DSM-5 Adult Attention-Deficit/Hyperactivity Disorder: Sex Differences in EEG Activity.

This study examined sex differences in the EEG of adults diagnosed with Attention-Deficit/Hyperactivity Disorder (AD/HD) according to DSM-5 criteria. Sixteen females and 16 males with AD/HD, and age- and sex-matched control groups, had an eyes-closed resting EEG recorded from 19 electrode sites. EEGs were Fast Fourier transformed and estimates for total power, absolute and relative power in the delta, theta, alpha, beta and gamma bands, and the theta/beta ratio, were analysed across nine cortical regions. Males with AD/HD, compared with male controls, had globally reduced absolute beta, globally elevated relative theta, and a larger theta/beta ratio. In contrast, no global effects emerged between females with and without AD/HD. Significant group interactions indicated that globally elevated relative theta and elevated frontal-midline theta/beta ratio noted in males with AD/HD differed significantly from results in females. There are statistically significant EEG differences in relative theta and the theta/beta ratio between males and females with and without AD/HD. These results indicate that AD/HD affects the EEG activity of males and females differently. This study helps confirm the need for further independent examination of AD/HD within female populations.

Time Effects on Resting EEG in Children With/Without AD/HD.

In this study we extend on behavioural evidence to examine the effect of time on EEG measures related to arousal and emotion/motivation in children with/without AD/HD. Thirty children with AD/HD and 30 age- and sex-matched controls participated. EEG was recorded during an eyes-closed resting condition and divided into three 2.5 min blocks after pre-processing. Time effects for absolute and relative alpha activity were found in healthy controls; these effects did not interact with AD/HD status. Interactions between time and AD/HD status were found for absolute theta, relative theta, and theta/beta ratio (TBR), with these EEG indices increasing over time in children with AD/HD. Moreover, IQ played a role in the interaction between time and AD/HD status. These results are consistent with predictions from both the optimal stimulation model and the delay aversion model, and suggest important methodological considerations for future EEG research in children with/without AD/HD.

The EEG Theta/Beta Ratio: A marker of Arousal or Cognitive Processing Capacity?

Attention-Deficit/Hyperactivity Disorder (AD/HD) is the most common psychiatric disorder of childhood and has been extensively researched using EEG technology. Within this literature, one of the most widely examined measures has been the theta/beta ratio. The theta/beta ratio was initially hypothesised to represent the arousal mechanism. However, subsequent research has shown this to be inaccurate and it was hypothesised that the ratio represents cognitive processing capacity. To examine that hypothesis, this study aimed to test the relationship between the P300 and the theta/beta ratio. The P300, absolute alpha and the theta/beta ratio were measured at Fz, Cz and Pz, and correlated in a group of 47 normal adults. A significant positive correlation was found between P300 latency and the theta/beta ratio. No relationship was found between P300 amplitude and the theta/beta ratio. P300 amplitude, but not latency, significantly correlated with alpha power. These results support the hypothesis that the theta/beta ratio is a marker of cognitive processing capacity.

Mindfulness meditation alters neural oscillations independently of arousal.

Neuroscience has identified that mindfulness meditation induces a state of relaxed alertness, characterised by changes in theta and alpha oscillations and reduced sympathetic arousal, although the underlying mechanisms remain unclear. This study aims to address this gap by examining changes in neural oscillations and arousal during mindfulness meditation using both traditional and data-driven methods. Fifty-two healthy young adults underwent electroencephalography (EEG) and skin conductance level (SCL) recordings during resting baseline and mindfulness meditation conditions, both conducted with eyes closed. The EEG data revealed a significant decrease in traditional alpha (8-13 Hz) amplitude during mindfulness meditation compared to rest. However, no significant differences were observed between conditions in traditional delta, theta, beta, or gamma amplitudes. Frequency Principal Components Analysis (fPCA) was employed as a data-driven approach, identifying six components consistent across conditions. A complex delta-theta-alpha component significantly increased during mindfulness meditation. In contrast, low alpha (~9.5 Hz) and low alpha-beta (~11 Hz) components decreased significantly during mindfulness meditation. No significant differences were observed between conditions in the delta, high alpha, and high alpha-beta components. Additionally, there were no significant differences in SCL between conditions, nor were there correlations between traditional alpha or fPCA components and SCL. These findings support the conceptualisation of mindfulness meditation as a state of relaxed alertness, characterised by changes in neural oscillations likely associated with attention and awareness. However, the observed changes do not appear to be driven by arousal.

Global alpha power fluctuations in a self-caught mind-wandering paradigm are independent of arousal.

