Attention-deficit/hyperactivity disorder (ADHD) is characterized by developmentally inappropriate levels of inattention and hyperactivity/impulsivity (American Psychiatric Association, 2013). ADHD is an early-onset, chronic, and prevalent disorder that is associated with a host of negative outcomes including behavior problems and risk-taking as well as functional impairment in virtually all major life domains such as the academic, occupational, and the social domains (Bunford, Brandt, et al., Reference Bunford, Brandt, Golden, Dykstra, Suhr and Owens2015; Bunford, Evans, et al., Reference Bunford, Evans and Langberg2018; Bunford et al., Reference Bunford, Kujawa, Dyson, Olino and Klein2021; Faraone et al., Reference Faraone, Biederman and Mick2006; Kieling et al., Reference Kieling, Kieling, Rohde, Frick, Moffitt, Nigg, Tannock and Castellanos2010; Le et al., Reference Le, Hodgkins, Postma, Kahle, Sikirica, Setyawan and Doshi2014). Accordingly, identification of predictors of prognosis and response to treatment is key. Evidence indicates that although clinical and demographic indices are weak predictors (Ball et al., Reference Ball, Stein and Paulus2014), electrophysiological measures are promising (Bunford, Kujawa, Fitzgerald, et al., Reference Bunford, Kujawa, Fitzgerald, Swain, Hanna, Koschmann, Simpson, Connolly, Monk and Phan2017; Hámori et al., Reference Hámori, File, Fiáth, Pászthy, Réthelyi, Ulbert and Bunford2023; Kujawa et al., Reference Kujawa, Weinberg, Bunford, Fitzgerald, Hanna, Monk, Kennedy, Klumpp, Hajcak, Phan, Swain, Monk, Hajcak and Phan2016).
In terms of resting electroencephalogram (EEG), data show ADHD is most consistently and reliably associated with atypical theta activity (Arns et al., Reference Arns, Conners and Kraemer2013; Barry et al., Reference Barry, Clarke and Johnstone2003); a large body of work indicates enhanced absolute (Arns et al., Reference Arns, Conners and Kraemer2013; Bresnahan et al., Reference Bresnahan, Anderson and Barry1999; Chabot & Serfontein, Reference Chabot and Serfontein1996, Clarke et al., Reference Clarke, Barry, McCarthy and Selikowitz1998, Reference Clarke, Barry, McCarthy and Selikowitz2001c; Defrance et al., Reference Defrance, Smith, Schweitzer, Ginsberg and Sands1996) and relative (Boutros et al., Reference Boutros, Fraenkel and Feingold2005) theta power in individuals with ADHD, and findings also show enhanced event-related frontal-midline theta synchronization (Guo et al., Reference Guo, Luo, Li, Chang, Sun and Song2020) as well as adult-like posterior theta lateralization (Guo et al., Reference Guo, Luo, Li, Chang, Sun and Song2020). Results also suggest elevated theta/alpha (TAR) (Clarke et al., Reference Clarke, Barry, McCarthy and Selikowitz2001c, Reference Clarke, Barry, McCarthy and Selikowitz2002) and elevated theta/beta (TBR) (Arns et al., Reference Arns, Conners and Kraemer2013; Monastra et al., Reference Monastra, Lubar, Linden, VanDeusen, Green, Wing, Phillips and Fenger1999, Reference Monastra, Lubar and Linden2001) ratios, though findings with proportion variables are less well-replicated (Koehler et al., Reference Koehler, Lauer, Schreppel, Jacob, Heine, Boreatti-Hümmer, Fallgatter and Herrmann2009).
Elevated theta in ADHD has been central to theories of cortical hypoarousal, developmental deviation, and a maturational lag (Barry et al., Reference Barry, Clarke and Johnstone2003). On one hand, the frontal focus of elevated theta indicates that cortical hypoarousal may disrupt anterior attentional network functioning. On the other hand, while theta activity decreases with age in neurotypical individuals, it is elevated in ADHD. This suggests the complementary interpretation that elevated theta reflects a maturational delay, especially of frontal, executive, and inhibitory networks (Barry et al., Reference Barry, Clarke and Johnstone2003). Of import, theta power decrease from pre- to post-treatment correspond to cognitive improvement in youth with ADHD (Lubar et al., Reference Lubar, Swartwood, Swartwood and O’Donnell1995), suggesting that theta activity is an index of processes that are potential mechanisms of response to treatment.
Theta oscillations (measured, typically, at 4–8 Hz) are generated across many brain structures, including the prefrontal-midline region of the cortex and the anatomically defined limbic system (Knyazev, Reference Knyazev2007). Across species, theta activity is associated with voluntary motor movement, learning, and memory (Knyazev, Reference Knyazev2007). In humans, theta activity has been linked primarily to affective processing, including affect regulation and discrimination of emotional stimuli (Nishitani, Reference Nishitani2003) and to attention (Mann et al., Reference Mann, Lubar, Zimmerman, Miller and Muenchen1992), including affectively motivated attention (Knyazev, Reference Knyazev2007; W. Zhang et al., Reference Zhang, Li, Liu, Duan, Wang and Shen2013). Although whether theta activity is an index of affect regulation or attention was unclear (Knyazev, Reference Knyazev2007) – and in some regards, arguably, remains to be (Loo et al., Reference Loo, McGough, McCracken and Smalley2018) – conceptually, frontal-midline theta corresponds to affect regulation (i.e., cognitive control; Cavanagh & Frank, Reference Cavanagh and Frank2014) whereas parietal theta corresponds to attention (i.e., affectively motivated attention; Aftanas et al., Reference Aftanas, Reva and Makhnev2008).
Alpha oscillations (measured, typically, at 7–13 Hz) are hypothesized to be generated by the thalamus and layer V pyramidal cells of the occipital cortex and have been observed to propagate from anterosuperior to posteroinferior cortex (Halgren et al., Reference Halgren, Ulbert, Bastuji, Fabó, Eross, Rey, Devinsky, Doyle, Mak-McCully, Halgren, Wittner, Chauvel, Heit, Eskandar, Mandell and Cash2019). Alpha oscillations have also been implicated in ADHD (Barry et al., Reference Barry, Clarke and Johnstone2003) and have also been linked to both inhibition and attention (Knyazev, Reference Knyazev2007; Saalmann et al., Reference Saalmann, Pinsk, Wang, Li and Kastner2012). Findings across a number of studies indicate attenuated relative alpha power in individuals with ADHD (Clarke et al., Reference Clarke, Barry, McCarthy and Selikowitz2001b, Reference Clarke, Barry, McCarthy and Selikowitz2002), including in posterior (Callaway et al., Reference Callaway, Halliday and Naylor1983) as well as frontal and central (Clarke et al., Reference Clarke, Barry, McCarthy and Selikowitz1998) sites. Observations of a task-related decrease in alpha power have been interpreted as alpha being a mechanism of inhibiting functions and input that are conflicting with or unnecessary to a given task (Knyazev, Reference Knyazev2007) and the involvement of the lower alpha band in attentional processes has been confirmed in neuroimaging (Laufs et al., Reference Laufs, Krakow, Sterzer, Eger, Beyerle, Salek-Haddadi and Kleinschmidt2003) and electrophysiological (Babiloni et al., Reference Babiloni, Miniussi, Babiloni, Carducci, Cincotti, Del Percio, Sirello, Fracassi, Nobre and Rossini2004; Dockree et al., Reference Dockree, Kelly, Roche, Hogan, Reilly and Robertson2004; Sauseng et al., Reference Sauseng, Klimesch, Stadler, Schabus, Doppelmayr, Hanslmayr, Gruber and Birbaumer2005) research.
ADHD is associated with deficits in both affect regulation and attention; atypicalities in affective processing, such as enhanced negative and positive dispositional affectivity (Bunford et al., Reference Bunford, Kujawa, Dyson, Olino and Klein2021; Lahey, Reference Lahey2009; Martel & Nigg, Reference Martel and Nigg2006), enhanced aggression (including physical, relational, and verbal (Bunford, Evans, & Wymbs, Reference Bunford, Evans and Wymbs2015)), and deficits in regulating negative and positive affect (Bunford, Evans, & Wymbs, Reference Bunford, Evans and Wymbs2015) are consistently documented features of the disorder. Indirectly, in boys, ADHD has also been linked to atypical connectivity of functional neuronal networks in frontal and occipital lobes (Nasab et al., Reference Nasab, Panahi, Ghassemi, Jafari, Rajagopal, Ghosh and Perc2022) and altered phase synchronization stability (Ansarinasab et al., Reference Ansarinasab, Panahi, Ghassemi, Ghosh and Jafari2022) during facial affect processing as well as atypical resting-state functional connectivity involving the affective network (Gao et al., Reference Gao, Shuai, Bu, Hu, Tang, Zhang, Li, Hu, Lu, Gong and Huang2019) and regions (Yu et al., Reference Yu, Liu, Chen, Cao, Zepf, Ji, Wu, An, Wang, Qian, Zang, Sun and Wang2020). Difficulties in directing and sustaining attention are among the most marked ADHD symptoms (American Psychiatric Association, 2013).
