Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 1.
Published in final edited form as: Psychophysiology. 2023 Dec 4;61(3):e14497. doi: 10.1111/psyp.14497

The Impact of Electrode Selection for Ocular Correction on the Reward Positivity and Late Positive Potential Components in Adolescents

Samantha Pegg 1,*, Anh Dao 1, Lisa Venanzi 1, Kaylin Hill 1, Autumn Kujawa 1
PMCID: PMC10922232  NIHMSID: NIHMS1949130  PMID: 38044748

Abstract

Electroencephalography (EEG) data processing to derive event-related potentials (ERPs) follows a standard set of procedures to maximize signal-to-noise ratio. This often includes ocular correction, which corrects for artifacts introduced by eye movements, typically measured by electrooculogram (EOG) using facial electrodes near the eyes. Yet, attaching electrodes to the face may be uncomfortable for some populations, best to avoid in some situations, and contribute to data loss. Eye movements can also be measured using electrodes in a standard 10–20 EEG cap. An examination of the impact of electrode selection on ERPs is needed to inform best practices. The present study examined data quality when using different electrodes to measure eye movements for ocular correction (i.e., facial electrodes, cap electrodes, and no ocular correction) for two well-established and widely studied ERP components (i.e., reward positivity, RewP; and late positive potential, LPP) elicited in adolescents (N = 34). Results revealed comparable split-half reliability and standardized measurement error (SME) between facial and cap electrode approaches, with lower SME for the RewP with facial or cap electrodes compared to no ocular correction. Few significant differences in mean amplitude of ERPs were observed, including the LPP to positive images when using facial compared to cap electrodes. Findings provide preliminary evidence of the ability to collect high-quality ERP data without facial electrodes. However, when using cap electrodes for EOG measurement and ocular correction, it is recommended to use consistent procedures across the sample or statistically examine the impact of ocular correction procedures on results.

Keywords: event-related potentials, ocular correction, EEG, reward, emotion, artifacts

Introduction

Event-related potentials (ERPs), derived from the electroencephalogram (EEG), measure neurophysiological responses to discrete events—such as a stimulus presentation or participant response—as a way to capture brain activity tied to a process of interest. The range of processes captured via ERPs vary widely, including perceptual processing, attentional selection, error monitoring, emotional reactivity, and reward responsiveness (Woodman, 2010). For example, the reward positivity (RewP) and late positive potential (LPP) are two commonly studied components that are elicited across development and appear to capture processes relevant for the emergence of psychopathology (e.g., Kujawa et al., 2018; Kujawa, Klein, et al., 2013; Kujawa, Weinberg, et al., 2013; Pegg et al., 2019). In addition to the range of processes measured, ERPs as a tool of measurement offer high levels of temporal precision (e.g., timescale on the order of milliseconds), experimental flexibility (e.g., tasks can be modified across a large range of developmental periods and abilities), and relative affordability in comparison to other neuroimaging methods. Given these strengths, research using ERP methods is burgeoning, such that a recent Google Scholar search of “event-related potential” yielded over 300,000 results.

Many processing steps are typically applied to the raw EEG data to derive ERPs with high signal-to-noise ratios (Luck, 2014). Eye blinks are a particular source of noise in the EEG signal because there is a large, constant potential termed the corneal-retinal potential on the order of 50 to 100 μV for approximately 200 to 400 ms deflected when the eyelids close (Luck, 2014). For this reason, ERP researchers often ask participants to withhold blinks, add intertrial intervals to allow for blinks, and use processing procedures after recording to correct for eye blinks and eye movement activity. There have been several methods developed to correct for ocular artifacts after the data are collected, including independent component analysis (ICA; e.g., Makeig et al., 1996), principal component analysis (PCA; e.g., Berg & Scherg, 1991), and regression-based approaches (e.g., Elbert et al., 1985; Gratton et al., 1983). One widely accepted regression-based approach was developed by Gratton and colleagues (1983). As explained in their seminal paper, ocular movements introduce noise to the EEG data, which, of course, ERP researchers try to reduce to best capture the signal of interest. Particularly concerning about ocular movements such as blinks is that they are systematic muscle movements which may even overlap with events of interest, thus reducing or changing measures of the ERP component and reducing the signal-to-noise ratio (Gratton et al., 1983). To correct for these muscle movements, Gratton and colleagues’ (1983) algorithm measures signal from the electrooculogram (EOG) and EEG electrodes, separated by that which is event-related and not event-related, creates correction factors by regressing the EEG onto EOG data, and then corrects raw EEG data by subtracting the EOG value scaled by the regression-derived correction factor. Strengths of this approach to ocular correction include that it (1) uses the data collected during the task for the ocular correction algorithm, rather than needing a pre-task calibration period; (2) computes correction factors for different types of ocular movements separately; and (3) considers event-related EOG and EEG signals before creation of the correction factor as to avoid overcorrection for ocular movements (Gratton et al., 1983).

The growing emphasis on reproducibility and rigor in psychological science has led to increased examination of ERP processing methods, scoring decisions, and data quality considerations. One way to look at the impact of data processing on ERPs is the standardized measurement error (SME). Recently developed by Luck and colleagues (2021), the SME is a useful metric for estimating the precision of the ERP measurement—that is, the reliability of the neural signal in the signal-to-noise ratio—and thus provides researchers with the ability to empirically evaluate the impact of data collection and processing procedures on data quality. Another useful and commonly used metric of internal consistency in ERP data is split-half reliability. For ERPs, split-half reliability is calculated by scoring the ERP data in two halves, often split by odd and even trials, and computing the correlation between these halves, adjusted using the Spearman-Brown prophecy formula to predict for the full length of the task (Clayson et al., 2021). Given the potential impact of differences in processing procedures on the overall magnitude of ERPs, split-half reliability can be a useful tool for measuring internal consistency.

Previous research has compared the impact of different ocular correction procedures on EEG data. For example, research has compared common ocular correction procedures, including ICA, PCA, and regression-based algorithms, using simulated and real data (Jung et al., 2000; Wallstrom et al., 2004). Other work has utilized eye tracking with EEG data to compare ICA and regression-based procedures in their ability to correct for different types of eye artifacts (Plöchl et al., 2012). Given that the Gratton’s regression-based algorithm is widely used in research on the RewP and LPP components (e.g., Bress et al., 2015; Cuthbert et al., 2000; Ethridge et al., 2017; Foti & Hajcak, 2008; Hill et al., 2023; Kessel et al., 2018) and its longstanding history as a method of ocular correction, we use this method in the present study to increase applicability to others’ work. Further, performance of methods that do not require measurement of EOGs, such as ICA, in ocular correction is dependent on the number of electrodes and requires relatively clean data (Jung et al., 2000; Makeig et al., 1996). There may be situations where a lower number of electrodes than would be ideal are needed, and Gratton’s algorithm would continue to be an especially useful approach for ocular correction.

