Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 May 1.
Published in final edited form as: Clin Neurophysiol. 2023 Feb 16;149:1–8. doi: 10.1016/j.clinph.2023.01.021

The role of single trial variability in event related potentials in children with attention deficit hyperactivity disorder

Anne B Arnett 1,2,*, Gaelle Gourdet 1, Virginia Peisch 1, Katherine Spaulding 1, Erica Ferrara 1, Vivian Li 1
PMCID: PMC10101921  NIHMSID: NIHMS1875251  PMID: 36841009

Abstract

Objective.

Children with attention deficit hyperactivity disorder (ADHD) show attenuated mean P3 component amplitudes compared to typically developing (TD) children. This finding may be the result of individual differences in P3 amplitudes, P3 latencies, and/or greater single trial variability (STV) in amplitude or latency, suggesting neural “noise.”

Methods.

Event related potentials (ERPs) from 75 children with ADHD and 29 TD children were recorded with electroencephalography (EEG). Caregivers provided ratings on child ADHD symptoms. Single-trial ERP amplitudes and latencies were extracted from the P3 component time window during a visual oddball task. Additionally, we computed individual-centered and trial-centered P3 amplitudes to account for inter-individual and inter-trial variability in the timing of the P3 peak.

Results.

In line with prior research, greater ADHD symptom severity was associated with reduced mean P3 amplitude. This correlation was no longer significant after correcting for inter-trial differences in P3 latency. In contrast, greater ADHD symptom severity was associated with reduced STV in P3 amplitude.

Conclusions.

Our results suggest that attenuated average P3 amplitude in ADHD samples is due to a consistent reduction in strength of the neurophysiological signal at the single trial level, as well as increased inter-trial variability in the timing of P3 peak amplitudes. The traditional method of extracting P3 amplitudes based on a single time window for all trials may not adequately capture variability in P3 latencies associated with ADHD.

Significance.

Inter- and intra-individual differences in brain signatures should be considered in models of neurobiological differences in neurodevelopmental samples.

Keywords: attention deficit hyperactivity disorder, EEG, evoked potentials, etiology, single trial variability

1. Introduction

Attention deficit hyperactivity disorder (ADHD) is a highly prevalent neurodevelopmental disorder associated with behavioral and cognitive dysregulation (Barbaresi et al., 2013, Nigg et al., 2020). In addition to symptoms of inattention, hyperactivity, and impulsivity, children with ADHD often present with variable performances on laboratory and real-world measures of executive functioning and cognitive control, including inhibitory control, processing speed, working memory, and planning (Arnett et al., 2021, Fair et al., 2012, Willcutt et al., 2005). These cognitive-behavioral deficits are believed to originate from reduced integrity of multiple large-scale neural networks, particularly frontal-striatal, frontal-parietal, somatosensory and default mode networks (Castellanos and Proal, 2012). Event related potential (ERP) methods have been used to identify cortical indices that are associated with differences in behavioral measures of executive functioning. One of the most robust findings pertains to the amplitude of the P300 (P3) ERP component. P3 amplitudes are attenuated in ADHD samples, specifically (Barry et al., 2003; Johnstone et al., 2013; Moavero et al., 2020), as well as among people with externalizing disorders more broadly (Carlson et al., 2007). In the current study, we focus on a P3 component that is associated with novelty detection and attention switching (Friedman et al., 2001) and thus highly relevant to cognitive-behavioral symptoms of ADHD (Keage et al., 2006).

Despite the sizeable body of literature documenting reduced P3 amplitude in ADHD, few studies have aimed to elucidate underlying mechanisms. A better understanding of the neurobiological origins for this finding could refine etiological models for ADHD and inform treatment approaches. Atypical or highly variable timing of the P3 onset or peak (i.e., latency) are potential sources of reduced amplitude. The most common method for extracting average ERP amplitude values is to identify a single time window for all individuals for each component of interest. The ERP amplitude for each individual is then computed within that time window without considering potential individual differences or variability across trials. Thus, if an individual’s peak latency is atypically fast or slow either across all trials (inter-individual differences) or on a subset of trials (single trial variability; STV), then the full morphology of the P3 component may not be captured by the a priori determined window in every trial. Mathematically, this would result in a decreased average ERP amplitude for that individual (Shucard et al., 2016).

STV in ERP components is sometimes referred to as the “neural noise theory” and has gained popularity in its attempt to draw associations between neural variability (“noise”) and behavioral symptoms (Butler et al., 2017). Thus far, the neural noise theory has primarily been applied to research on autism spectrum disorder (Milne, 2011; Kovarski et al., 2019), but has relevance for other neurodevelopmental disorders, such as ADHD. Consistent with this hypothesis, Burwell et al. (2014) reported that reduced temporal consistency (i.e., greater STV) of neural oscillations explained reduced P3 amplitude during an ERP task among adults with externalizing disorders. STV in both amplitude and latency of ERP components has received increasing attention as potential markers of individual neurocognitive differences (Mohr and Nagel, 2010, Ouyang et al., 2017).