Mind-wandering is characterised as the emergence of thought and emotions which shift attention away from a primary task. It is thought to consume up to 50 % of our waking lives and has several negative implications. Breath-counting is one task that has been utilised in conjunction with electroencephalography (EEG) to examine the brain states associated with mind-wandering. Research has consistently found reductions in alpha oscillations during periods of mind-wandering relative to breath-focus. It is possible that such fluctuations reflect an arousal mechanism warranting further investigation. Thirty-seven participants completed a 15 min breath-counting task, with simultaneous recording of EEG and skin conductance level (SCL). During this task participants were required to self-identify periods of mind-wandering via button-press. Event-related spectral perturbation (ERSP) analysis was used to quantify changes in global alpha power (8-13 Hz) relative to the button press. The -8 to -4 s period prior to button-press was assessed as mind-wandering, and the 4 to 8 s period following the button-press as breath-focus. Relative to breath-focus, mind-wandering was associated with a significant decrease in global alpha power and significant increase in SCL, consistent with perceptual decoupling theory. However, changes in global alpha power and SCL did not correlate. These results suggest arousal is not the primary mechanism underlying alpha changes observed during breath-counting, thus additional processes should be considered.

Age-Related Differences in Prestimulus EEG Affect ERPs and Behaviour in the Equiprobable Go/NoGo Task.

Detailed studies of the equiprobable auditory Go/NoGo task have allowed for the development of a sequential-processing model of the perceptual and cognitive processes involved. These processes are reflected in various components differentiating the Go and NoGo event-related potentials (ERPs). It has long been established that electroencephalography (EEG) changes through normal lifespan development. It is also known that ERPs and behaviour in the equiprobable auditory Go/NoGo task change from children to young adults, and again in older adults. Here, we provide a novel examination of links between in-task prestimulus EEG, poststimulus ERPs, and behaviour in three gender-matched groups: children (8-12 years), young adults (18-24 years), and older adults (59-74 years). We used a frequency Principal Component Analysis (f-PCA) to estimate prestimulus EEG components and a temporal Principal Component Analysis (t-PCA) to separately estimate poststimulus ERP Go and NoGo components in each age group to avoid misallocation of variance. The links between EEG components, ERP components, and behavioural measures differed markedly between the groups. The young adults performed best and accomplished this with the simplest EEG-ERP-behaviour brain dynamics pattern. The children performed worst, and this was reflected in the most complex brain dynamics pattern. The older adults showed some reduction in performance, reflected in an EEG-ERP-behaviour pattern with intermediate complexity between those of the children and young adults. These novel brain dynamics patterns hold promise for future developmental research.

Resting state brain oscillations and symptom profiles in attention deficit/hyperactivity disorder.

Our perspective on resting-state electroencephalogram (EEG) is that it provides a window into the substrate of cognitive and perceptual processing, reflecting the dynamic potential of the brain's current functional state. In an extended research program into the electrophysiology of attention deficit/hyperactivity disorder (AD/HD), we have examined resting-state EEG power and coherence, and event-related potentials (ERPs), in children, adolescents, and adults with the disorder. We sought initially to identify consistent AD/HD anomalies in these measures, relative to normal control subjects, and then to understand how these differences related to existing models of AD/HD. An emergent strand in this program has been to clarify the EEG correlates of "arousal" and to understand the role of arousal dysfunction as a core anomaly in AD/HD. To date, findings in this strand serve to rule out a commonly held dictum in the AD/HD field: that elevated theta/beta ratio is an indicator of hypo-arousal. In turn, this requires further work to elucidate the ratio's functional significance in the disorder. Our brain dynamics studies relating prestimulus EEG amplitude and phase states to ERP outcomes are expected to help in this regard, but we are still at a relatively early stage, currently examining these relationships in control children, in order to better understand normal aspects of brain dynamics before turning to children with AD/HD. This range of studies provides a framework for our recent work relating resting-state EEG anomalies, in individuals with AD/HD, to their symptom profile. This has had promising results, indicating links between increased inattention scores and reduced resting EEG gamma power. With resting-state EEG coherence, reduced left lateralized coherences across several bands have correlated negatively with inattention scores, while reduced frontal interhemispheric coherence has been correlated negatively with hyperactivity/impulsivity scores. Such linkages appear to provide encouraging leads for future EEG research in AD/HD.

The Effects of Concentrative Meditation on the Electroencephalogram in Novice Meditators.