Taken together, atypicalities in both theta and alpha activity have been linked to ADHD and both theta and alpha have been implicated in inhibitory, regulatory processes (in case of theta, in affect regulation, specifically) and in attention. Further, both affective dysregulation and inattention are key correlates of the ADHD phenotype. Yet, the complex interrelations across these variables, i.e., whether there are dissociable relations of theta and alpha with affect regulation and attention and how these associations are related to ADHD, is unknown.
For EEG markers to be truly valuable as biomarkers, their precise meaning (or affective-cognitive correlates) needs to be clearly understood. Relative to a host of research on the association between theta activity and attention in ADHD (Barry et al., Reference Barry, Clarke and Johnstone2003; Guo et al., Reference Guo, Luo, Li, Chang, Sun and Song2020), to our knowledge, no data are available on the association between theta activity and affective processing in ADHD and no studies have attempted to decipher the interrelations of theta and alpha activity with affective processing and attention in ADHD. Specifically, whether (1) frontal-midline and parietal theta and/ or alpha activity are differentially associated with affective processing and attention remains to be empirically examined.
Although empirical focus has historically been on assessment of between-group differences, e.g., comparing youth with ADHD to controls on a hypothesized correlate, such as theta power or synchronization, more recent data show that even in the absence of between-group differences in such correlates, the correlate or variable of interest may differentially relate to outcomes in youth with relative to youth without ADHD; the same characteristic may confer risk for a negative outcome in those with ADHD but not in those without. For example, in one research, emotional lability was positively associated with behavioral difficulties in children with ADHD but not in children without ADHD (Rosen et al., Reference Rosen, Walerius, Fogleman and Factor2015). In another research, although TBR did not differ between youth with and without ADHD, it was positively associated with inattention in youth with ADHD (D. W. Zhang et al., Reference Zhang, Li, Wu, Zhao, Song, Liu, Qian, Wang, Roodenrys, Johnstone, De Blasio and Sun2019). As a further example, across multiple studies, no differences in theta power during sleep were observed between children with and without ADHD (Scarpelli et al., Reference Scarpelli, Gorgoni, D’atri, Reda and De Gennaro2019). However, theta power during sleep is negatively associated with memory performance in youth with ADHD but positively associated with such performance in children and adults without ADHD (Scarpelli et al., Reference Scarpelli, Gorgoni, D’atri, Reda and De Gennaro2019). These data underscore the importance of moving beyond simple between-group comparisons to assessing differences between groups in the relations between hypothesized predictors and outcomes. As such, whether (1) frontal-midline and parietal theta and/ or alpha activity and affective processing and attention differ between adolescents with and without ADHD is a pertinent yet to-date unaddressed question.
Age and sex are both key third variables to consider in the context of the associations between theta and alpha activity with affective processing, attention, and ADHD; Regarding electrophysiological variables, EEG patterns are characterized by age-related differences (Segalowitz et al., Reference Segalowitz, Santesso and Jetha2010), including, as noted, a decrease in theta activity (Barry et al., Reference Barry, Clarke and Johnstone2003) and an increase in theta synchronization during adolescence (Uhlhaas & Singer, Reference Uhlhaas and Singer2011). In research where differences in the EEG across sexes in typically developing children were examined, findings were mixed, indicating no sex differences (Gasser et al., Reference Gasser, Verleger, Bächer and Sroka1988), differences suggestive of a maturational lag in girls (Harmony et al., Reference Harmony, Marosi, Díaz de León, Becker and Fernández1990), and differences showing males exhibit less absolute and relative theta and more relative alpha power (Clarke et al., Reference Clarke, Barry, McCarthy and Selikowitz2001a). Also, age- and sex-related differences are also evident in affective processing (Bunford, Evans, et al., Reference Bunford, Evans and Langberg2018) and in attention/inattention (Langberg et al., Reference Langberg, Epstein, Altaye, Molina, Arnold and Vitiello2008). Boys are considerably more – up to three times as frequently in community samples and nine times as frequently in clinical samples – often diagnosed with ADHD than girls (Skogli et al., Reference Skogli, Teicher, Andersen, Hovik and Øie2013). Differences in diagnostic rates may be a function, in part, of differences in disorder expression across boys and girls (Biederman et al., Reference Biederman, Mick, Faraone, Braaten, Doyle, Spencer, Wilens, Frazier and Johnson2002; Gaub & Carlson, Reference Gaub and Carlson1997; Quinn, Reference Quinn2008). Findings on the differences in prevalence and/ or manifestation of ADHD are coupled with results showing that certain biological (i.e., genetic) characteristics differentially confer risk for outcomes and symptoms in boys relative to girls with ADHD (Nymberg et al., Reference Nymberg, Jia, Lubbe, Ruggeri, Desrivieres, Barker, Büchel, Fauth-Buehler, Cattrell, Conrod, Flor, Gallinat, Garavan, Heinz, Ittermann, Lawrence, Mann, Nees, Salatino-Oliveira and Schumann2013) and, with regard to electrophysiological variables specifically, EEG absolute and relative power differences have been observed to be greater between males with and without ADHD than between females with and without ADHD, where total power, absolute alpha and beta, and relative delta and alpha were sensitive to differences between males but only TAR and TBR were sensitive to differences between females (Clarke et al., Reference Clarke, Barry, McCarthy and Selikowitz2001b).
Multi-method and –informant measurement, especially of complex and heterogeneous characteristics such as affective processing and regulation, have long been recommended for research on child and adolescent psychopathology (De Reyes & Kazdin, Reference De Reyes and Kazdin2005; Dirks et al., Reference Dirks, De Los Reyes, Briggs-gowan, Cella and Wakschlag2012; Mash & Hunsley, Reference Mash and Hunsley2005). In case of affective processing, self-report is key, given the largely internal and subjective nature of emotions (Bunford, Evans, & Wymbs, Reference Bunford, Evans and Wymbs2015). However, as children and adolescents with ADHD are often unreliable reporters of their behavior and functioning, augmenting self-report with observer-report (e.g., parent-report) as well as a physiological index is advantageous for capturing the different, multi-faceted aspects of the phenomenon. A physiological, event-related potential (ERP) measure of affective processing is the late positive potential (LPP), a sustained positivity in the ERP following presentation of affectively or motivationally relevant stimuli (Cuthbert et al., Reference Cuthbert, Schupp, Bradley, Birbaumer and Lang2000). The LPP reflects sustained attention toward or elaborative processing of affectively salient information as well as activation of the brain’s motivational systems (Cuthbert et al., Reference Cuthbert, Schupp, Bradley, Birbaumer and Lang2000; Hajcak et al., Reference Hajcak, Weinberg, MacNamara, Foti, Luck and Kappenman2011; Schupp et al., Reference Schupp, Flaisch, Stockburger and Junghöfer2006). Evincing its reliability and validity, the LPP is reliable across development (middle childhood through adolescence) (Kujawa, Klein, et al., Reference Kujawa, Klein and Proudfit2013), is sensitive to differences between pleasant/unpleasant and neutral pictures and words as well as between more and less intense stimuli (Hajcak et al., Reference Hajcak, MacNamara and Olvet2010), and is associated with self-report measures of both anxiety (Wessing et al., Reference Wessing, Rehbein, Romer, Achtergarde, Dobel, Zwitserlood, Fürniss and Junghöfer2015) and affect regulation (Hajcak et al., Reference Hajcak, Weinberg, MacNamara, Foti, Luck and Kappenman2011). Further, attending to increasingly arousing or emotional aspects of stimuli is associated with an enhancement in the LPP, whereas application of emotion regulation strategies is associated with an attenuation in the LPP (Hajcak & Nieuwenhuis, Reference Hajcak and Nieuwenhuis2006).