Among ADHD samples, research on inter-individual variability versus inter-trial variability in ERPs is limited, but generally supports a model wherein increased STV explains reduced group-level grand average amplitudes. In their study of adults with ADHD, Gonen-Yaacovi and colleagues (2016) reported that the ADHD group had greater STV in early sensory ERP components compared to a sample of healthy controls; although they did not measure later ERP components, such as the P3. Lazarro and colleagues (1997) found greater STV in mean amplitude, peak amplitude, and latency of a P3 component among unmedicated adolescents with ADHD relative to controls. Another early study found that STV in auditory evoked potentials differentiated children with ADHD from controls, and the greatest STV in both visual and auditory ERPs was characteristic of ADHD children who showed a positive response to amphetamine treatment (Buchsbaum and Wender, 1973). STV in steady state visual evoked potentials differentiated children with ADHD from healthy controls as well as children with dyslexia (Victor et al., 1993). Myatchin and colleagues (2012) found greater STV in ERP amplitudes among a small sample of children with ADHD compared to controls, but no difference in grand average ERP amplitude across groups. However, this study did not focus on the P3 or another related component. Rather, the authors analyzed the full waveform epoch, which could help account for the non-significant group differences in grand average ERP. In contrast, Hilger and colleagues (2020) found that greater ADHD symptom severity was associated with decreased P3 amplitude STV; however, this was a population sample and did not include a group of children with clinical ADHD (Hilger et al., 2020). Altogether, the overwhelming evidence suggests that children with ADHD have both reduced grand average P3 amplitudes and increased STV, when compared to TD samples. Yet, no study to date has directly examined associations between these two metrics in a well characterized pediatric sample.

The current study seeks to fill this knowledge gap by investigating the etiology of attenuated mean P3 amplitude to novel visual stimuli in a sample of school-aged children with ADHD and a comparison cohort of TD children. Specifically, we examine whether 1) reduced average P3 is associated with increased ADHD severity in our pediatric sample, and whether that effect reflects 2) individual differences in latency among youth with higher ADHD symptoms, and/or 3) STV in the timing and amplitude of the P3 response, i.e., neural noise. We model ADHD symptom severity as a continuous variable in order to maximize variance in the main study variables and increase statistical power.

2. Methods

2.1. Participants

Children between the ages of 7 and 11 years were enrolled in the study as either ADHD (n=107) or TD (n=34) participants. Exclusion criteria included a diagnosis of autism spectrum disorder (ASD), known genetic syndrome, intellectual disability or global developmental delay, perinatal complications (e.g., <32 weeks gestation, intracranial hemorrhage, prolonged NICU stay), prenatal exposure to substances, history of seizures, or colorblindness. ADHD diagnoses were confirmed by the supervising clinical psychologist on the project using standardized clinical interview, parent ratings, and/or direct observation. By caregiver report, TD participants had fewer than three DSM-5 ADHD symptoms and did not have an immediate family history of ADHD. Following enrollment, 22 children were excluded from analyses due to lack of confirmed ADHD or TD status (n=13), suspicion or later diagnosis of ASD (n=4), IQ < 70 (n=2), identification of epileptiform activity during testing (n=1), or failure to abstain from medications prior to the research visit (n=2). Fifteen additional participants were excluded from analyses due to achieving an accuracy of < 60% on the ERP experiment. The final sample (n=75 ADHD; n=29 control) is described in Table 1.

Table 1.

Sample Description

ADHD TD difference
N 75 29 -
Age in years M (SD) 9.28 (1.34) 8.83 (1.23) p = .106
Female 29% 34% p = .785
Non-White 39% 31% p = .619
Task Accuracy M (SD) 80% (9%) 85% (7%) p = .003
Full-Scale IQ M (SD) 109 (10) 118 (11) p = .001
Average ADHD Symptom Severity M (SD) 1.42 (0.62) −0.66 (0.59) p < .001
Individual P3 Trials M (SD) 25.72 (3.52) 27.31 (1.87) p = .004
Trial-Centered P3 Trials M (SD) 20.49 (3.21) 21.66 (2.72) p = .068
Open in a new tab

Note: ADHD = attention deficit hyperactivity disorder; TD = typically developing. M = mean; SD = standard deviation. Average ADHD symptom severity was calculated as the mean severity of parent ratings of 18 ADHD symptoms on the SWAN, with a range of −3 to 3; higher scores indicate greater ADHD symptom severity. Individual P3 and Trial-Centered P3 Trials = the number of trials included in the individual P3 amplitude or trial-centered P3 variables, respectively. Difference tests were independent samples t-tests for continuous variables (age, accuracy, novel trials, IQ); and chi-square tests for categorical variables (female, non-white). Bolded p values indicate statistically significant group differences at p < .05.

2.2. Procedures

Participants and a caregiver visited a university laboratory for a three-hour research visit that included electroencephalography (EEG) recording, neuropsychological testing, and caregiver surveys. Legal guardians provided written consent and participating children provided written assent, consistent with the university institutional review board (IRB) protocol. Prior to the visit, and if applicable, participants abstained from taking stimulant medications for at least 48 hours, and non-stimulant medications for a duration determined and approved by their prescribing physician.