Following investigations into the benefits of meditation on psychological health and well-being, research is now seeking to understand the mechanisms underlying these outcomes. This study aimed to identify natural alpha and theta frequency components during eyes-closed resting and concentrative meditation states and examined their differences within and between two testing sessions. Novice meditators had their EEG recorded during eyes-closed resting and concentrative meditation conditions, before and after engaging in a brief daily concentrative meditation practice for approximately one-month. Separate frequency Principal Components Analyses (f-PCA) yielded four spectral components of interest, congruent between both conditions and sessions: Delta-Theta-Alpha, Low Alpha, High Alpha, and Alpha-Beta. While all four components showed some increase in the meditation condition at the second session, only Low Alpha (∼9.5-10.0 Hz) showed similar increases while resting. These findings support the use of f-PCA as a novel method of data analysis in the investigation of psychophysiological states in meditation.

Age-related changes in the EEG in an eyes-open condition: II. Subtypes of AD/HD.

This study investigated age-related changes in the EEG of subtypes of Attention-Deficit/Hyperactivity Disorder (AD/HD) compared with neurotypical controls, using an eyes-open resting condition. Two hundred and twenty five children between the ages of 5 and 16 years participated in this study. Groups consisted of AD/HD of the combined (AD/HDcom) and inattentive (AD/HDin) types, which were compared with controls for each of three age ranges: Young (5-8 years), Middle (9-12 years), and Old (13-16 years). The EEG was recorded and analyzed using AMLAB hardware and software, and Fourier transformed to provide estimates for total power, and absolute and relative power in the delta, theta, alpha and beta bands. Compared to controls, the AD/HD groups had globally increased relative theta. Regional differences were found for absolute and relative alpha and beta. Compared to AD/HDcom, AD/HDin had globally reduced total power, absolute and relative theta, and absolute alpha. Regional differences only were found for absolute and relative delta, absolute beta, and relative alpha. No simple interactions were found for diagnostic factors with age. These results indicate that maturational effects can be observed between subtypes of AD/HD and controls in the eyes-open condition with similarities to those reported in eyes-closed conditions, although substantial differences are apparent in the maturation of fast wave activity, primarily alpha. These results provide evidence of maturational differences between subtypes of AD/HD in eyes-open conditions, and provide additional support for the suggestion that subtypes of AD/HD differ in severity rather than the nature of underlying neurological impairment.

Age-related changes in the EEG in an eyes-open condition: I. Normal development.

This study investigated age related changes in the EEG of normal children in an eyes-open condition, in order to provide developmental norms for the study of children and adolescents with Attention-Deficit/Hyperactivity Disorder (AD/HD) - see our companion paper (Mason et al., submitted). EEG was recorded at 17 sites from 75 children (63 boys and 12 girls, in the approximately 5:1 ratio common in AD/HD) between the ages of 5 and 16 years. They comprised three groups, each of 25 children (21 boys and 4 girls), divided into Young (5-8 years), Middle (9-12 years), and Old (13-16 years). The EEG was recorded during an eyes-open resting condition and Fourier transformed to provide estimates for total power, and absolute and relative power in the delta, theta, alpha and beta bands. Total power and absolute delta, theta, alpha, and beta decreased with increasing age, as did relative delta, while relative alpha increased with increasing age. Changes occurred faster in the posterior regions for total power, absolute theta and alpha, relative theta, and in frontal regions for absolute and relative beta. Some lateral developmental effects differed with band. These results indicate that maturation effects observed in the eyes-open EEG show some similarities to those reported in eyes-closed conditions, although substantial differences are apparent in the maturation of fast wave activity, particularly alpha. The data provide simple age-norms for eyes-open investigations of EEG differences in young clinical groups, particularly AD/HD. They encourage further investigations of the activational effects of this simple eyes-closed/eyes-open manipulation, which may aid understanding of the energetics of behaviour.

Children with attention-deficit/hyperactivity disorder and autistic features: EEG evidence for comorbid disorders.

Attention-deficit/hyperactivity disorder (AD/HD) is the most common psychiatric disorder of childhood, although AD/HD is rarely the only diagnosis given to these children. Within the literature there is some debate as to whether it is valid to diagnose AD/HD with autism as a comorbid disorder, since the present diagnostic systems exclude the diagnosis of both disorders in the same child. The aim of this study was to determine whether electroencephalography (EEG) differences exist between two groups of children diagnosed with AD/HD, one scoring high (AD/HD+) and one scoring low (AD/HD-) on a measure of autism. The EEG was recorded during an eyes-closed resting condition from 19 electrodes, and Fourier transformed to provide absolute and relative power estimates in delta, theta, alpha and beta bands. Compared to age- and sex-matched controls, the AD/HD- group had increased absolute power in all frequency bands, somewhat higher relative theta activity and decreased relative delta. In comparison to the AD/HD- group, patients with autistic features (AD/HD+) had a number of qualitative differences in the beta and theta bands. These results indicate the presence of two comorbid conditions in the AD/HD+ group, which suggests that AD/HD and autism can occur in the same individual.