Current study
Our goals were to address the identified gaps in knowledge and, in a carefully phenotyped sample of middle-late adolescents with and without ADHD matched for age and sex, assess whether resting state (RS) theta and alpha power or event-related theta and alpha synchronization during affect regulation (measured at frontal-midline, centroparietal and bilateral parietal regions) (1) differ between adolescents with and without ADHD; (2) are differentially associated with an ERP measure of affective processing, the LPP, and parent- and self-reported measures of affective processing and attention, given ADHD status and sex, and (3) are differentially lateralized, given ADHD status and sex. We expected that
(1) adolescents with ADHD would exhibit enhanced RS theta power, enhanced theta ERS during affect regulation, and attenuated RS alpha power, relative to adolescents without ADHD;
(2) in adolescents without ADHD, (i) greater (frontal-midline) RS theta power and greater theta ERS during affect regulation will be associated with greater LPP (Lapomarda et al., Reference Lapomarda, Valer, Job and Grecucci2022) and with lower parent- and self-reported affective dysregulation and self-reported affectivity (Lapomarda et al., Reference Lapomarda, Valer, Job and Grecucci2022) and (ii) greater RS alpha power will be associated with better attention and thus lower IA and with greater LPP and with lower parent- and self-reported affective dysregulation and self-reported affectivity (Wang et al., Reference Wang, Dong, Sun, Wang, Zhang, Xue, Ren, Lv, Yuan and Yu2022) but
in adolescents without ADHD, associations described in (i) and (ii) would be either in an opposite direction or weaker (Rosen et al., Reference Rosen, Walerius, Fogleman and Factor2015; Scarpelli et al., Reference Scarpelli, Gorgoni, D’atri, Reda and De Gennaro2019; D. W. Zhang et al., Reference Zhang, Li, Wu, Zhao, Song, Liu, Qian, Wang, Roodenrys, Johnstone, De Blasio and Sun2019); and
(3) adolescents with ADHD would exhibit “adult-like” theta lateralization but adolescents without ADHD would exhibit less or no lateralization (Guo et al., Reference Guo, Luo, Li, Chang, Sun and Song2020).
Method
Procedures
Data were collected in the context of a larger longitudinal project, the Budapest Longitudinal Study of ADHD and Externalizing Disorders (BLADS) study, aimed at identifying behavioral and biological protective and risk factors of behavior problems and functional impairments in adolescents exhibiting a range of ADHD symptoms but oversampled for ADHD. Data analyzed in the current study were obtained during the first (cognitive and clinical) and second (EEG) assessment sessions during baseline measurements.
Adolescents between the ages of 14–17 years were recruited mainly from public middle-, technical and vocational-, and high schools as well as two child and adolescent psychiatry clinics in Budapest, Hungary. In case of schools, research staff visited classrooms and presented on the opportunity to participate in a research program. In case of clinics, research staff distributed an e-mail and fliers with information on the research program. Exclusionary criteria were cognitive ability at or below the percentile rank corresponding to an FSIQ of 80 across administered indices; autism spectrum disorder (severity ≥ 2); neurological illness; and having visual impairment as defined by impaired vision < 50 cm, unless corrected by glasses or contact lenses.
Parents and participants provided written informed consent (and assent) and then participants underwent a series of tests, including assessment of cognitive ability and a structured clinical interview (to assess and establish all but ADHD and externalizing diagnoses), an EEG measurement, and completion of questionnaires across two sessions. Parents also completed a series of questionnaires using the Psytoolkit platform (Stoet, Reference Stoet2010, Reference Stoet2017) and the Qualtrics software, Version June 2020–March 2021 (Qualtrics, Provo, UT). This research was approved by the National Institute of Pharmacy and Nutrition (OGYÉI/17089-8/2019) and has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
ADHD diagnoses were determined using parent report on the ADHD Rating Scale-5 (ARS-5) (DuPaul et al., Reference DuPaul, Power, Anastopoulos and Reid2016). For an ADHD diagnosis, adolescents had to meet a total of five (in case of youth ≥ 17 years old) or six (in case of youth < 17 years old) (or more) of the Diagnostic and Statistical Manual of Mental Disorders (5th ed; DSM-5) (American Psychiatric Association, 2013) inattention or hyperactivity/impulsivity symptoms and exhibit impairment (i.e., rating of 2 = moderate impairment or 3 = severe impairment) in at least three areas of functioning (G. DuPaul, personal communication, July 19, 2021).
Participants
Participants included in the current study were 132 age- and gender-matched adolescents between the ages of 14–17 years (M age = 14.944 years; SD = .802), n = 66 met criteria for ADHD. The majority (95.46%) identified as White and 4.54% identified as part of an ethnic minority group in Hungary. Average cognitive ability was in the 63rd percentile (SD = 22.253), with estimated VCI percentile rank: M = 66.074, SD = 25.380, estimated PRI percentile rank M = 59.853, SD = 26.248. Participants were from an above-average socioeconomic background based on parental income (average family net income fell in the 5 001-700 000 HUF/month range, with average net income in Hungary being 289 000 HUF/month) (Központi Statisztikai Hivatal, 2021).
Of 66 adolescents with ADHD, n = 33 (50%) were medication-naïve at the time of assessments. Of those who currently used ADHD medication (n simulants = 12 (66%) and n nonstimulants = 6 (33%)), 9 took a ≥ 24-hr medication hiatus prior to (EEG) testing, 2 did not, and 6 did not indicate whether they took a hiatus.
Measures
Clinician administered measures
Wechsler Intelligence Scales. To estimate cognitive ability, abbreviated versions of the Wechsler Intelligence Scales were used. The Wechsler Intelligence Scale for Children, Fourth Edition (Wechsler, Reference Wechsler2003) for youth under the age of 17 and the Wechsler Adult Intelligence Scale – Fourth Edition (Wechsler, Reference Wechsler2008) for youth 17 and above. Two subtests were administered of the Perceptual Reasoning subscales (PR), Matrix Reasoning and Picture Concepts (WISC), and Matrix Reasoning and Visual Puzzles (WAIS). Two subtests were administered of the Verbal Comprehension subscales (VC), Similarities and Vocabulary (both WISC and WAIS). These subtests allow for calculation of a Perceptual Reasoning Index (PRI) and Verbal Comprehension Index (VCI) estimate and percentile ranks corresponding to these estimates were used as indices of cognitive ability.
Adolescent self-report measures
Positive and Negative Affect Schedule (PANAS; Watson, Clark, Tellegen, et al., Reference Watson, Clark, Tellegen, Tellegan, Tellegen and Tellegan1988 ). The PANAS is a 20-item self-report measure of state and/or trait positive and negative affect, comprised of two subscales, the positive affect (PA) subscale, reflecting the extent to which a person feels enthusiastic, active and alert, and a negative affect (NA) subscale, reflecting a general dimension of subjective distress and a variety of aversive mood states such as anger, contempt, disgust, fear, guilt, and nervousness. Respondents rate the extent to which they are experiencing each mood state “right now” (i.e., state version) or “during the past two weeks” (i.e., trait version) on a five-point Likert-type response format scale (1 – ‘very slightly or not at all’ to 5 – ‘very much’). Higher scores on the PA and NA subscales indicate greater positive and negative affect, respectively. Prior findings indicate that PANAS scales have good internal consistency (αs ranging from .86 to .90 for PA and from .84 to .87 for NA) and good convergent and discriminant associations with distress and psychopathology measures of the underlying affectivity factors (e.g., Beck Depression Inventory [BDI], Hopkins Symptom Checklist [HSCL], STAI) (Watson, Clark, & Tellegen, Reference Watson, Clark and Tellegen1988). The Hungarian translation also demonstrated acceptable psychometric properties, including good internal consistency (PA α = .82, NA α = .83 [alpha values are provided only to the second decimal in the source article]) (Gyollai et al., Reference Gyollai, Simor, Köteles, Demetrovics, Gyollai, Simor, Koteles, Demetrovics, Gyollai, Simor, Köteles and Demetrovics2011).
In the current sample, the PANAS-trait was administered internal consistency of the subscales was ≥acceptable, with Cronbach’s alpha values as follows: NA = .788; PA = .804. In the current study, data from the PA and NA subscales were analyzed.