2.2.1. EEG Recording

EEG was recorded with Magstim-EGI equipment. Specifically, we used a high-density, 128-channel Hydrocel geodesic sensor net and Netstation Acquisition software version 4.5.6, integrated with a 400-series high impedance amplifier (Magstim-EGI; Plymouth, MN). The vertex electrode was used as a reference. Signals were analog filtered (0.1 Hz high-pass, 100 Hz elliptical low-pass), amplified and digitized with a sampling rate of 1000 Hz. Electrode impedances were reduced to below 50 kOhms at the start of the session and saline solution was used to re-wet electrodes throughout to maximize signal-to-noise ratio. The precise timing of visual stimulus presentation was recorded using a Cedrus Stimtracker (Cedrus Corporation, San Pedro, CA).

2.2.2. EEG Processing

EEG data were processed after acquisition using EEGLAB 15 and ERPLab v8.0 functions in MATLAB R2018b. We excluded eye electrodes and 14 rim channels from analyses. Next, data were downsampled to 250 Hz and bandpass filtered at 0.3–80 Hz. Electrical line noise was removed in the 55–65 Hz range using the Cleanline plugin for EEGLAB. Bad channels were detected using a threshold of 3 standard deviations from the normalized spectra. Rejected channels were then interpolated into the dataset prior to average referencing. Experimenter-marked 1.5 second segments in which the child was overly active (e.g., talking, looking around the room) or not attending to the task were removed. Next, artifactual portions of the continuous EEG were identified and removed using a standard spectrum thresholding algorithm. Lastly, extended independent component analysis (ICA) was run with primary component analysis dimension reduction to identify and subsequently remove artifactual independent components (e.g., eye blinks, muscle artifact, cardiac signal). After ERP segmentation, data were lowpass filtered at 40 Hz and baseline corrected with a 300ms pre-stimulus baseline. Only novel trials in which the participant did not impulsively press a button were included. The number of novel trials removed due to impulsive responses did not differ between ADHD (M = 1.53, SD = 2.15) and TD (M = 1.50, SD = 1.92) groups (t[97] = 0.07, p = .943. After data cleaning, the control group had more retained trials than the ADHD group on average (p = .004).

2.3. Measures

2.3.1. ERP Task

Following five minutes of baseline EEG, participants completed an 8-minute visual ERP experiment. This was a dual-task paradigm adapted from (Jonkman et al., 2000) in which stimuli for a 1-back visual working memory task were presented alternately with stimuli for a passive visual oddball task. Participants were presented with 1-back target visual stimuli (red, orange, green, or blue rectangles) one at a time, against a black background. Children were given instructions for a 1-back working memory test, wherein they were told to respond to two consecutive rectangles of the same color with a right-hand button press and two consecutive rectangles of differing colors with a left-hand button press (Figure 1). A response was required for all 1-back target stimuli and there was equal probability of right- and left-hand responses. After each 1-back stimulus, an oddball-task stimulus was presented which did not require a response. Oddball task stimuli were presented in random order and included standard (white bracket on a black background; 60%), deviant (white bracket oriented in the opposite direction; 20%) and novel (white line drawings of animals and vehicles, each presented only once; 20%) stimuli. Participants were told not to respond to the oddball task stimuli. Thus, each trial included a target stimulus requiring a response, followed by an oddball task stimulus requiring no response. Participants were given up to three practice sets of 10 trials, followed by 140 test trials broken into three blocks. Stimulus duration was 300ms and the interstimulus interval varied randomly from 0.8–1.4 seconds.

Figure 1.

Figure 1.

Open in a new tab

Schematic of the event related potential (ERP) task adapted from Jonkman et al. (2000) and used in the current study. Visual stimuli alternated between those relevant to a 1-back task (colored rectangles) and those relevant to a passive visual oddball task. In the current study, responses to the novel stimuli in the oddball task were analyzed.

2.3.2. ERP Components

Consistent with prior literature (Debener et al., 2005, Ratsma et al., 2001) and based on visual examination of waveforms and topographical plots (see Figure 2), the novel stimulus P3 was maximal at the midline parietal electrode cluster (Pz, P3, P4) during a time window of 280 – 450 ms. An example of single trial waveforms for ADHD and TD participants is shown in Supplementary Figure 1. Peak latency within the P3 time window for each trial was identified consistent with methods described by Lazzaro and colleagues (1997). Specifically, the peak latency for each trial was defined as the timing of the highest amplitude point for which there were at least four consecutive increasing amplitude sample points (i.e., 16 ms) immediately prior and at least four consecutive decreasing amplitude sample points immediately following. This ensured a true peak was selected rather than a local maximum. Eighty percent of the peaks identified were the highest point in the P3 time window. If a peak was not identified for a particular trial, that trial was not included in the trial-centered analyses. The number of trials with peaks did not differ between ADHD and TD groups (see Trial-Centered P3 Trials in Table 1).