EOG data for the Gratton ocular correction method are typically collected from electrodes placed on the participant’s face in close proximity to the eyes. Specific procedures for EOG vary by EEG system, but it is common that 3 to 4 electrodes are placed on the participant’s face surrounding one or both eyes to capture ocular movements (e.g., about 1 cm above, below, and to the outside of one or both eyes). However, placement of facial electrodes is not always possible, feasible, or preferable, depending on experimental context. For example, some populations may be sensitive to electrodes placed on the face (e.g., people with sensory sensitivities), distracted or annoyed by such equipment (e.g., young children), or have limited facial surface area to fit the electrodes between the cap and the eyes (e.g., infants). Additionally, in contrast to electrodes in an elastic cap, facial EOG electrodes have to be attached to the skin with adhesive, which poses the risk of additional skin irritation or sensory aversion. Facial electrodes can also be difficult to obtain clean EOG signal from due to facial products (e.g., lotion, make-up) and can detach during EEG assessments, resulting in data loss when ERP processing relies on these data. In addition, recent situational factors have further limited researchers’ capacity to collect EOG data from facial electrodes. For example, it was suggested that fewer electrodes be used in ERP studies to reduce risk of illness transmission and increase safety in the COVID-19 pandemic (Simmons & Luck, 2020). Because EOG data is not typically of primary interest, these electrodes were candidates for exclusion from data collection procedures. Additionally, ocular electrodes are particularly challenging to place while participants wear face masks and often require direct face-to-face contact.

Alternatively, electrodes embedded within a typical EEG cap on the scalp may provide data for EOG measurement without the additional need for ocular electrodes placed on the face. Specifically, in the 10–20 system, Fp1 falls on the forehead just above the participant’s left eye which may be conducive to measuring vertical eye movements, while FT9 and FT10 fall near the left and right temples and may capture horizontal eye movements. These cap electrode sites are in close proximity to where conventional facial electrodes are placed, and as such, may be sufficient for measuring the EOG and correcting for ocular movements when facial electrodes are not feasible (see Figure 1). In the first study to our knowledge to test this possibility, we examined the effects of an EOG-based regression algorithm for ocular correction (Gratton et al., 1983) on two well-established reward and emotion ERPs, the RewP and LPP, which are commonly examined across development and offer insight into emotional processes and the emergence of psychopathology.

Figure 1. Front (A) and right side (B) views of scalp and facial electrodes used in ocular correction approaches. (C) Topographical plot of cap electrodes. Green circles indicate cap electrodes used in analyses. Orange circles indicate facial electrodes used in analyses.

Figure 1

Open in a new tab

Note: A side view of the left HEO facial electrode placement is not included because it is a mirror image of the right HEO facial electrode placement.

The RewP component, sometimes also referred to as the feedback negativity, is characterized as a relative positive-going deflection in the ERP waveform that peaks around 300 ms at frontocentral sites following positive compared to negative or neutral feedback (Proudfit, 2015). The RewP is a reliable measure of reward responsiveness that is thought to reflect a reward prediction error signal that drives reinforcement learning (Bress et al., 2015; Holroyd & Coles, 2002). The LPP component is characterized as a sustained positive deflection in the ERP waveform beginning about 300 ms often measured at centroparietal sites following stimulus onset that is larger for emotional stimuli compared to neutral stimuli (Cuthbert et al., 2000; MacNamara & Hajcak, 2010; Weinberg & Hajcak, 2010). The LPP is a reliable measure that is thought to reflect sustained attention and processing of emotional stimuli (Bondy et al., 2018; Hajcak et al., 2012). The RewP and LPP reliably emerge across development in simple reward tasks and in response to viewing pleasant and unpleasant images, respectively (Kujawa et al., 2018; Kujawa, Klein, et al., 2013; Kujawa, Weinberg, et al., 2013; Pegg et al., 2019). These components also provide an opportunity to examine the impact of ocular correction approaches on two reliable components with distinct spatial distributions that may be differentially impacted by ocular correction.

Given the utility of measuring these components to examine questions related to developmental psychopathology, adolescence is a critical time to assess the best practices of measurement for these components. Adolescence is an important and distinct developmental period in which rates of depression dramatically increase (Hankin et al., 1998; Kessler et al., 2001). It is a time in which research examining alterations in processes, including reward responsiveness and processing of emotional stimuli, is conducted to better understand pathways to depression. Additionally, having alternative options for electrode selection in measuring eye blinks and eye movements may be particularly relevant for developmental and longitudinal research. Thus, examining the impact of selection of electrodes in ocular correction approaches earlier in development, rather than adulthood, may be particularly relevant for informing best practices in developmental neuroscience research.

In the present study, we compared the impacts of electrode selection for ocular correction using Gratton’s algorithm on the RewP and LPP components with the goal of informing the handling of ERP data when facial electrodes may not be feasible or comfortable for participants. The three approaches for ocular correction consisted of standard collection and processing procedures using facial electrodes, a new approach using cap electrodes, and no ocular correction. This investigation was conducted using the 10–20 system given its common use to expand the feasibility and utility of this ocular correction approach for researchers who may find it useful to their work. Data were measured with an adolescent sample collected prior to the COVID-19 pandemic.

We used several measures to examine the extent to which ocular correction impacted the RewP and LPP components. Specifically, as measures of internal consistency and data quality, we examined split-half reliability and SME, respectively, across ocular correction approaches. We also compared the number of blinks detected between the facial and cap electrode approaches to assess whether the selection of electrodes for ocular correction differed in terms of ability to detect these changes in the EEG signal. Finally, we tested mean amplitudes of the RewP and LPP across approaches to assess whether differences were observed in scored components. We hypothesized that use of ocular correction (i.e., either cap or facial electrodes) would outperform data processed without ocular correction. Further, we anticipated that ocular correction using cap electrodes would provide high quality data, but that data quality may be better when using facial electrodes, given their specificity for measuring EOG and their closer proximity to the eyes.

Method

Participants

Participants were a subsample of adolescents aged 14–17 years old participating in a larger study of depression risk. The larger study included adolescents with a current depressive disorder diagnosis (i.e., currently depressed, n = 5 in this analysis), no personal history of clinical depression but a biological maternal history of a depressive disorder diagnosis (i.e., high risk, n = 12 in this analysis), or no personal or maternal history of clinical depression (i.e., low risk, n = 17 in this analysis). Exclusion criteria for the larger study included adolescent diagnoses of autism spectrum disorders or developmental disorders, intellectual disability, and/or history of depression without a current depressive episode at the intake interview, as well as adolescent or maternal lifetime mania and/or psychotic disorders outside of mood disorder features.