Difficulties in Emotion Regulation Scale (DERS; Gratz & Roemer, Reference Gratz and Roemer2004 ). The DERS is a 36 item self-report measure of ED, comprised of six subscales, Nonacceptance of Emotional Responses (Nonacceptance, e.g., When I’m upset, I become angry with myself for feeling that way), Difficulties Engaging in Goal-Directed Behavior (Goals, e.g., When I’m upset, I have difficulty concentrating), Impulse Control Difficulties (Impulse, e.g., When I’m upset, I become out of control), Lack of Emotional Awareness (Awareness, e.g., When I’m upset, I acknowledge my emotions), Limited Access to Emotion Regulation Strategies (Strategies, e.g., When I’m upset, I believe that wallowing in it is all I can do), and Lack of Emotional Clarity (Clarity, e.g., I have difficulty making sense out of my feelings). Items are rated on a five-point Likert-type response format scale (1 – ‘Almost Never’ to 5 – ‘Almost Always’), with higher scores indicating greater difficulty with emotion regulation. Prior findings indicate the DERS has acceptable psychometric properties, including good internal consistency, good test–retest reliability, and adequate construct and predictive validity in multiple adolescent samples (Adrian et al., Reference Adrian, Zeman, Erdley, Lisa, Homan and Sim2009; Bunford, Evans, Becker, et al., Reference Bunford, Evans, Becker and Langberg2015; Bunford, Evans, et al., Reference Bunford, Evans and Langberg2018; Vasilev et al., Reference Vasilev, Crowell, Beauchaine, Mead and Gatzke-Kopp2009; Weinberg & Klonsky, Reference Weinberg and Klonsky2009). In addition, the DERS exhibited robust correlations with psychological problems reflecting ED (Weinberg & Klonsky, Reference Weinberg and Klonsky2009) and physiological measures of ED (Vasilev et al., Reference Vasilev, Crowell, Beauchaine, Mead and Gatzke-Kopp2009). The Hungarian translation also demonstrated acceptable psychometric properties, including good internal consistency (all αs > .70) as well as construct and convergent validity with the Zung Self-rated Depression Scale (Kökönyei et al., Reference Kökönyei, Urbán, Reinhardt, Józan and Demetrovics2014).
In the current sample, internal consistency of the subscales was ≥acceptable, with Cronbach’s alpha values as follows: Awareness = .806; Clarity = .816; Goals = .879; Impulse = .871; Non-Acceptance = .842; Strategies = .874; Total DERS: .928. In the current study, data from the Total score were analyzed.
Parent-report measures
ADHD Rating Scale-5 (ARS 5; DuPaul et al., Reference DuPaul, Power, Anastopoulos and Reid2016 ). The ARS-5 is a 30-item parent- and teacher-report measure of the past 6-month presence and severity of DSM-5 ADHD symptoms (nine inattentive symptom items and nine hyperactivity/impulsivity symptom items) and functional impairment across six domains: relationship with significant others (family members for the home version), relationship with peers, academic functioning, behavioral functioning, homework performance and self-esteem (2 × 6 impairment items, with one set corresponding to inattention and one to hyperactivity/impulsivity). Parents and teachers rate items on a four-point scale ranging in case of symptoms from 0 (never or rarely) to 4 (very often) and in case of impairment from 0 (“no problem”) to 3 (“severe problem”), with higher scores indicating more severe symptoms and impairment. The ARS-5 is comprised of two symptoms scales, Inattention and Hyperactivity-Impulsivity, and a Total Scale. The ARS-5 is suitable for ages 5-17 years, with separate forms for children (5–10 years) and adolescents (11–17 years) and age-appropriate and DSM-5 compatible descriptions of symptoms. In the current study, the adolescent home (i.e., parent-report) version was used. The ARS-5 has well-established reliability of the adolescent, home version (e.g., internal consistency and 6-week test-retest reliability) and validity (i.e., factor structure; concurrent validity and predictive validity and clinical utility) (DuPaul et al., Reference DuPaul, Power, Anastopoulos and Reid2016).
For purposes of the current study, the English version of the ARS-5 was translated into Hungarian following evidence-based guidelines: (1) the English version was translated into Hungarian by three independent translators; (2) these three translations were combined into a single “summary translated” measure by a fourth independent translator, reconciling all discrepancies across the three translations/ors; (3) the “summary” was back-translated into English by two additional independent translators and (4) the two back-translations were combined into a single “summary back-translated” measure by members of the research team, reconciling all discrepancies in a manner that the “summary back-translation” measure best matches the Hungarian “summary translated” measure. This “summary back-translated” questionnaire was sent to the original author(s) who provided the research team with feedback and ultimately approved the translated measure (G. DuPaul, personal communication, June 5, 2020). In the current sample, internal consistency of the ARS-5 was ≥acceptable, with a Cronbach’s alpha value of .916 for the Inattention (IA) subscale and .806 for the Hyperactivity/impulsivity (H/I) subscale. In the current study, data from both subscales and the Total score were used for diagnostic purposes and, to index attention/inattention in statistical analyses, data from the IA subscale were analyzed.
Difficulties in Emotion Regulation Scale Parent-Report (DERS-P; Bunford et al., Reference Bunford, Dawson, Evans, Ray, Langberg, Owens, DuPaul and Allan2020 ). The DERS-P is a 29 item parent-report measure of ED, comprised of four subscales, Attuned (e.g., “My child pays attention to how he/she feels”), Catastrophize (e.g., “When my child is upset, he/she believes that he/she will end up feeling very depressed”), Distracted (e.g., “When my child is upset, he/she has difficulty concentrating”), and Negative Secondary (“When my child is upset, he/she feels ashamed with him/herself for feeling that way”). Items are rated on a five-point Likert-type response format scale (1 – ‘Almost Never’ to 5 – ‘Almost Always’), with higher scores indicating greater difficulty with emotion regulation. Prior findings indicate the DERS-P has acceptable psychometric properties, including acceptable internal consistency, convergent, concurrent, and incremental validity in adolescents with and without ADHD (Bunford et al., Reference Bunford, Dawson, Evans, Ray, Langberg, Owens, DuPaul and Allan2020). As items are re-worded versions (from e.g., “When I am upset” to “When my child is upset”), a Hungarian translation of the DERS-P was created by taking items of the Hungarian translation of the DERS (Kökönyei et al., Reference Kökönyei, Urbán, Reinhardt, Józan and Demetrovics2014) and applying the same re-wording as was done for the English version.
In the current sample, internal consistency of the subscales was ≥acceptable, with Cronbach’s alpha values as follows: Attuned = .914; Catastrophize = .939; Distracted = .907; Negative Secondary = .901; Total DERS-P: .948. In the current study, data from the Total score were analyzed.
Experimental paradigm
Resting-state EEG measurement
A six-minute interval was used to record eyes-open RS theta and alpha power at the end of the EEG assessment session. Adolescents were instructed to look at a fixation cross for 2 × 3 min while their head was placed on a chin rest.
ERP measurement
In the larger study, the Doors task (Dunning & Hajcak, Reference Dunning and Hajcak2007; Foti & Hajcak, Reference Foti and Hajcak2009; Kujawa, Smith, et al., Reference Kujawa, Smith, Luhmann and Hajcak2013; Kujawa et al., Reference Kujawa, Proudfit and Klein2014, Reference Kujawa, Carroll, Mumper, Mukherjee, Kessel, Olino, Hajcak and Klein2018) was used to probe initial responsiveness to reward attainment and here, a portion of the task was conceptualized as probing affect regulation. The task consisted of 120 trials in total, presented in two blocks of 30 trials/condition. Participants were told that on each trial, they could either gain 100 or lose 50 (HUF). At the beginning of each trial, a fixation mark (+) appeared for 900 ms. Then, participants were presented with an image of two doors for 3000 ms and asked to choose one door by pressing the number 7 or 8 on the keypad (for the left and the right door, respectively). Finally, after a short delay (1100 ms with a jitter of ±50 ms), feedback was presented for 1500 ms on the screen. Gain was indicated by a green “↑” and loss was indicated by a red “↓”. The duration of the intertrial interval was 2000 ms with a jitter of ± 250 ms. In a single block, 30 gain and 30 loss trials were presented in random order.
To maximize effectiveness of the experimental paradigm, participants were told that the virtual money they accumulated can be exchanged for snacks (candy, chips, etc.), with more virtual money exchangeable for more desirable snack options (as ranked by the participant prior to the tasks).
We conceptualized that an emotion regulation process occurs after feedback stimuli are presented on the screen; whether an adolescent wins or loses, he/she has to regulate his/her affective response to such feedback as the next trial is upcoming.
EEG data acquisition and processing
EEG data were recorded and processed as described previously (Bunford et al., Reference Bunford, Hámori, Nemoda, Angyal, Fiáth, Sebők-Welker, Pászthy, Ulbert and Réthelyi2023; Hámori et al., Reference Hámori, Rádosi, Pászthy, Réthelyi, Ulbert, Fiáth and Bunford2022, Reference Hámori, File, Fiáth, Pászthy, Réthelyi, Ulbert and Bunford2023). Briefly, continuous EEG was acquired with a 64-channel BrainAmp DC system equipped with actiCAP active electrodes (Brain Products GmbH, Gilching, Germany) and digitized at a sampling rate of 1000 Hz and 16-bit resolution. Impedances were kept under 10 kΩ, and the FCz electrode was used as online reference. One electrooculogram electrode was placed below the left eye and another lateral to the outer canthus of the right eye.