Figure 2.

Figure 2.

Open in a new tab

Top: scalp topography during the standard P3 component window (280–450ms) for ADHD (left) and TD (right) groups. Bottom: ERP waveforms across standard, deviant and novel stimuli for ADHD (left) and TD (right) groups. Shading indicates the P3 component window from which peak and grand average metrics were extracted. ADHD = participants with attention deficit hyperactivity disorder. TD = typically developing participants. ERP = event related potential.

2.3.3. Individual- and Trial-Level P3 Variables

Trial-level and individual-level P3 component variables of interest were computed. First, individual P3 amplitude was calculated using traditional methods, by extracting the mean amplitude in the P3 time window (280–450 ms), averaged across all trials for each individual, regardless of whether a trial-level peak was identified (Figure 3a). Second, individual-centered P3 amplitude was defined as the mean amplitude in a 170 ms window surrounding the individual’s average peak P3 latency (Figure 3b). In this way, individual-centered P3 amplitude accounted for the possibility that an individual’s average P3 component morphology was not captured by the standard 280–450 ms time window. Third, a trial-centered P3 amplitude was calculated that corrected for inter-trial variability in the timing and morphology of the P3 component (Figure 3c). Specifically, P3 amplitudes were extracted from each trial for which a peak was identified, in the 170 ms window around the peak P3 latency at each trial. These amplitudes were averaged across all trials to derive an individual’s trial-centered P3 amplitude. Individual P3 amplitude STV, trial-centered P3 amplitude STV, and P3 latency STV were each calculated as an individual’s standard deviation of that respective variable across trials. Lastly, we calculated the number of trials for which a P3 peak was not identified, and the number of trials for which the traditional P3 amplitude was negative, as additional candidate measures of cortical noise.

Figure 3.

Figure 3.

Open in a new tab

P3 amplitudes were calculated in three ways. A) Mean P3 amplitude extracted from a single time window across all trials, consistent with traditional methods. B) Individual P3 amplitude was extracted as the average amplitude across trials within a time window defined by the individual’s average P3 latency. C) Trial-centered P3 amplitudes were extracted for each trial, from a time window defined by the peak P3 amplitude on that trial.

2.3.4. Intellectual Ability

Full-scale IQ was derived from the two-subtest (Vocabulary and Matrix Reasoning) administration of the Wechsler Abbreviated Scale of Intelligence, Second Edition (WASI-II).

2.3.5. ADHD Symptoms

ADHD symptoms were assessed using the Strengths and Weaknesses of ADHD Symptoms and Normal Behavior (SWAN) Scale, which asked caregivers to compare their child’s behavior to that of other same-aged children on a 7-point Likert scale (Lakes et al., 2012). Items 1–9 correspond to symptoms of inattention while items 10–18 correspond to symptoms of hyperactivity/impulsivity. Total severity scores were calculated by averaging the responses for all 18 items. Responses were re-scored on a scale of −3 to 3, such that higher ratings reflected greater ADHD symptom severity.

2.4. Data Analyses

Analyses were conducted in R Studio version 2022.02.3. Variables were examined for normality and confirmed to have skew and kurtosis values < |3|. One data point with an outlying trial-centered P3 amplitude greater than 50 microvolts was removed from the data. All analyses included the full sample of children. ADHD symptom severity, rather than diagnosis, was included as a continuous variable to capture the full symptom variance, maximize statistical power, and acknowledge that ADHD is best conceptualized as a dimensional construct (Arnett et al., 2013, Elton et al., 2014). First, we tested bivariate Pearson correlations to determine whether age, IQ, or number of ERP trials were associated with any of the ERP variables and should therefore be included as covariates in our model. None of the potential covariates related to any of the P3 metrics, although the association between IQ and individual P3 STV approached significance (r = −0.17, p = .057). All other correlation coefficients were not statistically significant (absolute r’s < 0.17, p’s > .086). Thus, we used two-tailed bivariate Pearson correlations to test our hypotheses, without inclusion of covariates. P-values were adjusted for multiple comparisons within each hypothesis using the Hochberg method (Hochberg, 1988).

3. Results

3.1. ADHD Symptom Severity and Individual P3 Amplitude

First, we examined whether greater ADHD symptom severity was indeed associated with attenuated mean P3 amplitudes, calculated in the traditional way. Consistent with prior research, lower individual P3 amplitude was correlated with increased ADHD symptom severity (r = −0.23, p = .019).

3.2. Latency Hypothesis

Next, we tested the hypothesis that higher ADHD symptoms were associated with extreme low or high latency, indicating that the P3 morphology would not be well described by traditional means of extracting the P3 component. In support of this hypothesis, we found a negative correlation between ADHD symptom severity and average P3 latency (r = −0.24, p = .012). Next, we examined whether accounting for individual differences in average P3 latency, by centering the P3 window on the individual’s average latency, would eliminate the association between mean P3 amplitude and ADHD symptom severity. Contrary to this hypothesis, the correlation between individual-centered P3 amplitude and ADHD symptom severity remained statistically significant in the expected direction: r = −0.27, p = .012. Altogether, these results indicate that although children with high ADHD symptoms appear to have shorter average P3 latencies, this does not account for P3 amplitude attenuation.