For the current study, there were 34 participants (Mage = 15.00, SD = 1.02 years) with usable data on either task, including usable VEO, HEO, Fp1, FT9, and FT10 data, collected before the COVID-19 pandemic for analyses. In terms of sex, 73.5% of the sample identified as female. Regarding race, 8.8% of the sample identified as African American or Black, 5.9% of Asian descent, 79.4% as White or Caucasian, and 5.9% identified as another race. All participants identified as not Hispanic or Latino/a/x. Of these 34 participants, for the monetary reward task data, 4 were missing data due to (1) error at data collection (n = 1), (2) noisy mastoid data (n = 1), (3) being an outlier case using the Grubbs’ test (Grubbs, 1950) (n = 1), and (4) noisy data at Fp1 (n = 1). For the emotional interrupt task data, 7 participants were missing data due to (1) noisy VEO data (n = 4), (2) low number of segments (n = 1), (3) being an outlier case using the Grubbs’ test (Grubbs, 1950) (n = 1), and (4) not completing the task during the EEG visit (n = 1).

Procedure

Study procedures were approved by the Vanderbilt University Institutional Review Board. Consent was obtained according to the Declaration of Helsinki prior to study enrollment. Diagnoses were confirmed with the Kiddie Schedule of Affective Disorders and Schizophrenia for adolescents (Kaufman et al., 1997) and the Structured Clinical Interview for DSM-5 for mothers (First et al., 2016). Adolescents were then invited to campus to complete the EEG assessment. For this portion of the study, participants received $50 for completing the diagnostic interviews and questionnaires and $30 for completing the EEG assessment. If participants completed all parts of the study, they could earn up to $140 total.

Measures

Monetary Reward Task

Participants completed a simple guessing reward task (i.e., Doors task), a well-established monetary reward task that has been found to reliably elicit RewP across development (Ethridge et al., 2017; Hill et al., 2023; Kujawa et al., 2015), while continuous EEG data were recorded. The task is frequently used to examine the effects of depressive symptoms on neural responses to rewards (Proudfit, 2015). Participants were informed that they could win up to $5 based on their performance in the task. At the start of each trial, participants were presented with two doors on a computer screen. They were instructed to press the right or left mouse button to choose one of the doors to guess which had a prize behind it. After participants made their selection, a fixation cross was displayed for 1,000 ms, followed by feedback about each selection, presented for 1,500 ms. The task included two types of feedback: a green upward arrow, indicating a win of $0.50, and a red downward arrow, indicating a loss of $0.25. Following the presentation of feedback, a fixation cross was shown again on the screen for 1,000 ms, and participants were prompted to click for another round. Participants completed a practice block prior to the actual task to ensure understanding. The actual task had 30 win and 30 loss trials that were presented in a random order. At the end of the task, all participants were informed that they won the full $5 amount.

Emotional Interrupt Task

Participants also completed the emotional interrupt task (Mitchell et al., 2006), a well-established task designed to elicit the LPP in response to emotionally salient images. Prior studies have shown that this task reliably elicits LPP across development (Bondy et al., 2018; Kessel et al., 2018; Kujawa et al., 2012). The version of the task used in the current study includes interpersonally relevant images that have been previously validated in young adults (Dickey et al., 2021). Developmentally appropriate positive social (i.e., parents and teens hugging, teens having fun), negative social (i.e., bullying, teens arguing with their parents), and neutral (i.e., nature scenes, objects) images included in the task were selected from validated picture sets and stock photos so that 15 images were shown for each category. At the beginning of each trial, a fixation cross was presented for 800 ms, followed by an image displayed for 1,000 ms. Next, a right or left arrow was shown for 150 ms, and participants were instructed to click the left or right mouse button to indicate the direction of the arrow as soon as it appeared on the screen to ensure attention to the task. The same image was then shown on the screen for 400 ms, followed by a fixation cross screen for 1,500–2,000 ms. Participants completed six practice trials before starting the task, and the actual task included two blocks with images presented once per block for a total of 90 trials. The LPP is measured following the first image presentation, and only correct trials were included in analyses, consistent with previous work on this task (Dickey et al., 2021).

EEG data collection

EEG data were continuously recorded with a 32-electrode cap actiCHamp system from BrainProducts (Munich, Germany) based on a standard 10–20 layout with Cz as the online reference. Vertical and horizontal EOG (VEO and HEO, respectively) were measured using facial electrodes and compared to measurements acquired from cap electrodes Fp1, FT9, and FT10. For VEO, electrodes were placed approximately 1 cm above and below either the left or right eye to measure blinks, with the center of the above eye VEO facial electrode placed approximately 2 cm from the center of either Fp1 or Fp2. For HEO, electrodes were placed approximately 1 cm from the outer corners of the eyes to capture horizontal saccades, with the center of each HEO electrode placed approximately 5 cm from the centers of FT9 and FT10 (see Figure 1). Bipolar electrodes were referenced to an electrode placed on the back of the neck of the participant, per the BrainProducts bipolar-to-auxiliary adapter design, which was set to a resolution of 1 μV and gradient of 0.1 mV/μV. Impedances were reduced to below approximately 30 kΩ. A 24-bit resolution and sampling rate of 1000 Hz were used to digitize the EEG recordings.

EEG data processing

Data were processed offline using BrainVision Analyzer, Version 2.1.2 (BrainVision Analyzer, 2019). Data were filtered and re-referenced to the averaged left and right mastoid recordings (TP9/TP10). For the reward task data, a bandpass filter was used with a low-cutoff of 0.1 Hz and high-cutoff of 30 Hz, and a 60 Hz notch filter applied to VEO and HEO1. For the emotional interrupt task data, a bandpass filter was used with a low-cutoff of 0.01 Hz (to avoid attenuating later stages of the LPP) and high-cutoff of 30 Hz. The reward task data were segmented into 1,000 ms segments, which began 200 ms before stimulus onset and ended 800 ms after. The emotional interrupt task data were segmented into 1,200 ms segments, also beginning 200 ms before the stimulus onset and ending 1,000 ms after stimulus onset.

EEG data were processed with three processing trees, one for each set of electrodes used in ocular correction and one to omit ocular correction. In the standard processing iteration, ocular correction was conducted using Gratton’s algorithm (Gratton et al., 1983) with VEO and HEO, hereinafter referred to as the “facial electrodes” approach. The second ocular correction approach, referred to as the “cap electrodes” approach, used Fp1 with common reference in place of VEO and FT9 with FT10 as the reference in place of HEO. Figure 1 depicts the electrode placement to compare distances between facial electrodes and those in the cap. The third processing iteration omitted ocular correction completely, going from segmentation to artifact rejection.