The FieldTrip MATLAB toolbox was used for offline processing of the EEG data (Oostenveld et al., Reference Oostenveld, Fries, Maris and Schoffelen2011). Hamming-windowed sinc finite impulse response filters were used to filter the EEG (for details on filter parameters, see Supplement). Bad channels were removed (M±SD: 1.84 ± 1.57 channels, range: 0–8), then interpolated at a later stage of the preprocessing. The infomax independent component analysis (ICA) algorithm (Bell & Sejnowski, Reference Bell and Sejnowski1995) was applied, then components related to blinks, eye movements and transient or persistent noise artifacts (M ± SD: 3.45 ± .98 components, range: 1–8) were removed. The ICA-cleaned data was then high-pass filtered at 0.1 Hz and re-referenced to the average of TP9 and TP10 electrodes located at the left and right mastoids, respectively. The FCz electrode was included in the group of active electrodes. The continuous EEG was then epoched from −200 ms to 1000 ms around the stimuli (cue or target). Epochs containing high muscle activity or meeting the following criteria were automatically rejected: a voltage step of more than 50 μV between data points, a voltage difference of 300 μV within a trial, and a voltage difference of <.50 μV within 100 ms intervals (Bunford, Kujawa, Fitzgerald, et al., Reference Bunford, Kujawa, Fitzgerald, Swain, Hanna, Koschmann, Simpson, Connolly, Monk and Phan2017; Bunford, Kujawa, Swain, et al., Reference Bunford, Kujawa, Swain, Fitzgerald, Monk and Phan2017; Kujawa et al., Reference Kujawa, Proudfit, Kessel, Dyson, Olino and Klein2015, Reference Kujawa, Weinberg, Bunford, Fitzgerald, Hanna, Monk, Kennedy, Klumpp, Hajcak, Phan, Swain, Monk, Hajcak and Phan2016). Next, a final visual inspection was applied to remove remaining epochs with artifacts. Following, trials were baseline corrected to the first 200 ms of the epochs.
Resting-state spectral analyses. The EEGLAB Matlab toolbox was used for spectral analyses. Four regions of interest (ROI) were defined as follows: Frontal-midline theta (4–7 Hz) and alpha (8–12 Hz) were scored at Fz FC1, FC2, C1, Cz, and C2; parietal theta and alpha at CP4, CP6, P4, and P6 (right) and CP5, CP3, P5, and P3 (left); and a centroparietal, event-related region (i.e., using the same electrodes as for the LPP) theta and alpha at CP1, CPz, CP2, P1, Pz, P2, and POz.
Resting state EEG data were concatenated and 120 triggers were randomly placed into the recording. The recording was cut after each trigger into 1000–3000 ms-long epochs (same time window as for the LPP) and data in each epoch were convolved with a set of Morlet wavelets using 2–7 cycles increased linearly from 1–30 Hz in .5 Hz frequency steps, and then the mean of all epochs was used for decibel normalization. Time frequency/power spectrum values for channels of interest were obtained as results. Next, the mean of channels for each ROI was calculated, followed by calculation of the means of frequency bands, theta (4–7 Hz) and alpha (8–12 Hz) at a time window of 1600–2000 ms, for statistical analyses.
Event-related spectral analysis. Analyses for data obtained during the Doors task were comparable to those applied to data obtained during rest, except that each participant’s raw RS data were used for normalization purposes. Recording obtained during the Doors task was cut to 1000–3000 ms-long, post-feedback (i.e., after the chosen door either “wins” or “loses” as indicated by a green “↑” or a red “↓”) windows and the same spectral analyses as to the RS data, were applied. Using this method, desynchronization/synchronization could be defined as an average decrease/increase of power spectrum values from rest to during affect regulation. Obtained desynchronization and synchronization values are in dB.
ERP analyses. ERPs were averaged for each participant and for each condition, from the pre-processed EEG (i.e., final output of the pre-processing workflow) as follows. (1) The EEG was epoched from −200 ms to 3000 ms around feedback stimuli. (2) To ensure proper operation of our automatic artifact rejection algorithm, trials were low-pass filtered at 45 Hz (order: 294; transition width: 11.3 Hz). (3) Epochs containing high muscle activity (detected during step (3) of pre-processing) were removed. (4) An automatic artifact rejection method implemented in Matlab was used to reject additional trials containing artifacts. Artifact removal was based on the following criteria: (i) a voltage step of more than 50 μV between data points, (ii) a voltage difference of 300 μV within a trial, and (iii) a voltage difference of less than .50 μV within 100 ms intervals (Bunford, Kujawa, Fitzgerald, et al., Reference Bunford, Kujawa, Fitzgerald, Swain, Hanna, Koschmann, Simpson, Connolly, Monk and Phan2017; Bunford, Kujawa, Swain et al., Reference Bunford, Kujawa, Swain, Fitzgerald, Monk and Phan2017; Kujawa et al., Reference Kujawa, Proudfit, Kessel, Dyson, Olino and Klein2015, Reference Kujawa, Weinberg, Bunford, Fitzgerald, Hanna, Monk, Kennedy, Klumpp, Hajcak, Phan, Swain, Monk, Hajcak and Phan2016). (5) We performed a final visual evaluation to detect and remove remaining epochs with artifacts (6) Next, trials were baseline corrected using the 200 ms time interval prior to the stimulus onset. (7) After that, for each participant and for each condition, we computed the ERP averages, then these averages were low-pass filtered at 30 Hz (order: 442; transition width: 7.5 Hz). (8) As a final step, for each component, grand average ERP waveforms were calculated from individual ERP averages. As such, based on chosen electrodes and time windows, one ERP value per condition was calculated for each participant.
Given prior data with youth with anxiety disorders and other psychiatric symptoms indicating that the effects of psychopathology on the LPP were most apparent 1000–3000 ms after stimulus onset (Bunford, Kujawa, Swain, et al., Reference Bunford, Kujawa, Swain, Fitzgerald, Monk and Phan2017; Leutgeb et al., Reference Leutgeb, Schäfer, Köchel, Scharmüller and Schienle2010), we used the 1000–3000 ms post-feedback time window to index the LPP. Consistent with earlier adult (Stange et al., Reference Stange, MacNamara, Barnas, Kennedy, Hajcak, Phan and Klumpp2017) and child studies (Bunford, Kujawa, et al., Reference Bunford, Kujawa, Fitzgerald, Monk and Phan2018, ; Bunford, Kujawa, Swain, et al., Reference Bunford, Kujawa, Swain, Fitzgerald, Monk and Phan2017), where the LPP was scored at CP1, CP2, Cz, and Pz and at O1, O2, Oz, PO3, PO4, P3, P4, and Pz, electrodes for LPP scoring were: CP1, CPz, CP2, P1, Pz, P2, and POz. Analyses were conducted on the gain minus loss difference score.
Variables included in analyses were RS frontal-midline, centroperietal, and left and right parietal theta; RS frontal-midline, centroperietal, and left and right parietal alpha; event-related frontal-midline, centroperietal, and left and right parietal theta synchronization; and event-related frontal-midline, centroperietal, and left and right parietal alpha synchronization; and the LPP.
Analytic plan
Data were analyzed using IBM SPSS Statistics (version 28.0.1.0), through a two-step process involving analysis of variance (ANOVA) and bivariate correlations.
Normality was assessed using the Kolmogorov–Smirnov (with Lilliefors correction) test and visual inspection of diagnostic plots (density and Q–Q plots, histograms). Normality was violated in case of RS: centroparietal, left and right parietal theta; centroparietal, and left and right parietal theta synchronization; RS left parietal alpha; left and right parietal alpha synchronization; negative affectivity, self-report DERS, and IA.
For Aim 1, youth with and without ADHD were compared on key predictor variables, i.e., RS theta and alpha power and event-related theta and alpha synchronization (and also on outcome variables: affectivity and affect regulation as indexed by the PANAS NA, PANAS PA and the self- and parent-report DERS, and the LPP), using a one-way ANOVA in case of normally distributed and a Kruskal–Wallis test in case of non-normally distributed variables.
For Aim 2, bivariate correlations (Pearson’s r where both variables were normally distributed and Spearman’s rho where at least one was non-normally distributed) were computed for RS theta and alpha power and event-related theta and alpha synchronization variables and affectivity, affect regulation, and the LPP separately for groups with and without ADHD (with 95% confidence intervals [CIs] around the r values obtained with 1000 bootstrap resamples). Correlations with p < .05 in either group were chosen for further analysis, where selected r value-pairs (in the with and without ADHD groups) were transformed into z scores (Fisher’s r to z transformation) which were compared for statistical significance. Obtained p values were Benjamini–Hochberg corrected for false discovery rate (FDR).