3.3. Neural Noise

3.3.1. Latency.

To test the neural noise hypothesis, we first examined bivariate correlations between peak P3 latency STV and both ADHD symptom severity and mean P3 amplitude (Table 2). Peak P3 latency STV was associated with increased ADHD symptoms (r = 0.25, p = .033), but not with mean P3 amplitude (r = 0.02, p = .821). To further probe this hypothesis, we next examined whether the association between ADHD symptom severity and mean P3 amplitude could be accounted for by correcting for STV in peak P3 latency, i.e. by centering the P3 window around the P3 latency at each trial. Consistent with the neural noise theory, the correlation between ADHD symptom severity and trial-centered P3 amplitude was not statistically significant after correction for multiple comparisons (r = −0.20, p = .096). However, as can be seen in Figure 4, the effect size was similar to that for the correlation between ADHD symptom severity and mean P3 amplitude.

Table 2.

Correlations Between P3 Amplitude Indices of Neural Noise, ADHD Symptom Severity and Traditional P3 Amplitude

Individual P3 Amplitude STV Trial-Centered P3 Amplitude STV Number of Trials with Negative P3 Amplitude Number of Trials without P3 Peak
ADHD Symptom Severity −0.23, p = .048 −0.24, p = .036 0.24, p = .038 0.00, p = .999
Mean P3 Amplitude 0.21, p = .150 0.27, p = .033 −0.82, p < .001 −0.06, p = 1.00
Open in a new tab

Note: values are two-tailed Pearson correlations; p-values are adjusted for four comparisons using the Hochberg (1998) method.

Figure 4.

Figure 4.

Open in a new tab

Scatter plot depicting the linear associations between ADHD symptom severity and P3 amplitude computed with traditional methods (solid black line) versus trial-centered P3 amplitude (dashed gray line).

3.3.2. Amplitude.

Next, we examined bivariate associations between four indices of neural noise reflected by variability in P3 amplitude (individual P3 amplitude STV; trial-centered P3 amplitude STV; number of negative amplitude trials; and number of trials without an identifiable P3 peak) and both ADHD symptom severity and mean P3 amplitude (see Table 2). Contrary to the neural noise hypothesis, higher ADHD symptom severity was associated with reduced STV in individual P3 amplitude, calculated using both traditional methods (i.e., a single window for all trials) and the trial-centered approach. Mean P3 amplitude was positively correlated with trial-centered P3 amplitude STV, but not P3 amplitude STV calculated using the traditional method of a single time window. Additionally, children with higher ADHD symptom severity had more trials on which the average amplitude in the standard P3 window was negative, and this was related to lower average P3 amplitude overall. Finally, there was no association between the number of trials for which a clear P3 peak could not be identified and ADHD symptoms or mean P3 amplitude. Altogether, greater ADHD symptom severity and lower P3 amplitude were associated with consistently reduced P3 amplitudes across trials.

4. Discussion

In the current study, we attempted to replicate and explain one of the most consistent findings in the pediatric ADHD literature: attenuated P3 amplitudes. Our review of the extant literature suggested a paucity of research examining mechanisms underlying this robust group difference. We considered two potential explanations for this effect: inter-individual variability in peak P3 latency and neural noise. Across our combined ADHD and TD sample, we did not find support for the atypical latency hypothesis and found partial support for the neural noise hypothesis.

Increased ADHD symptom severity was related to reduced (i.e., faster) average peak P3 latency, suggesting that the P3 morphology for these children may not be captured in a traditional P3 window. However, the association between reduced P3 amplitude and greater ADHD symptom severity remained statistically significant even after we corrected for differences in mean P3 latency by adapting the P3 window to each individual. Thus, although children with higher ADHD symptoms in our sample had shorter peak P3 latencies, their mean P3 morphologies appeared to be captured by the standard P3 window. It should be noted that our sample predominantly comprised children with ADHD; thus, the selection of the standard P3 window likely favored this group and our results might not be replicated in a sample that includes more TD children.

We tested the neural noise hypothesis by examining STV in peak P3 latency as well as STV in P3 amplitudes. Consistent with the neural noise theory, greater peak P3 latency STV was correlated with higher ADHD symptom severity, and the association between ADHD symptom severity and P3 amplitude was no longer significant when peak P3 latency STV was accounted for, i.e., by using trial-centered P3 amplitudes. Interestingly, peak P3 latency STV was not correlated with mean P3 amplitude. This combination of findings suggests that the effect of latency STV may be specific to children with high ADHD symptoms.

We did not find support for the hypothesis that variability or neural noise pertaining to P3 amplitudes explains the association between ADHD severity and attenuated mean P3. Instead, greater ADHD symptom severity and reduced traditional P3 amplitude were both associated with consistently reduced P3 amplitudes across trials, including more trials on which the P3 amplitude was negative. Thus, our trial level data support the extant research on grand average waveforms, wherein higher ADHD symptoms are related to attenuated neurophysiological response during the P3 phase.