Following ocular correction, automatic artifact rejection was applied using the following parameters: maximal allowed voltage step of 50 μV/ms, maximal allowed difference of values of 175 μV in intervals of 400 ms, minimal allowed amplitude of −200 μV and maximal allowed amplitude of 200 μV, and lowest allowed activity of 0.5 μV in intervals of 100 ms. No channels were interpolated to ensure consistency across ocular correction approaches. The reward task data were then separated into win and loss conditions with averages computed per condition and baseline corrected 200 ms before feedback onset. The emotional interrupt task data were separated into positive, negative, and neutral conditions. Each condition segment was then averaged and baseline corrected 200 ms before stimulus onset. ERPs were scored based on prior literature (DeCicco et al., 2014; Foti & Hajcak, 2008; Krigolson, 2018; Kujawa et al., 2012; Proudfit, 2015; Speed et al., 2015; Stange et al., 2017) and visual inspection of the grand averaged data. Specifically, the RewP was scored at 250–350 ms at Cz (Figure 2), and the LPP was scored at 400–1,000 ms at an occipital pooling of O1, O2, and Oz (Figure 3). After artifact rejection, the average number of segments across conditions and ocular correction approaches ranged from 29.97 to 30.00 at Cz for the RewP and from 28.22 to 28.41 for the LPP at O1, O2, and Oz. The number of segments at Cz did not differ between ocular correction approaches for RewP for either the win or loss condition, and there were no significant differences in the number of segments per ocular correction approach, condition, or electrode (i.e., O1, O2, or Oz) for the LPP, ps > .237. To calculate SME, individual trial data were processed and baseline corrected 200 ms before feedback/image onset then exported for further analysis.

Figure 2. Event-related potential waveform and scalp distributions for each ocular correction approach and corresponding scalp distributions for the reward positivity.

Figure 2

Open in a new tab

Figure 3. Event-related potential waveform and scalp distributions for each ocular correction approach and corresponding scalp distributions for the late positive potential.

Figure 3

Open in a new tab

Data Analysis

Split-half reliability was examined for each ocular correction approach by calculating the Spearman-Brown coefficient for averages of even and odd trials for each condition to compare internal consistency of the components across approaches. The number of detected blinks were extracted from BrainVision Analyzer to examine differences between the facial and cap electrode ocular correction approaches in detecting blinks. These values were then compared using paired samples t-tests. SME was computed to examine how variable an individual’s performance was across trials (Luck et al., 2021). To examine effects of ocular correction approach on SME, a 2 (condition: win, loss) x 3 (ocular correction approach: facial electrodes, cap electrodes, and no ocular correction) repeated-measures ANOVA for the reward task data, and a 3 (condition: positive, negative, neutral) x 3 (ocular correction approach) repeated-measures ANOVA for the interpersonal emotional interrupt task data was conducted. Significant effects were further probed using pairwise comparisons.

Lastly, to examine effects on overall RewP and LPP magnitude, a 2 (condition: win, loss) x 3 (ocular correction approach: facial electrodes, cap electrodes, and no ocular correction) repeated-measures ANOVA for the reward task and a 3 (condition: positive, negative, neutral) x 3 (ocular correction approach) repeated-measures ANOVA for the interpersonal emotional interrupt task data were conducted. Pairwise comparisons were conducted to interpret significant effects. For all ANOVAs, Greenhouse-Geisser correction was applied where assumptions of sphericity were violated. To be transparent in presenting all potential impacts of ocular correction methods on ERPs, we did not correct for multiple comparisons in any analyses.

Results

Split-half Reliability

Spearman-Brown coefficients for each condition and ocular correction approach are presented in Table 1. For the RewP, split-half reliability was excellent across ocular correction approaches for RewP to win (Spearman-Brown coefficients: .93-.94) and loss feedback (Spearman-Brown coefficients: .92-.93). For the LPP, split-half reliability was also good across ocular correction approaches for LPP to positive (Spearman-Brown coefficients: .81-.83), negative (Spearman-Brown coefficients: .85-.89), and neutral images (Spearman-Brown coefficients: .81-.85). Overall, for both RewP and LPP, split-half reliability values were similar or the same within condition for each ocular correction approach. This suggests that both ERPs were reliably elicited across ocular correction approaches.

Table 1.

Spearman-Brown coefficients for split-half reliability by event-related potential and ocular correction approach

Facial electrodes Cap electrodes No ocular correction
Spearman-Brown Coefficient Spearman-Brown Coefficient Spearman-Brown Coefficient
RewP
Win .93 .93 .94
Loss .93 .92 .92
LPP
Positive .81 .83 .81
Negative .85 .89 .86
Neutral .85 .81 .83
Open in a new tab

Note: LPP = late positive potential; RewP = reward positivity

Blinks Detected

For the RewP, the facial electrodes ocular correction approach detected more blinks (M = 39.70, SD = 18.82) compared to the cap electrodes (M = 34.80, SD = 19.26), t(29) = 2.59, p = .015 (two-sided), Cohen’s d = 0.47. Similarly for the LPP, the facial electrodes ocular correction approach detected more blinks (M = 28.37, SD = 22.31) compared to the cap electrodes (M = 15.30, SD = 18.10), t(26) = 5.14, p < .001 (two-sided), Cohen’s d = 0.99.

Standardized Measurement Error

Means and standard deviations for SME values for the RewP and LPP are presented in Table 2. Lower values indicate higher data quality (Luck et al., 2021). For RewP, a 2 (condition: win, loss) x 3 (ocular correction approach: facial electrodes, cap electrodes, no ocular correction) repeated-measures ANOVA was conducted. There was no main effect of condition, F(1, 29) = 0.21, p = .653, partial η2 = 0.01. There was a significant main effect of ocular correction approach, F(1.19, 34.43) = 18.98, p < .001, partial η2 = 0.40. Pairwise comparisons indicated that SME values were significantly lower for cap electrodes compared to facial electrodes, F(1, 29) = 23.77, p < .001, partial η2 = 0.45. SME values were significantly lower for facial electrodes compared to no ocular correction, F(1, 29) = 11.44, p = .002, partial η2 = 0.28. SME values were significantly lower for cap electrodes compared to no ocular correction, F(1, 29) = 25.96, p < .001, partial η2 = 0.47. There was no condition x ocular correction approach interaction, F(1.24, 35.88) = 1.50, p = .234, partial η2 = 0.05.

Table 2.

Means and standard deviations for standardized measurement error (SME) by event-related potential and ocular correction approach

Facial electrodes Cap electrodes No ocular correction
SME M (SD) SME M (SD) SME M (SD)
RewP
Win 2.26 (0.44) 2.20 (0.46) 2.42 (0.55)
Loss 2.31 (0.57) 2.24 (0.57) 2.41 (0.62)
LPP
Positive 1.87 (0.70) 1.80 (0.66) 1.78 (0.68)
Negative 1.73 (0.47) 1.66 (0.40) 1.67 (0.42)
Neutral 1.74 (0.61) 1.76 (0.70) 1.71 (0.61)
Open in a new tab

Note: LPP = late positive potential; RewP = reward positivity

For the LPP, a 3 (condition: positive, negative, neutral) x 3 (ocular correction approach) repeated-measures ANOVA was conducted. There was no main effect of condition, F(2, 52) = 0.78, p = .463, partial η2 = 0.03; no main effect of ocular correction approach, F(2, 52) = 2.36, p = .104, partial η2 = 0.08; and no condition x ocular correction approach interaction, F(2.95, 76.63) = 1.19, p = .320, partial η2 = 0.04.