For Aim 3, to compare left and right RS theta and alpha power and left and right event-related theta and alpha synchronization and thus estimate lateralization, the probability of superiority measure (Delaney & Vargha, Reference Delaney and Vargha2002; Ruscio, Reference Ruscio2008), denoted by Aw, was calculated (given non-normal distribution of three of four compared variables) for boys and girls with and without ADHD, using the following formula for Aw: Aw = [#(p > q) + .5#(p = q)/npnq] and for converting Aw to d-metric: dA = √2Φ − 1(Aw), where Φ is the normal cumulative distribution function.
Data analyzed in this study are available at [https://github.com/Bunfordlab/Takacs-et-al-Theta-and-alpha-activity-article].
Results
Event-related spectral perturbation of EEG power across boys and girls with and without ADHD at frontal-midline, centroparietal, and left and right parietal sites are depicted in Figures 1, 2, and 3. LPP scalp distributions and grand average waveforms are shown in Figure 4.
Figure 1. Event-related spectral perturbation of frontal-midline EEG power across boys and girls with and without ADHD. Figure depicts event-related synchronization of frontal-midline EEG power (scored in the 1000–3000 ms post-feedback time window, at Fz Fc1, Fc2, C1, Cz, and C2, with the average of theta band calculated in the 4–7 Hz frequency range (bottom rectangle) and the average of alpha calculated in the 8–12 Hz frequency range (top rectangle) both in the 1600–2000 ms time window) for (A) boys and (B) girls with ADHD as well as (C) boys and (D) girls without ADHD. Visual inspection indicates that in adolescents with ADHD, relative to boys (A), girls (B) exhibit a greater increase in theta and a greater decrease in alpha power in the time window of interest whereas in adolescents without ADHD, the opposite pattern is observable such that relative to girls (D), boys exhibit a greater increase in theta power in the time window of interest. Across groups, boys with ADHD exhibit the smallest increase in theta power in the time window of interest. Note. n = 66 adolescents with and 66 without ADHD, with n = 48 boys and 18 girls in each group.
Figure 2. Event-related spectral perturbation of centroparietal EEG power across boys and girls with and without ADHD. Figure depicts event-related synchronization of centroparietal EEG power (scored in the 1000–3000 ms post-feedback time window, at CP1, Cpz, Cp2, P1, Pz, P2, and Poz, with the average of theta band calculated in the 4–7 Hz frequency range (bottom rectangle) and the average of alpha calculated in the 8–12 Hz frequency range (top rectangle), both in the 1600–2000 ms time window) for (A) boys and (B) girls with ADHD as well as (C) boys and (D) girls without ADHD. Visual inspection indicates in adolescents with ADHD, relative to boys (A), girls (B) exhibit a greater increase in theta and a greater decrease in alpha power whereas in adolescents without ADHD, in both boys (C) and girls (D), there appears only a slight increase in theta power (that is nevertheless greater than that apparent in boys with ADHD) in the time window of interest. Note. n = 66 adolescents with and 66 without ADHD, with n = 48 boys and 18 girls in each group.
Figure 3. Event-related left and right parietal EEG power across boys and girls with and without ADHD. Figure depicts event-related left (left, across figures) and right (right, across figures) parietal EEG power synchronization (scored in the 1000–3000 ms post-feedback time window, at CP4, Cp6, P4, P6 (right) and CP5, CP3, P5, P3 (left), with the average of theta band calculated in the 4–7 Hz frequency range (bottom rectangle) and the average of alpha calculated in the 8–12 Hz frequency range (top rectangle), both in the 1600–2000 ms time window) for (A) boys and (B) girls with ADHD as well as (C) boys and (D) girls without ADHD. Visual inspection indicates, in the time window of interest, both boys (A) and girls (B) with ADHD exhibit a decrease in alpha power, with a greater decrease in the right hemisphere and boys with ADHD showing greatest decrease in the right hemisphere. Girls but not boys with ADHD also exhibit an increase in theta power. In adolescents without ADHD, boys (C) exhibit a comparable pattern as adolescents with ADHD with regard to a right parietal decrease in alpha power. They also exhibit, unlike boys with ADHD, a slight increase in theta power. Girls (D) do not show differences in parietal alpha power but do show such differences in parietal theta, driven more by the left side. Note. n = 66 adolescents with and 66 without ADHD, with n = 48 boys and 18 girls in each group.
Figure 4. Late positive potential (LPP) scalp distributions and grand average waveforms across boys and girls with and without ADHD. Figure depicts scalp distributions to gain, loss, and the gain minus loss difference in the 1000–3000 ms time window as well as ERPs (scored at CP1, Cpz, Cp2, P1, Pz, P2, and Poz) for (A) boys and (B) girls with ADHD as well as (C) boys and (D) girls without ADHD. Note. n = 66 adolescents with and 66 without ADHD, with n = 48 boys and 18 girls in each group.
Clinical, demographic, and behavioral performance descriptives across groups are reported in Table 1.
Table 1. Clinical, demographic, and behavioral performance descriptives across groups
Note. SES = socioeconomic status based on net family income/month coded as 1 = <50,000HUF, 2 = 50,001–99,000HUF, 3 = 100,000–150,000HUF, 4 = 150,001–200,000HUF, 5 = 200,001–300,000HUF, 6 = 300,001–500,000HUF, 7 = 700,001–800,000HUF, 8 = 800,001–1,000,000HUF, 9 = 1,000,000–1,200,000HUF, 10 = >1,200,000; PRI = perceptual reasoning index; VCI = verbal comprehension index; IA = parent-rated inattention symptoms on the ADHD Rating Scale–5; H/I = parent-rated hyperactivity/impulsivity symptoms on the ADHD Rating Scale–5; ODD = parent-rated oppositional defiant disorder symptoms on the Disruptive Behavior Disorders Rating Scale adapted to align with DSM-5; CD = parent-rated conduct disorder symptoms on the Disruptive Behavior Disorders Rating Scale adapted to align with DSM-5; Anxiety = Anxiety problems T score on the Youth Self Report; Depression = Depressive problems T score on the Youth Self Report; NA = negative affectivity on the Positive and Negative Affect Schedule; PA = positive affectivity on the Positive and Negative Affect Schedule; DERS = Difficulties in Emotion Regulation Scale scores; DERS-Parent = parent-rated Difficulties in Emotion Regulation Scale scores; behavioral performance = latency of response during the Doors task in ms.
Diff=between-groups comparisons using one-way ANOVAs for PA and DERS-Parent and Kruskal–Wallis tests for the remaining variables. In case of SES, families of girls without ADHD exhibited greater SES than boys with and boys without ADHD (psBonferroni-adjusted ≤ .043). In case of VCI, girls without ADHD exhibited higher VCI than boys with ADHD (p Bonferroni-adjusted = .020). In case of IA, H/I, and ODD symptoms, girls and boys with ADHD exhibited greater IA (psBonferroni-adjusted < .001), H/I (psBonferroni-adjusted < .001), and ODD (psBonferroni-adjusted ≤ .002) scores than girls and boys without ADHD. In case of NA and DERS, girls with ADHD exhibited greater NA (p Bonferroni-adjusted = .006) and DERS (p Bonferroni-adjusted = .003) than boys without ADHD. In case of DERS-Parent, girls and boys with ADHD exhibited greater scores (psTukey HSD < .035) than girls and boys without ADHD. None of the remaining between-groups comparisons were significant.
In the current sample, the LPP was marginally significantly associated with self-reported affective dysregulation total score (rho = .153, p = .081) but was not associated with the remaining affectivity or affect regulation variables: NA, PA, or parent-reported affective dysregulation total score (ps > .164). The LPP was not associated with inattention (p = .356).
Examination of the association between the LPP and affective dysregulation subscale scores revealed the LPP was associated with DERS Nonacceptance (r = .201, p = .021) and marginally significantly with DERS Clarity (r = .147, p = .092) and Strategies (r = .159, p = .068) (but not the remaining DERS subscales (ps > .154)).
Aim 1: Theta and alpha activity differences between youth with and without ADHD
Adolescents with and without ADHD did not differ on any power spectrum or ERP variables (ps > .086), except for RS frontal-midline alpha power, F(1, 126) = 6.514, p = .012; youth with ADHD exhibited lower alpha power (M = .256 [95%CI = −.071; .611], SD = 1.440 [95%CI = 1.152; 1.660]) than youth without ADHD (M = .943 [95%CI = .535; 1.351], SD = 1.603 [95%CI =1.248; 1.928]), with a medium effect size (Cohen’s D = −.451).
