
The ERN is a negative voltage with a maximum amplitude near FCz that occurs on error trials (usually in tasks where the error is obvious to the participant). It usually begins shortly before the actual buttonpress response and peaks 50-100 ms after the response. To get a good ERN, you need participants to make enough errors that you have a decent number of trials for averaging, but not so many errors that they’re not really putting any effort into doing the task correctly. For an excellent review of the ERN, see Gehring et al. (2012).
In the ERP CORE, we used a version of the Eriksen flankers task (Eriksen, 1995) to elicit these components. In the present chapter, we’ll focus on the LRP. As illustrated in Figure 10.1.A, each stimulus array contained a central target arrow that pointed leftward or rightward with equal probability, surrounded by two flanking arrows on each side. Participants were instructed to make a rapid buttonpress response for each array, pressing with the left index finger if the central arrow pointed leftward and with the right index finger if the central arrow pointed rightward. They were instructed to ignore the flankers, which pointed in the same direction as the central target arrow on 50% of trials (called Compatible trials) and pointed in the opposite direction on the other 50% (called Incompatible trials). Each array was presented for 200 ms, and successive arrays were separated by an interstimulus interval of 1200–1400 ms. To make sure we had a good number of error trials, participants were told to speed up if they were making errors on fewer than 10% of trials and to slow down if they were making errors on more than 20% of trials.
In the ERP CORE paper (Kappenman et al., 2021), we focused our LRP and ERN analyses on ERPs that were time-locked to the response (i.e., the response rather than the stimulus was used as time zero when the data were epoched and averaged). The LRP was isolated by examining trials with left-hand versus right-hand responses. Figure 10.1.B shows the grand average ERPs, with one waveform for the ipsilateral hemisphere (C3 for left-hand response trials averaged with C4 for right-hand response trials) and another waveform for the contralateral hemisphere (C4 for left-hand response trials averaged with C3 for right-hand response trials). The LRP is a negative voltage over the contralateral hemisphere that is superimposed on the other brain activity. To isolate the LRP, we construct a contralateral-minus-ipsilateral difference wave (Figure 10.1.C). You can see that the LRP in this difference wave starts to head in a negative direction approximately 120 ms prior to the response, and it peaks shortly before the response. (The slight positive voltage in the difference wave prior to -150 ms is likely an artifact of the 0.1 Hz high-pass filter that was applied to the continuous EEG at an early step in the analysis.)
The ERN is examined by comparing trials with correct trials and trials with incorrect trials (Figure 10.1.D). A sharp negative wave is superimposed on the positive voltage that would otherwise occur, peaking shortly after the response. This is then followed by a more positive voltage on the error trials than on the correct trials (called the error positivity or P E ). The ERN and P E are often isolated from the other brain activity by making an error-minus-correct difference wave (Figure 10.1.E).
In the main LRP analyses in the ERP CORE paper (Kappenman et al., 2021), we collapsed across compatible and incompatible trials for the sake of simplicity. In the present chapter, we’re going to look at the LRP separately for these trial types, focusing on stimulus-locked instead of response-locked averages. We’ll consider only the correct trials. Many prior experiments have found that response times (RTs) are slowed on incompatible trials relative to compatible trials, and this is mainly because information about the flankers “leaks through” to response selection mechanisms. On compatible trials, this helps to activate the correct response. On incompatible trials, however, the incorrect response may be activated, which slows down the response (and often leads to errors). This can sometimes be observed in the LRP waveform, which may be contralateral to the incorrect hand briefly before becoming contralateral to the correct hand on incompatible trials (Gratton et al., 1988). In addition, the onset latency of the LRP is delayed on incompatible trials relative to compatible trials.
We didn’t do any of these analyses for the ERP CORE paper, but I thought they would be interesting to do in the present chapter. For one thing, they’ll give us an opportunity to look at several different measures of amplitude and latency. For another, we won’t know the results until we do the analyses, so there will be some drama!