ERP Amplitude

Means and standard deviations for the mean amplitudes of the RewP and LPP are presented in Table 3. As expected, the RewP was enhanced for win compared to loss feedback across ocular correction approaches (main effect of condition: F(1, 29) = 21.40, p < .001, partial η2 = 0.43). The LPP was also modulated by condition across ocular correction approaches as expected (main effect of condition: F(2, 52) = 12.39, p < .001, partial η2 = 0.32), with the LPP being enhanced overall for both positive (F(1,26) = 12.31, p = .002, partial η2 = 0.32) and negative (F(1,26) = 18.49, p < .001, partial η2 = 0.42) compared to neutral images.

Table 3.

Means and standard deviations for mean amplitudes by event-related potential (ERP) and ocular correction approach

Facial electrodes Cap electrodes No ocular correction
ERP M (SD) ERP M (SD) ERP M (SD)
RewP
Win 15.03 (7.24) 15.17 (6.74) 16.23 (7.60)
Loss 11.66 (7.09) 12.21 (6.53) 12.64 (7.16)
LPP
Positive 6.31 (4.99) 6.74 (5.01) 6.81 (4.96)
Negative 7.02 (4.80) 7.36 (4.63) 7.58 (4.60)
Neutral 4.09 (4.87) 4.04 (4.57) 4.27 (4.57)
Open in a new tab

Note: LPP = late positive potential; RewP = reward positivity

For the RewP, a 2 (condition: win, loss) x 3 (ocular correction approach) repeated-measures ANOVA was conducted. There was a significant main effect of ocular correction approach, F(1.59, 46.17) = 5.78, p = .009, partial η2 = 0.17. There was also a condition x ocular correction approach interaction, F(1.43, 41.37) = 4.52, p = .027, partial η2 = 0.14. To probe this significant interaction, the repeated measures ANOVA was conducted within each condition.

For the RewP to loss feedback, there was a main effect of ocular correction approach, F(1.64, 47.48) = 3.66, p = .041, partial η2 = 0.11. Pairwise comparisons revealed a difference between facial electrodes and cap electrodes at a trend level, F(1, 29) = 4.16, p = .051, partial η2 = 0.13, with mean amplitudes being more positive for cap compared to facial electrodes. There was a significant difference between facial electrodes and no ocular correction, F(1, 29) = 6.66, p = .015, partial η2 = 0.19, such that mean amplitudes for the no ocular correction approach were more positive compared to facial electrodes. There was no significant difference between cap electrodes and no ocular correction, F(1, 29) = 1.04, p = .317, partial η2 = 0.03.

For the RewP to win feedback, there was a main effect of ocular correction approach, F(1.51, 43.74) = 8.17, p = .002, partial η2 = 0.22. Pairwise comparisons revealed no significant difference between facial electrodes and cap electrodes, F(1, 29) = 0.39, p = .538, partial η2 = 0.01. There was a significant difference between facial electrodes and no ocular correction, F(1, 29) = 12.97, p = .001, partial η2 = 0.31, such that mean amplitudes for the no ocular correction approach were more positive compared to facial electrodes. There was also a significant difference between cap electrodes and no ocular correction, F(1, 29) = 7.28, p = .012, partial η2 = 0.20, such that mean amplitudes for the no ocular correction approach were more positive compared to cap electrodes.

For the LPP, a 3 (condition: positive, negative, neutral) x 3 (ocular correction approach) repeated-measures ANOVA was conducted. There was no main effect of ocular correction approach, F(1.41, 36.64) = 2.85, p = .087, partial η2 = 0.10. There was a significant condition x ocular correction approach interaction, F(4, 104) = 2.74, p = .033, partial η2 = 0.10.

For the LPP to neutral images, there was no main effect of ocular correction approach, F(1.26, 32.76) = 0.59, p = .483, partial η2 = 0.02. For the LPP to positive images, there was a significant main effect of ocular correction approach, F(2, 52) = 4.82, p = .012, partial η2 = 0.16. There was a significant difference between the facial electrodes and cap electrodes, F(1, 26) = 4.96, p = .035, partial η2 = 0.16, such that the LPP with the cap electrodes approach was significantly more positive compared to the facial electrodes approach. There was also a significant difference between facial electrodes and no ocular correction, F(1, 26) = 10.87, p = .003, partial η2 = 0.30, such that the LPP with no ocular correction was significantly more positive compared to facial electrodes approach. There was no significant difference between cap electrodes and no ocular correction, F(1, 26) = 0.18, p = .676, partial η2 = 0.01.

For the LPP to negative images, there was a significant main effect of ocular correction approach, F(1.63, 42.36) = 4.32, p = .026, partial η2 = 0.14. Pairwise comparisons found no significant difference between facial electrodes and cap electrodes, F(1, 26) = 2.29, p = .142, partial η2 = 0.08. There was a significant difference between facial electrodes and no ocular correction, F(1, 26) = 14.34, p < .001, partial η2 = 0.36, such that no ocular correction was significantly more positive compared to the facial electrodes approach. No significant difference was found between cap electrodes and no ocular correction approaches, F(1, 26) = 1.25, p = .275, partial η2 = 0.05.

Discussion

EEG data are commonly collected with electrodes placed on the face near the eyes to measure the EOG and adjust the EEG data for common eye movement and blink artifacts. However, this method can be limited in feasibility for research with some populations, such as young children or people with sensory sensitivities, and certain contexts, such as needing to minimize time in close proximity to the researcher or equipment preparation time. In this study, we examined the extent to which data quality and reward- and emotion-related ERP (i.e., RewP and LPP) magnitudes in adolescents are impacted between different ocular correction electrode selections with a commonly used regression-based ocular correction procedure (Gratton et al., 1983). Specifically, we compared facial electrodes (i.e., VEO and HEO) versus cap electrodes located near the eyes (i.e., Fp1 in place of VEO and FT9 with reference FT10 for HEO) to identify eye blinks and eye movements. We also compared both approaches to no ocular correction.