Groups also differed on NA: H(1) = 5.382, p = .020, DERS: H(1) = 8.160, p = .004, and DERS-P: F(1, 126) = 35.669, p < .001) scores, such that youth with ADHD exhibited greater NA (M = 20.877 [95%CI = 18.986; 22.656], SD = 7.324 [95%CI = 6.042; 8.470]), and greater affect dysregulation (DERS: M = 85.569 [95%CI = 80.473; 90.771], SD = 21.958 [95%CI = 18.048; 25.446]; DERS-P: (M = 86.538 [95%CI = 82.165; 90.614], SD = 17.583 [95%CI = 13.952; 20.316])) than youth without ADHD (NA: (M = 17.841 [95%CI = 16.457; 19.226], SD = 5.626 [95%CI = 4.478; 6.630]); DERS: (M = 74.936 [95%CI = 70.107; 80.192], SD = 20.803 [95%CI = 15.355; 25.083]); DERS-P: (M = 65.603 [95%CI = 60.263; 70.877], SD = 21.902 [95%CI = 17.996; 25.156]). These differences corresponded to a medium effect size for NA (Cohen’s d = .464) and self-report DERS (Cohen’s d = .497) and a large effect size for parent-report DERS (Cohen’s d = 1.054).
Aim 2: Associations between indices of theta and alpha activity and measures of affective processing and attention in youth with and without ADHD
For correlations between alpha and theta RS and synchronization variables with affective processing, age, and sex in the entire sample, see Tables S1 and S2.
Correlations of event-related right parietal theta synchronization and PA differed between adolescents with and without ADHD (p = .018) (Table 2). In youth with ADHD, event-related right parietal theta synchronization was positively associated with PA whereas in youth without ADHD, it was negatively associated with PA (at a trend level) (Table 2).
Table 2. Differences across groups in relations between EEG variables and rating scale measures of affective and motivational processing
Note. ADHD = attention-deficit/hyperactivity disorder; diff = r values were transformed into z scores (i.e., Fisher’s r to z transformation), which were compared for statistical significance. All correlations calculated using Pearson’s r, unless otherwise indicated. ctrl = controlling for. med = current ADHD medication status. †Correlations calculated using Spearman’s rho. *p values are Benjamini–Hochberg corrected for FDR within frequency band (i.e., alpha [7 comparisons] and theta [9 comparisons] variables).
Figure 1 indicated boys with ADHD exhibit different event-related frontal-midline theta synchronization during affect regulation than girls with and youth without ADHD. To quantify visual inspection, a two-way ANOVA was conducted and indicated no main effect by ADHD or sex (ps > .301) but a marginally significant ADHD*sex interaction effect (F(1, 128) = 2.923, p = .090) on event-related frontal-midline theta synchronization. Given small sample sizes especially for subsamples involving girls (ns = 18) follow-up calculations of effect size were conducted. Boys with (M = .715; SD = 1.445) and without (M = 1.418; SD = 1.635) ADHD differed from each other to a moderate extent (d corrected for small samples = .449). Girls with (M = .9982; SD = 1.898) and without (M = .578; SD = 1.278) ADHD differed to a small extent (d corrected for small samples = .247). The difference between boys and girls without ADHD was also moderate (d corrected for small samples = .564) but that between boys and girls with ADHD was negligible (d corrected for small samples = .156). (All other pairwise comparisons were small or negligible).
Given these findings, comparison of correlations across groups were repeated by controlling for sex.
Partial correlations of right parietal theta synchronization with PA differed between adolescents with and without ADHD (p = .036) (Table 2). In youth with ADHD, right parietal theta synchronization was positively associated with PA whereas in youth without ADHD, it was negatively associated with PA (at a trend level).
Partial correlations of RS right parietal alpha power differed between adolescents with and without ADHD (p = .049) (Table 2). In youth with ADHD, RS right parietal alpha power was negatively, whereas in youth without ADHD, it was not associated with the LPP (Table 2).
Aim 3: Theta and alpha lateralization across boys and girls with and without ADHD
Lateralization was apparent in RS parietal theta and alpha power in girls with ADHD, in RS parietal theta power in boys with ADHD, and in event-related parietal theta synchronization in girls without ADHD (Table 3). In girls with ADHD, power was greater in the right hemisphere, in boys with ADHD, power was greater in the left hemisphere, and in girls without ADHD, synchronization was greater in the left hemisphere (Table 3).
Table 3. Parietal theta and alpha lateralization across boys and girls with and without ADHD
Note. dm = d-metric, interpretable along the same cutpoints as Cohen’s d: .2<d < .5 = a small effect, d ≥ .5 a medium effect, and d ≥ .8 a large effect. Differences that are at least small are in bold.
As current ADHD medication status was associated with RS right parietal alpha power (r = .183, p = .036), the effect of medication status on Aim 1 and 2 findings involving RS right parietal alpha power was evaluated. Controlling for medication status, adolescents with and without ADHD did not differ on RS right parietal alpha power (p = .737) but adolescents with and without ADHD did differ in terms of the association between RS right parietal alpha power and the LPP (Table 2). In youth with ADHD, RS right parietal alpha power was negatively, whereas in youth without ADHD, it was not associated with the LPP.
Discussion
Our main research questions were whether RS or event-related synchronization (ERS) of theta and alpha activity are differentially associated with an ERP index and parent- and self-report measures of affective processing and a parent-report measure of inattention and whether observed associations differ between adolescents with and without ADHD. We also examined between-group differences in- and in lateralization of these EEG measures.
Findings indicated no between-group differences in RS or ERS of theta and alpha activity across frontal-midline, centroparietal, and parietal sites, with the exception of RS frontal-midline alpha power, which, consistent with earlier results (Ter Huurne et al., Reference Ter Huurne, Onnink, Kan, Franke, Buitelaar and Jensen2013; Vollebregt et al., Reference Vollebregt, Zumer, ter Huurne, Buitelaar and Jensen2016) was lower in youth with relative to without ADHD. The absence of group differences in RS theta activity is inconsistent with a body of work indicating atypical RS theta activity in ADHD and may reflect a developmental effect where youth with ADHD may have either developed a compensatory mechanism to counter a deficit or, as we measured middle-late adolescents, the maturational lag between youth with and without ADHD may have decreased to an extent where simple between-group differences are not detectable. Regarding ERS of theta, earlier findings indicate during a visual spatial attention task, 8–13-year-old children with ADHD exhibit elevated frontal-midline theta synchronization relative to children without ADHD (Guo et al., Reference Guo, Luo, Li, Chang, Sun and Song2020) whereas we found no differences in RS theta activity or, during affect regulation, in theta synchronization, between adolescents with and without ADHD. Differences across studies may be explainable, at least in part, by differences in experimental paradigms; Guo et al. applied a cognitively demanding visual spatial attention task and the elevated frontal-midline theta synchronization they observed may reflect a compensatory mechanism that counters attenuated attention arousal to achieve behavioral performance that is comparable to children without ADHD (Guo et al., Reference Guo, Luo, Li, Chang, Sun and Song2020). Here, we applied a simple guessing task that may have not elicited such compensatory cognitive mechanisms.
When examining associations in adolescents with and without ADHD separately, first, because visual inspection indicated boys with ADHD may exhibit different event-related frontal-midline theta synchronization during affect regulation than girls with and youth without ADHD, we conducted follow-up analyses and found that both boys with ADHD and girls with ADHD differed (exhibited lower synchronization) from their age- and sex-matched counterparts without ADHD but the difference between boys with and without ADHD was much larger than that between girls with and without ADHD. As noted, there continues to be debate about whether ADHD is less frequently diagnosed in girls because it is actually less prevalent in girls or it is less frequently diagnosed because girls with ADHD exhibit behavioral manifestations that are considerably less disruptive and, as a result, are less often referred for ADHD assessment. Relative to boys, girls with ADHD exhibit fewer hyperactive/impulsive but more inattentive symptoms (Biederman et al., Reference Biederman, Kwon, Aleardi, Chouinard, Marino, Cole, Mick and Faraone2005; Gaub & Carlson, Reference Gaub and Carlson1997; Gershon, Reference Gershon2002) and more commonly meet criteria for the inattentive presentation (Hinshaw et al., Reference Hinshaw, Owens, Sami and Fargeon2006). It may be for these reasons that teachers more often refer boys than girls for treatment for ADHD, even when showing equal levels of impairment (Sciutto et al., Reference Sciutto, Nolfi and Bluhm2004). An alternative and related explanation is that because of these differences in manifestation, boys with ADHD may differ more from boys without ADHD than the degree to which girls with ADHD differ from girls without ADHD. The pattern of findings obtained here is consistent with this hypothesis, at least with regard to frontal-midline theta synchronization during affect regulation. Certainly, there is also indication in the broader literature that childhood disorders are, although more prevalent in males, are more severe in females (Eme, Reference Eme1992) and this counters the above interpretation.