Our data were not able to clarify the neurobiological etiology of reduced peak P3 amplitude or P3 latency STV, and several explanations should be considered. ERPs are generated by synchronized post-synaptic potentials in a group of neurons (Luck, 2006). The strength of the ERP signal at the scalp is influenced by a number of factors, including the orientation and depth of the neural generators, the thickness of the skull, the number of neural generators that are activated, and their synchronicity (Cardenas et al., 2005). Thus, the etiology of individual differences in ERP amplitudes is likely multifactorial. Reduced P3 amplitudes and greater variability in peak P3 latency in association with greater ADHD severity may reflect reduced synchronization of neural generators, potentially due to atypical organization of the cortex, and/or reduced availability of post-synaptic excitatory and inhibitory neurotransmitters (Luck, 2006). This would be consistent with findings that catecholamine-enhancing psychostimulant medications increase the amplitude of P3 ERP components (Peisch et al., 2021). Alternately, this finding could reflect greater engagement of localized, rather than global, neural networks among children with ADHD (Arnett et al., 2022, Ostlund et al., 2021). To fully understand the neurobiological etiology of reduced neurophysiological signal in children with ADHD, future investigations should consider integrating EEG with brain imaging methods that have high spatial resolution, such as functional magnetic resonance imaging (fMRI). Experimental designs that manipulate catecholamine levels (e.g., testing children on and off stimulant medications) could also expand what is known about neurobiological mechanisms.

Prior research has indicated that greater ERP latency STV occurs in the context of increased cognitive load, particularly during working memory tasks (Shucard et al., 2016). We used a passive oddball task integrated with a 1-back visual working memory task. Children with ADHD often struggle on tests of working memory (Kofler et al., 2020), and in the current study they performed significantly worse on the ERP task. However, the current analyses only included ERP trials on which participants provided a correct response. Thus, our findings may indicate that on these correct trials, children with higher ADHD symptoms or increased grand average P3 amplitudes were engaging more cognitive resources, both of which were associated with greater peak P3 latency STV.

Our results support the continued use of traditional ERP grand average amplitude calculations in neurodevelopmental research. However, our findings also highlight the need to consider heterogeneity in the neurocognitive mechanisms associated with individual presentations of ADHD. First, it is important to acknowledge that although we conceptualized STV as a measure of intra-individual variability, it also reflects inter-individual differences. In other words, SDs of P3 peak amplitude and latency also reflect person-level differences in the degree of inter-trial consistency in the P3 response. Second, a potential source of heterogeneity in the neurobiology of ADHD could be the degree to which low amplitude versus inconsistent P3 latency relate to an individual’s ADHD symptoms. These distinct associations could inform etiological models of neurobiological mechanisms as well as development of individualized treatment guidelines for children with ADHD and their families.

There are inherent limitations to single trial ERP analyses. Specifically, the morphology of an ERP component is most evident when multiple trials are averaged together. At the individual trial level, a clear P3 component may not be easily identified. We attempted to account for this by using an algorithm to identify P3-like peak amplitudes within the P3 time window. However, even with these methodological approaches, it cannot be guaranteed that the neurophysiological activity we captured reflects the same brain process during each trial. Given the novelty of our findings, replication in independent samples is required. Future research could deploy different tasks, including those that are less cognitively demanding and require additional aspects of executive functioning. In our sample, a number of participants (n=15) were not able to complete the task with at least 60% accuracy; exclusion of these participants may have biased our results. In future work, we plan to investigate whether individual differences in average and STV of ERP responses are linked to variability in cognitive and behavioral performance among children. Given that psychostimulant medications have been shown to normalize P3 amplitude as well as reduce ADHD symptoms (Peisch et al., 2021), it will be important to investigate whether stimulants modify individual P3 amplitude or consistency of P3 latency. While we suspect that increased availability of catecholamines through psychostimulant treatment could serve to both strengthen and increase reliability of the timing of the P3 signal, if the effect was specific to one metric or another, this would provide critical clues into the neurobiology of ADHD as well as the mechanisms of action for psychostimulants.

In sum, we aimed to advance understanding of the cortical processes associated with symptoms of ADHD. The P3 ERP component is generated by neural networks that support attention orienting, novelty detection, and attention switching. Our findings suggest that those networks are not only weaker among children with ADHD, but that they may also be subject to greater moment-by-moment variability in neural excitation and inhibition. At the symptom level, neural variability may help account for vulnerabilities in consistent and age-appropriate attention and behavior regulation.

Supplementary Material

1

Highlights.

  • Children with high ADHD symptom severity consistently have reduced amplitude of the grand average P3 event related potential.

  • This association is largely explained by reduced peak amplitude in the P3 time window across trials.

  • Single trial variability in peak P3 latency across trials explains additional variance in the association between ADHD symptoms and grand average P3.