For the RewP, split-half reliability was excellent across ocular correction approaches for both win and loss conditions. The facial electrode approach detected more blinks compared to the cap electrodes approach. SME values were significantly lower for the cap electrode approach compared to facial electrode approach, and significantly lower for both facial and cap electrode approaches compared to no ocular correction, suggesting that data quality was better for the cap electrodes compared to facial electrodes and for either ocular correction approach compared to no correction. Mean amplitudes for the RewP did not significantly differ between facial and cap electrode approaches across conditions. Mean amplitudes for both RewP to losses and wins were significantly more positive for no ocular correction compared to the facial electrode approach, and also more positive compared to the cap electrode approach for the RewP to wins condition.

Similar to the RewP, split-half reliability was also good across ocular correction approaches for the LPP to positive, negative, and neutral images. The facial electrode approach also detected more blinks compared to cap electrodes. SME values did not significantly differ between conditions or ocular correction approaches, suggesting comparable data quality across the tested methods. Mean amplitudes for the LPP did not significantly differ between the facial and cap electrode approaches, except for the LPP to positive images condition where values were more positive for the cap electrode approach compared to the facial electrode approach. Mean amplitudes were also significantly more positive for no ocular correction compared to the facial electrode approach for the LPP to positive and negative images.

For both ERPs, split-half reliability was comparable across ocular correction approaches, suggesting that the RewP and LPP can be reliably elicited regardless of the electrodes selected for ocular correction. Results across both ERPs suggest that data quality was comparable across ocular correction approaches as measured by SME. It is also worth noting that SME values were significantly lower when using the cap electrodes compared to facial electrodes for the RewP. Although surprising, this finding further supports that it is possible to obtain high quality data using cap electrodes in place of facial electrodes in ocular correction. Further, SME values for both the RewP and LPP in the present sample were similar to or slightly higher than those observed in a recent study with a larger adult sample, where SME values for RewP win and loss conditions on a similar monetary reward task ranged from 1.80–1.98, (Clayson et al., 2021), as well as to the P3b component measured in a larger adult sample, where SME values ranged from 0.81–1.83 (Zhang & Luck, 2023). As Luck and colleagues (2021) describe, it is unclear what would quantify a “good” SME value in general. Results of the present study suggest that either set of electrodes used for ocular correction between cap and facial electrodes appeared comparable for the RewP and LPP components and were improved for the RewP compared to data processed without ocular correction.

When examining differences between ocular correction approaches for the mean amplitudes, the only significant difference was for the LPP to positive images between facial and cap electrodes, such that mean amplitudes were more positive for cap electrodes compared to facial electrodes. These differences may be related in part to blink detection. Given that the facial electrode approach detected significantly more blinks compared to cap electrodes for both ERPs, it is possible that this greater number of detected and corrected blinks attenuated the data to a greater extent when using facial electrodes compared to cap electrodes in Gratton’s algorithm (Gratton et al., 1983). On the other hand, the additional detected blinks for facial electrode compared to cap electrode approaches for ocular correction may have also artificially increased the individual conditions for the RewP and LPP. It is important to note that it is unclear which mean amplitude value would be considered the more “true” value, given the lack of availability of normative values across development for these two ERPs and likely variability across systems, tasks, and processing parameters.

Although we tested ocular correction approaches for two commonly studied ERPs in the reward and emotion ERP literature, both the RewP and LPP are relatively large ERP components in terms of amplitude and appear later in the ERP waveform. It is possible that smaller or earlier components, such as early sensory and attentional processing components like N1 or P2, may be impacted more by differences in processing approaches. Further, Gratton’s algorithm was used in ocular correction across the facial electrode and cap electrode approaches (Gratton et al., 1983), which needs a difference between two channels (e.g., top and bottom of eye to capture vertical movements, right and left of eye to capture horizontal movements). Using facial electrodes, a pair of electrodes is often used to capture this difference, while the cap electrodes do not have a channel below the eye when using Fp1 to measure vertical eye movements. A common reference can be used instead, but this is a difference between facial and cap electrode options worth noting. Also, the propagation factor applied in this algorithm is decreased from frontal to parietal electrodes (Gratton et al., 1983). ERPs scored farther from the source of eye blinks and eye movements, such as the LPP, may not be as impacted by ocular correction given their location on the scalp. However, the RewP was scored at a central site (i.e., Cz) closer to the source of eye blinks and eye movements, and results with this component suggested consistent reliability across ocular correction approaches, so this may not be of great concern. Future work should test other ERP components that occur earlier in the ERP waveform, that are smaller in magnitude, and that are typically scored closer to the face to examine potential impacts of differences in ocular correction approaches between facial and cap electrodes to replicate and expand the current findings.

Although the 10–20 system and electrode layout used in the present study is commonly used in EEG research, it is important to note that this layout may differ from those used by other researchers. We encourage researchers to be thoughtful in their choice of electrodes used in ocular correction based on their cap and layout. We recommend selecting electrodes closest to where VEO and HEO facial electrodes would be placed, testing the impact of processing decisions on data quality, and using a consistent approach within their studies or account for processing differences in analyses. It may be beneficial for future work to also test whether electrode density influences ocular correction procedures given the 32-electrode scheme used in the present study. Additionally, the sample was comprised of adolescents, and future work is needed to test the use of different electrodes to measure eye movements for ocular correction across development, particularly in infants and younger children; eliminating facial electrodes in EEG research with young participants may minimize burden for children and their caregivers as well as enhance feasibility if results are comparable between facial and cap electrodes in these groups. Future work could also employ other methods of investigating implications of ocular correction procedures. This could include simulating data to compare the influence of different types of eye blinks and eye movements on data quality across ocular correction approaches, which may provide insight into the influence of these artifacts on multiple ERP components in a controlled way.

There are several instances where using facial electrodes may be limited in feasibility, such as needing to reduce time and close contact with participants during the COVID-19 pandemic, or may not be feasible to use, such as when working with young children or other sensitive populations who may find facial electrodes uncomfortable or difficult to wear during EEG data collection. Results from the present study provide preliminary support for the use of cap electrodes in place of facial electrodes for ocular correction to preserve as much data as possible for data analysis while accounting for eye movements. We recommend that researchers use consistent ocular correction approaches for all participants in a study, when possible, given the potential impacts on overall ERP magnitude. When unexpected challenges introduce variability in ocular correction approaches within a sample, we recommend that researchers examine the impact of the use of different electrodes in ocular correction on the overall ERP waveforms and include this as a covariate in analyses to further understand its impact on findings. In addition, it is important to report the choice of electrodes for ocular correction approaches in Method sections. With this additional tool available for ERP research, using cap electrodes for ocular correction may improve statistical power in ERP analyses by reducing data loss and allowing for inclusion of more participants. At the same time, this approach should be implemented cautiously pending further evidence in support, and we recommend that ERP researchers continue to examine the impact of processing decisions like ocular correction electrode selection on data quality in their own studies.

Impact statement.