Second, when comparing correlations across groups, controlling for sex, there were no differential associations of ERS and RS theta and alpha power with affective processing and attention. Lack of differential association and especially absence of an association between theta and alpha activity measures with inattention may be due to differences in methods. In prior studies (where such differentiation and association were observed) (Cavanagh & Frank, Reference Cavanagh and Frank2014; Diao et al., Reference Diao, Qi, Xu, Fan and Yang2017; Kawasaki & Yamaguchi, Reference Kawasaki and Yamaguchi2012), attention was experimentally manipulated and measured in the laboratory. Here, attention was measured employing parent-report on a rating scale. Whereas the former method likely captures considerably more basic, homogeneous aspects and manifestations of differences in attention, the latter method captures more heterogeneous, observable aspects, and manifestations of attention.
There was, however, a double disassociation between ERS theta and RS alpha power with affectivity and elaborate affective/motivational processing. Controlling for sex, in adolescents with ADHD, event-related right parietal theta synchronization was positively associated with positive affectivity and RS right parietal alpha power was negatively associated with elaborate affective/motivational processing. These associations were in the opposite direction (at a trend level) in case of the theta-affective processing relation and were nonsignificant in case of the alpha-affective processing relation.
These results have both general and specific conceptual implications. Regarding general implications, our findings highlight the importance of moving beyond assessing simple between-group comparisons (e.g., comparing youth with to youth without ADHD) to – as has been done in an emerging body of work (Rosen et al., Reference Rosen, Walerius, Fogleman and Factor2015; Scarpelli et al., Reference Scarpelli, Gorgoni, D’atri, Reda and De Gennaro2019; D. W. Zhang et al., Reference Zhang, Li, Wu, Zhao, Song, Liu, Qian, Wang, Roodenrys, Johnstone, De Blasio and Sun2019) – the assessment of between-group differences in the relations across the characteristics of interest. The first approach (simple between-group comparisons) assumes that characteristics operate in isolation and thus that if groups differ on a relevant variable (e.g., ADHD diagnosis), they will also differ on the characteristics of interest (e.g., RS or event-related theta synchronization). Yet, developmental psychopathology research consistently shows that characteristics do not operate in isolation and that there is a complex interplay between various protective and risk factors (Cicchetti & Rogosch, Reference Cicchetti and Rogosch1996). The second approach (comparison of between-group differences in the relations across characteristics) appreciates this complexity; although groups may not differ on a relevant correlate or susceptibility trait, they may differ in how those conditions or traits interact and operate. Regarding specific implications, these data show that in adolescents with ADHD, event-related right parietal theta synchronization and RS right parietal alpha power may be employed as electrophysiological indices of behaviorally different ADHD-related affective characteristics that have been previously shown to be prognostically relevant, i.e., affectivity (Bunford et al., Reference Bunford, Kujawa, Dyson, Olino and Klein2021). Event-related theta synchronization during affect regulation following monetary gain or loss may be a biological marker of individual differences in dispositional affectivity whereas RS alpha power is a marker of elaborate affective/motivational processing.
Of note, the LPP has been conceptualized as reflecting not only elaborate affective/motivational processing but also sustained attention towards information that is affectively and motivationally salient (Bunford, Kujawa, et al., Reference Bunford, Kujawa, Fitzgerald, Monk and Phan2018; Kujawa, Klein, et al., Reference Kujawa, Klein and Proudfit2013). In this framework, in adolescents with ADHD, there was arguably a disassociation between ERS and RS theta and alpha power with affective processing and attention. As with any ERP, the cognitive function behind the LPP, the meaning of the LPP is largely determined by the experimental task during which it is elicited and measured. the Doors task by nature is appropriate for manipulating motivational processes; as such, even if emphasis is on the sustained attention interpretation of the LPP, the sustained attention it reflects is motivationally dependent or modulated. Further, in the current sample, the LPP was more closely related to rating scale measures of affective processing than of attention, underscoring an elaborate affective/motivational processing interpretation.
Finally, we found lateralization predominantly in adolescents with ADHD; in girls, RS parietal theta and alpha power was greater in the right hemisphere and in boys, RS parietal alpha power was greater in the left hemisphere. One exception was that in girls without ADHD, event-related parietal theta synchronization was greater in the left hemisphere. Prior results show that during a visual spatial attention task, 8–13-year-old children with ADHD exhibited adult-like posterior theta lateralization, such that in children with ADHD, the theta modulation index (i.e., changes in power to different stimuli) was greater in the right relative to the left hemisphere, whereas there was no right–left difference in children without ADHD (Guo et al., Reference Guo, Luo, Li, Chang, Sun and Song2020). In another study employing the same paradigm, in children with ADHD, alpha modulation was not attenuated in the right but was attenuated in the left hemisphere (Guo et al., Reference Guo, Luo, Wang, Li, Chang, Sun and Song2019). In combination with the current findings, data show an abnormal unilateral advantage in the parieto-occipital area in children with ADHD but the exact manifestation of such advantage – hyper- vs. hypoactivity and specificity to the left vs. the right hemisphere across theta and alpha – is unclear. Nevertheless, Guo et al., attributed such “adult-like” theta lateralization in children with ADHD to atypical neurodevelopment in the disorder where the premature development of lateralized theta modulation compensates for attention deficits and thereby promotes (close to typical) behavioral performance.
Directions for future research
It will be key to evaluate whether our findings generalize across development, i.e., they apply to children and adults with ADHD, especially given that at least in terms of the manifestation of ADHD symptoms, those become more differentiable, more dissimilar from each other from childhood through adolescents and into adulthood (Martel et al., Reference Martel, Levinson, Langer and Nigg2016). This increased differentiation may also characterize the differential relations between alpha and theta activity and inattention and affective processing in this population.
It will also be key to examine whether these findings replicate in independent samples, especially results obtained on differences across sexes, as the boy and girl subsamples were moderate in size.
Assessment of attention/ inattention via self-report, especially in children and adolescents with ADHD is not recommended (Pelham et al., Reference Pelham, Fabiano and Massetti2005). However, in studies aiming to address research questions similar to the ones evaluated here, augmenting parent-report with a biological, e.g., ERP, measure of attention – as we did for affective processing – will be an important extension and replication of our results.
As there is indication that EEG activity may differ across ADHD presentations (Reference Clarke, Barry, McCarthy and Selikowitz2001b; Barry et al., Reference Barry, Clarke and Johnstone2003, Clarke et al., Reference Clarke, Barry, McCarthy and Selikowitz2001c), it will be important to determine whether such presentations – the inattentive vs. the hyperactive/impulsive or the combined presentations – have any bearing on the findings observed here.
ADHD pharmacotherapy is associated with differences in cognitive performance and neural processing even when such therapy has been discontinued (Schlochtermeier et al., Reference Schlochtermeier, Stoy, Schlagenhauf, Wrase, Park, Friedel, Huss, Lehmkuhl, Heinz and Ströhle2011), the association between ADHD medication status and the relation between the herein examined EEG indices and affective processing and attention will need to be evaluated.
Conclusion
We examined, for the first time, whether – and at which sites – theta and alpha activity are differentially associated with affective processing and attention and whether these differential relations are modulated by ADHD status.
When ADHD status and sex are accounted for, in adolescents with ADHD, there is a double disassociation between ERS theta and RS alpha power with affectivity and elaborate affective/motivational processing. In either adolescents with or in adolescents without ADHD, there is no disassociation of ERS and RS theta and alpha power with affective processing and attention.
Findings also showed that as behaviorally, boys with ADHD differ from their age and sex matched peers without ADHD in event-related frontal-midline theta synchronization more than girls do.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0954579423000639
Data availability statement
Datasets and codes generated and/or analyzed for the current study are available from the corresponding author on reasonable request.
Funding statement
This research was funded by a Hungarian Academy of Sciences Momentum (“MTA Lendület”) Grant awarded to NB (#LP2018-3/2018). During the preparation of this article, TM was supported by a Hungarian Academy of Sciences Hungarian Brain Research Program (“NAP 3.0”) Grant awarded to NB (HAS-ELRN NAP2022-I-2/2022) and TB was supported by a Hungarian Academy of Sciences János Bolyai Grant (BO/00237/19/2).
Competing interests
None.
Ethical standard
This research was approved by the National Institute of Pharmacy and Nutrition (OGYÉI/17089-8/2019) and has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