Acknowledgments

This research was funded by grants to A.B.A. from the National Institute of Mental Health (K99MH116064-01A1 and R00MH116064-01A1). We would like to thank the IDDRC funded by NIH P50 HD105351. The funders were not involved in the study design, collection, analysis or interpretation of data; nor were they involved in the writing of the manuscript or the decision to submit the article for publication.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Disclosures

The authors have no competing interests to declare.

References

  1. Arnett AB, Fearey M, Peisch V, Levin AR. Absence of Dynamic Neural Oscillatory Response to Environmental Conditions Marks Childhood Attention Deficit Hyperactivity Disorder. J Child Psychol Psychiatry 2022; doi: 10.1111/jcpp.13645 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arnett AB, McGrath LM, Flaherty B, Pennington B, Willcutt E. Heritability and Clinical Characteristics of Neuropsychological Profiles of ADHD. J Atten Disord 2021; 26(11): 1422–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arnett AB, Pennington BF, Friend A, Willcutt EG, Byrne B, Samuelsson S, et al. The SWAN captures variance at the negative and positive ends of the ADHD symptom dimension. J Atten Disord 2013;17(2):152–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986; 51: 1173–1182. [DOI] [PubMed] [Google Scholar]
  5. Barbaresi WJ, Colligan RC, Weaver AL, Voigt RG, Killian JM, Katusic SK. Mortality, ADHD, and psychosocial adversity in adults with childhood ADHD: a prospective study. Pediatrics 2013;131(4):637–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Barry RJ, Johnstone SJ, Clarke AR. A review of electrophysiology in attention-deficit/hyperactivity disorder: II. Event-related potentials. Clin Neurophysiol 2003;114(2):184–98. [DOI] [PubMed] [Google Scholar]
  7. Buchsbaum M, Wender P. Average evoked responses in normal and minimally brain dysfunctioned children treated with amphetamine: A preliminary report. Arch Gen Psychiatry 1973;29(6):764–70. [DOI] [PubMed] [Google Scholar]
  8. Burwell SJ, Malone SM, Bernat EM, Iacono WG. Does electroencephalogram phase variability account for reduced P3 brain potential in externalizing disorders? Clin Neurophysiol 2014;125(10):2007–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Butler JS, Molholm S, Andrade GN, Foxe JJ. An Examination of the Neural Unreliability Thesis of Autism. Cereb Cortex. 2017;27(1):185–200. doi: 10.1093/cercor/bhw375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cardenas VA, Chao LL, Blumenfeld R, Song E, Meyerhoff DJ, Weiner MW, et al. Using automated morphometry to detect associations between ERP latency and structural brain MRI in normal adults. Hum Brain Mapp 2005;25(3):317–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Carlson SR, McLarnon ME, Iacono WG. P300 amplitude, externalizing psychopathology, and earlier-versus later-onset substance-use disorder. J Abnorm Psychol 2007;116(3):565. [DOI] [PubMed] [Google Scholar]
  12. Castellanos FX, Proal E. Large-scale brain systems in ADHD: beyond the prefrontal-striatal model. Trends Cogn Sci 2012;16(1):17–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Debener S, Makeig S, Delorme A, Engel AK. What is novel in the novelty oddball paradigm? Functional significance of the novelty P3 event-related potential as revealed by independent component analysis. Brain Res Cogn Brain Res 2005;22(3):309–21. [DOI] [PubMed] [Google Scholar]
  14. Elton A, Alcauter S, Gao W. Network connectivity abnormality profile supports a categorical‐dimensional hybrid model of ADHD. Hum Brain Mapp 2014;35(9):4531–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fair DA, Bathula D, Nikolas MA, Nigg JT. Distinct neuropsychological subgroups in typically developing youth inform heterogeneity in children with ADHD. Proc Natl Acad Sci U S A 2012;109(17):6769–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Friedman D, Cycowicz YM, Gaeta H. The novelty P3: an event-related brain potential (ERP) sign of the brain’s evaluation of novelty. Neurosci Biobehav Rev 2001;25(4):355–73. [DOI] [PubMed] [Google Scholar]
  17. Gonen-Yaacovi G, Arazi A, Shahar N, Karmon A, Haar S, Meiran N, et al. Increased ongoing neural variability in ADHD. Cortex 2016;81:50–63. [DOI] [PubMed] [Google Scholar]
  18. Hilger K, Sassenhagen J, Kühnhausen J, Reuter M, Schwarz U, Gawrilow C, et al. Neurophysiological markers of ADHD symptoms in typically-developing children. Sci Rep 2020;10(1):1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hochberg Y A sharper Bonferroni procedure for multiple tests of significance. Biometrika. 1988. 75(4):800–2. [Google Scholar]
  20. Johnstone SJ, Barry RJ, Clarke AR. Ten years on: a follow-up review of ERP research in attention-deficit/hyperactivity disorder. Clin Neurophysiol 2013;124(4):644–657. [DOI] [PubMed] [Google Scholar]