This study examines the effects of using cap compared to facial electrodes to measure eye movements for ocular correction in event-related potential (ERP) research. We test the impact of electrode selection on data quality and mean amplitudes of reward- and emotion-related ERPs in adolescents. Findings revealed comparable data quality and some differences in overall ERP magnitude, suggesting cap electrodes may be used if needed. A consistent approach to ocular correction within a study is recommended.

Funding statement:

This project was supported by a Klingenstein Third Generation Foundation Fellowship and Brain and Behavior Research Foundation Katherine Deschner Family Young Investigator Grant awarded to AK. KEH was supported in part from a National Institute for Mental Health training grant T32 MH18921 during completion of this work. SP was supported by National Institute for Mental Health grants T32 MH18921 and F31 MH127817 during completion of this work.

Footnotes

Conflict of interest disclosure: The authors report no competing financial or non-financial interests that are directly or indirectly related to the work submitted for publication.

Ethics approval statement: The study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Vanderbilt University Institutional Review Board.

1

In the monetary reward task only, we were initially concerned about external sources of noise that differentially impacted VEO and HEO given the BrainProducts design, so the 60 Hz notch filter was applied only to these electrodes. To examine potential impacts of this processing decision, we reprocessed the data for this task without the notch filter and computed bivariate correlations between the RewP data using facial electrodes for ocular correction with and without the 60 Hz notch filter applied to VEO and HEO. The RewP values were highly correlated across conditions (rs > .995).

CrediT statement:

Samantha Pegg: conceptualization, data curation, formal analysis, methodology, investigation, writing – original draft, writing – review and editing

Anh Dao: formal analysis, data curation, project administration, visualization, writing – original draft, writing – review and editing

Lisa Venanzi: formal analysis, visualization, writing – original draft, writing – review and editing

Kaylin Hill: methodology, writing – original draft, writing – review and editing

Autumn Kujawa: conceptualization, funding acquisition, methodology, investigation, supervision, writing – original draft, writing – review and editing

Data and code availability:

The data that support the findings of this study are openly available in OSF at https://osf.io/796ma/?view_only=588654e549bd4aeea71892ac78d42db2.