  21. Jonkman LM, Kemner C, Verbaten MN, Van Engeland H, Camfferman G, Buitelaar JK, Koelega HS. Attentional capacity, a probe ERP study: differences between children with attention-deficit hyperactivity disorder and normal control children and effects of methylphenidate. J. Psychophysiol 2000;37(3):334–346. [PubMed] [Google Scholar]
  22. Keage HA, Clark CR, Hermens DF, Kohn MR, Clarke S, Williams LM, et al. Distractibility in AD/HD predominantly inattentive and combined subtypes: the P3a ERP component, heart rate and performance. J Integr Neurosci 2006;5(01):139–58. [DOI] [PubMed] [Google Scholar]
  23. Kofler MJ., Singh LJ, Soto EF, Chan ESM, Miller CE, Harmon SL, Spiegel JA. Working memory and short-term memory deficits in ADHD: A bifactor modeling approach. Neuropsychology 2020; 34(6): 686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kovarski K, Malvy J, Khanna RK, Arsène S, Batty M, Latinus M. Reduced visual evoked potential amplitude in autism spectrum disorder, a variability effect? Transl Psychiatry 2019;9:341. 10.1038/s41398-019-0672-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lakes KD, Swanson JM, Riggs M. The reliability and validity of the English and Spanish strengths and weaknesses of ADHD and normal behavior rating scales in a preschool sample: Continuum measures of hyperactivity and inattention. J Atten Disord 2012;16(6):510–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lazzaro I, Anderson J, Gordon E, Clarke S, Leong J, Meares R. Single trial variability within the P300 (250–500 ms) processing window in adolescents with attention deficit hyperactivity disorder. Psychiatry Res 1997;73(1–2):91–101. [DOI] [PubMed] [Google Scholar]
  27. Luck SJ. The Operation of Attention--Millisecond by Millisecond--Over the First Half Second. In: Öğmen H, Breitmeyer BB, editors. The first half second: The microgenesis and temporal dynamics of unconscious and conscious visual processes: MIT Press; 2006. p. 187–206. [Google Scholar]
  28. Milne E Increased intra-participant variability in children with autistic spectrum disorders: evidence from single-trial analysis of evoked EEG. Front Psychol. 2011;2:51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Moavero R, Marciano S, Pro S, De Stefano D, Vigevano F, Curatolo P, Valeriani M. Event-Related Potentials in ADHD Associated With Tuberous Sclerosis Complex: A Possible Biomarker of Symptoms Severity? Front Neurol 2020;11:546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mohr PN, Nagel IE. Variability in brain activity as an individual difference measure in neuroscience? J Neurosci 2010;30(23):7755–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Myatchin I, Lemiere J, Danckaerts M, Lagae L. Within-subject variability during spatial working memory in children with ADHD: an event-related potentials study. Eur Child Adolesc Psychiatry 2012;21(4):199–210. [DOI] [PubMed] [Google Scholar]
  32. Nigg JT, Karalunas SL, Feczko E, Fair DA. Toward a Revised Nosology for ADHD Heterogeneity. Biol Psychiatry Cogn Neurosci Neuroimaging 2020; 5(8): 726–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ostlund BD, Alperin BR, Drew T, Karalunas SL. Behavioral and cognitive correlates of the aperiodic (1/f-like) exponent of the EEG power spectrum in adolescents with and without ADHD. Dev Cogn Neurosci 2021;48:100931. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ouyang G, Hildebrandt A, Sommer W, Zhou C. Exploiting the intra-subject latency variability from single-trial event-related potentials in the P3 time range: a review and comparative evaluation of methods. Neurosci Biobehav Rev 2017;75:1–21. [DOI] [PubMed] [Google Scholar]
  35. Peisch V, Rutter T, Wilkinson CL, Arnett AB. Sensory processing and P300 event-related potential correlates of stimulant response in children with attention-deficit/hyperactivity disorder: A critical review. Clin Neurophysiol 2021; 132(4):953–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ratsma JE, van der Stelt O, Schoffelmeer AN, Westerveld A, Boudewijn Gunning W. P3 event-related potential, dopamine D2 receptor A1 allele, and sensation-seeking in adult children of alcoholics. Alcohol Clin Exp Res 2001;25(7):960–7. [PubMed] [Google Scholar]
  37. Shucard DW, Covey TJ, Shucard JL. Single trial variability of event-related brain potentials as an index of neural efficiency during working memory. International Conference on Augmented Cognition: Springer; 2016. p. 273–83. [Google Scholar]
  38. Victor JD, Conte MM, Burton L, Nass RD. Visual evoked potentials in dyslexics and normals: failure to find a difference in transient or steady-state responses. Vis Neurosci 1993;10(5):939–46. [DOI] [PubMed] [Google Scholar]
  39. Willcutt EG, Doyle AE, Nigg JT, Faraone SV, Pennington BF. Validity of the executive function theory of attention-deficit/hyperactivity disorder: a meta-analytic review. Biol Psychiatry 2005;57(11):1336–46. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS1875251-supplement-1.pdf (428.6KB, pdf)

RESOURCES