References

  1. Berg P, & Scherg M (1991). Dipole modelling of eye activity and its application to the removal of eye artefacts from the EEG and MEG. Clinical Physics and Physiological Measurement, 12(A), 49–54. 10.1088/0143-0815/12/A/010 [DOI] [PubMed] [Google Scholar]
  2. Bondy E, Stewart JG, Hajcak G, Weinberg A, Tarlow N, Mittal VA, & Auerbach RP (2018). Emotion processing in female youth: Testing the stability of the late positive potential. Psychophysiology, 55(2), e12977. 10.1111/psyp.12977 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bress JN, Meyer A, & Proudfit GH (2015). The stability of the feedback negativity and its relationship with depression during childhood and adolescence. Development and Psychopathology, 27(4 Pt 1), 1285–1294. 10.1017/S0954579414001400 [DOI] [PubMed] [Google Scholar]
  4. Clayson PE, Brush CJ, & Hajcak G (2021). Data quality and reliability metrics for event-related potentials (ERPs): The utility of subject-level reliability. International Journal of Psychophysiology, 165, 121–136. 10.1016/j.ijpsycho.2021.04.004 [DOI] [PubMed] [Google Scholar]
  5. Cuthbert B, Schupp H, Bradley M, Birbaumer N, & Lang P (2000). Brain potentials in affective picture processing: Covariation with autonomic arousal and affective report. Biological Pyschology, 52(2), 95–111. 10.1016/s0301-0511(99)00044-7 [DOI] [PubMed] [Google Scholar]
  6. DeCicco JM, O’Toole LJ, & Dennis TA (2014). The late positive potential as a neural signature for cognitive reappraisal in children. Developmental Neuropsychology, 39(7), 497–515. 10.1080/87565641.2014.959171 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dickey L, Pegg S, & Kujawa A (2021). Neurophysiological responses to interpersonal emotional images: Associations with symptoms of depression and social anxiety. Cognitive, Affective & Behavioral Neuroscience, 21(6), 1306–1318. 10.3758/s13415-021-00925-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Elbert T, Lutzenberger W, Rockstroh B, & Birbaumer N (1985). Removal of ocular artifacts from the EEG — A biophysical approach to the EOG. Electroencephalography and Clinical Neurophysiology, 60(5), 455–463. 10.1016/0013-4694(85)91020-X [DOI] [PubMed] [Google Scholar]
  9. Ethridge P, Kujawa A, Dirks MA, Arfer KB, Kessel EM, Klein DN, & Weinberg A (2017). Neural responses to social and monetary reward in early adolescence and emerging adulthood. Psychophysiology, 54(12), 1786–1799. 10.1111/psyp.12957 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. First MB, Williams JBW, Karg RS, & Spitzer RL (2016). Structured Clinical Interview for DSM-5-Clinician Version (SCID-5-CV). American Psychiatric Association. [Google Scholar]
  11. Foti D, & Hajcak G (2008). Deconstructing reappraisal: Descriptions preceding arousing pictures modulate the subsequent neural response. Journal of Cognitive Neuroscience, 20(6), 977–988. 10.1162/jocn.2008.20066 [DOI] [PubMed] [Google Scholar]
  12. Gratton G, Coles MG, & Donchin E (1983). A new method for off-line removal of ocular artifact. Electroencephalography and Clinical Neurophysiology, 55(4), 468–484. 10.1016/0013-4694(83)90135-9 [DOI] [PubMed] [Google Scholar]
  13. Grubbs FE (1950). Sample criteria for testing outlying observations. The Annals of Mathematical Statistics, 21(1), 27–58. 10.1214/aoms/1177729885 [DOI] [Google Scholar]
  14. Hajcak G, Weinberg A, MacNamara A, & Foti D (2012). ERPs and the study of emotion. In Luck SJ & Kappenman ES (Eds.), The Oxford handbook of event-related potential components (pp. 441–472). Oxford University Press. [Google Scholar]
  15. Hankin BL, Abramson LY, Moffitt TE, Silva PA, McGee R, & Angell KE (1998). Development of depression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1), 128–140. 10.1037/0021-843X.107.1.128 [DOI] [PubMed] [Google Scholar]
  16. Hill KE, Dickey L, Pegg S, Dao A, Arfer KB, & Kujawa A (2023). Associations between parental conflict and social and monetary reward responsiveness in adolescents with clinical depression. Research on Child and Adolescent Psychopathology, 51(1), 119–131. 10.1007/s10802-022-00949-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Holroyd CB, & Coles MGH (2002). The neural basis of human error processing: Reinforcement learning, dopamine, and the error-related negativity. Psychological Review, 109(4), 679–709. 10.1037/0033-295X.109.4.679 [DOI] [PubMed] [Google Scholar]
  18. Jung TP, Makeig S, Humphries C, Lee TW, McKeown MJ, Iragui V, & Sejnowski TJ (2000). Removing electroencephalographic artifacts by blind source separation. Psychophysiology, 37(2), 163–178. 10.1111/1469-8986.3720163 [DOI] [PubMed] [Google Scholar]
  19. Kaufman J, Birmaher B, Brent D, Rao U, Flynn C, Moreci P, Williamson D, & Ryan N (1997). Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): Initial reliability and validity data. Journal of the American Academy of Child & Adolescent Psychiatry, 36(7), 980–988. 10.1097/00004583-199707000-00021 [DOI] [PubMed] [Google Scholar]
  20. Kessel EM, Nelson BD, Kujawa AJ, Hajcak G, Kotov R, Bromet EJ, Carlson GA, & Klein DN (2018). Hurricane Sandy exposure alters the development of neural reactivity to negative stimuli in children. Child Development, 89(2), 339–348. 10.1111/cdev.12691 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kessler RC, Avenevoli S, & Merikangas KR (2001). Mood disorders in children and adolescents: An epidemiologic perspective. Biological Psychiatry, 49(12), 1002–1014. 10.1016/s0006-3223(01)01129-5 [DOI] [PubMed] [Google Scholar]
  22. Krigolson OE (2018). Event-related brain potentials and the study of reward processing: Methodological considerations. International Journal of Psychophysiology, 132, 175–183. 10.1016/j.ijpsycho.2017.11.007 [DOI] [PubMed] [Google Scholar]
  23. Kujawa A, Carroll A, Mumper E, Mukherjee D, Kessel EM, Olino T, Hajcak G, & Klein DN (2018). A longitudinal examination of event-related potentials sensitive to monetary reward and loss feedback from late childhood to middle adolescence. International Journal of Psychophysiology, 132(Pt B), 323–330. 10.1016/j.ijpsycho.2017.11.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kujawa A, Klein DN, & Hajcak G (2012). Electrocortical reactivity to emotional images and faces in middle childhood to early adolescence. Developmental Cognitive Neuroscience, 2(4), 458–467. 10.1016/j.dcn.2012.03.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kujawa A, Klein DN, & Proudfit GH (2013). Two-year stability of the late positive potential across middle childhood and adolescence. Biological Psychology, 94(2), 290–296. 10.1016/j.biopsycho.2013.07.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kujawa A, Proudfit GH, Kessel EM, Dyson M, Olino T, & Klein DN (2015). Neural reactivity to monetary rewards and losses in childhood: Longitudinal and concurrent associations with observed and self-reported positive emotionality. Biological Psychology, 104, 41–47. 10.1016/j.biopsycho.2014.11.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kujawa A, Weinberg A, Hajcak G, & Klein DN (2013). Differentiating event-related potential components sensitive to emotion in middle childhood: Evidence from temporal-spatial PCA. Developmental Psychobiology, 55(5), 539–550. 10.1002/dev.21058 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Luck SJ (2014). An introduction to the event-related potential technique, second edition. The MIT Press. [Google Scholar]
  29. Luck SJ, Stewart AX, Simmons AM, & Rhemtulla M (2021). Standardized measurement error: A universal metric of data quality for averaged event-related potentials. Psychophysiology, 58(6), e13793. 10.1111/psyp.13793 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. MacNamara A, & Hajcak G (2010). Distinct electrocortical and behavioral evidence for increased attention to threat in generalized anxiety disorder. Depresssion and Anxiety, 27(3), 234–243. 10.1002/da.20679 [DOI] [PubMed] [Google Scholar]
  31. Makeig S, Bell AJ, Jung TP, & Sejnowski TJ (1996). Independent component analysis of electroencephalographic data. In Touretzky D, Mozer M, & Hasselmo M(Eds.), Advances in neural information processing systems (Vol. 8, pp. 145–151). The MIT Press. [Google Scholar]
  32. Mitchell DGV, Richell RA, Leonard A, & Blair RJR (2006). Emotion at the expense of cognition: Psychopathic individuals outperform controls on an operant response task. Journal of Abnormal Psychology, 115(3), 559–566. 10.1037/0021-843X.115.3.559 [DOI] [PubMed] [Google Scholar]
  33. Pegg S, Dickey L, Mumper E, Kessel E, Klein DN, & Kujawa A (2019). Stability and change in emotional processing across development: A 6-year longitudinal investigation using event-related potentials. Psychophysiology, 56(11), e13438. 10.1111/psyp.13438 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Plöchl M, Ossandón JP, & König P (2012). Combining EEG and eye tracking: Identification, characterization, and correction of eye movement artifacts in electroencephalographic data. Frontiers in Human Neuroscience, 6, 278. 10.3389/fnhum.2012.00278 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Proudfit GH (2015). The reward positivity: From basic research on reward to a biomarker for depression. Psychophysiology, 52(4), 449–459. 10.1111/psyp.12370 [DOI] [PubMed] [Google Scholar]
  36. Simmons AM, & Luck SJ (2020). Protocol for reducing COVID-19 transmission risk in EEG research. Research Square, rs.3.pex-974. 10.21203/rs.3.pex-974/v2 [DOI] [Google Scholar]
  37. Speed BC, Nelson BD, Perlman G, Klein DN, Kotov R, & Hajcak G (2015). Personality and emotional processing: A relationship between extraversion and the late positive potential in adolescence. Psychophysiology, 52(8), 1039–1047. 10.1111/psyp.12436 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stange JP, MacNamara A, Barnas O, Kennedy AE, Hajcak G, Phan KL, & Klumpp H (2017). Neural markers of attention to aversive pictures predict response to cognitive behavioral therapy in anxiety and depression. Biological Psychology, 123, 269–277. 10.1016/j.biopsycho.2016.10.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Wallstrom GL, Kass RE, Miller A, Cohn JF, & Fox NA (2004). Automatic correction of ocular artifacts in the EEG: A comparison of regression-based and component-based methods. International Journal of Psychophysiology, 53(2), 105–119. 10.1016/j.ijpsycho.2004.03.007 [DOI] [PubMed] [Google Scholar]
  40. Weinberg A, & Hajcak G (2010). Beyond good and evil: The time-course of neural activity elicited by specific picture subtypes. Emotion, 10(6), 767–782. 10.1037/a0020242 [DOI] [PubMed] [Google Scholar]
  41. Woodman GF (2010). A brief introduction to the use of event-related potentials in studies of perception and attention. Attention, Perception & Psychophysics, 72(8), 2031–2046. 10.3758/APP.72.8.2031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Zhang G, & Luck SJ (2023). Variations in ERP data quality across paradigms, participants, and scoring procedures. Psychophysiology, 60(7), e14264. 10.1111/psyp.14264 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are openly available in OSF at https://osf.io/796ma/?view_only=588654e549bd4aeea71892ac78d42db2.

RESOURCES