EEG coherence in children with attention-deficit/hyperactivity disorder: differences between good and poor responders to methylphenidate.

This retrospective study investigated differences in regional derivations of EEG coherence between good and poor responders to methylphenidate (MPH) in children (aged 8-12 years) with the combined type of attention-deficit/hyperactivity disorder (AD/HD). Participants included groups of good and poor male MPH responders and an aged-matched group of male controls. An eyes-closed, resting electroencephalogram (EEG) was recorded from 21 electrode sites. Coherence was calculated from eight intrahemispheric and eight interhemispheric electrode pairs, for the delta, theta, alpha and beta frequency bands. Compared with controls, the AD/HD participants had enhanced laterality over short-medium inter-electrode distances, and elevated frontal interhemispheric coherences, in the theta band. Good MPH responders had higher intrahemispheric coherences than poor MPH responders over short-medium and long inter-electrode distances in the beta band. Enhanced laterality at short-medium inter-electrode distances suggests that the AD/HD children may have a developmental lag in short-axonal connections in the left hemisphere. Elevated frontal interhemispheric theta coherence consistently indicates some frontal dysfunction in AD/HD. The beta coherence differences found between good and poor MPH responders could indicate that good MPH responders have some type of structural dysfunction associated with cortical connections involved in attention/arousal.

Resting state EEG power research in Attention-Deficit/Hyperactivity Disorder: A review update.

This article reviews the eyes-open and eyes-closed resting electroencephalogram (EEG) literature for Attention-Deficit/Hyperactivity Disorder (AD/HD) from 2002 to 2019. This time has seen substantial increase in interest in resting state EEG, with investigations moving from a comparison of an AD/HD sample to a control group, to research investigating a wide range of theoretical and clinical aspects of the disorder. This has included investigations of subgroups based on their EEG profile, sex differences in the EEG, increasing interest in the EEG of adults, and the effects of comorbid disorders on the EEG of people with AD/HD. Research has further investigated the hyper- and hypo- arousal models of AD/HD, as well as the developmental deviation model. From a clinical perspective, a growing body of literature is emerging trying to ascertain if the EEG can be used as a diagnostic test, particularly the theta/beta ratio. While these advances have been made, there is widespread use of both the eyes-open and eyes-closed resting paradigms as being interchangeable, despite the two paradigms differ on a number of important factors. There is also lack of independent replication within the literature, which is needed to consolidate many of the findings that have been published.

Preferred EEG brain states at stimulus onset in normal ageing: Explorations in a fixed interstimulus interval Go/NoGo task.

Some 40 years ago, Erol Basar began an interesting line of brain dynamics research on the phase of oscillatory EEG activity. Stimuli presented at a fixed interstimulus interval (ISI) were reported to produce a "preferred phase angle" at stimulus onset in the ongoing EEG oscillations, so that cortical negativity occurred more often than expected. Our previous work has confirmed this phenomenon in young adults, and linked it to enhancements in stimulus-elicited ERP components and behaviour. The present study sought evidence for this phenomenon in older adults. Twenty healthy independent-living participants (5 males) aged 59.8 to 74.8 years (M = 68.2) completed a fixed ISI equiprobable auditory oddball task. EEG phases in four traditional bands were computed separately for each artefact-free Go and NoGo trial. These were used to group accepted trials according to the phase quartile at stimulus onset in each band, forming 16 mean ERPs for each participant at each phase quartile/band combination, separately for Go and NoGo. The 32 mean ERPs were decomposed by separate Principal Component Analyses, and the N1 and P3 components were examined as a function of EEG phase at stimulus onset in each band. Differential preferential occurrence of phase states was found in all bands, generally confirming prior results in young adults despite important age differences in both EEG spectra and ERP morphology. Substantial effects on N1 and P3 components, as well as prestimulus CNVs, are discussed. The stability of this phenomenon across age indicates its important role in perception and cognition.

Natural alpha frequency components in resting EEG and their relation to arousal.

OBJECTIVE: Global EEG alpha activity is negatively correlated with skin conductance level (SCL), supporting alpha as an inverse marker of arousal. Frequency Principal Components Analysis (f-PCA) of resting EEG amplitude spectra has demonstrated natural components in the alpha band of healthy persons. This is a preliminary exploration of whether such components differ with arousal, possibly underpinning the anomalous ADHD hypoarousal link to reduced alpha. METHOD: Twenty-seven right-handed undergraduate students participated in three 2 minute blocks of resting eyes-open/closed EEG and SCL: EO1, EC, EO2. For each condition, mean EEG spectra were submitted to separate f-PCAs. RESULTS: The inverse alpha/SCL relationship was confirmed for band amplitudes. EO had two alpha components; both correlated negatively with SCL. EC alpha contained four components, but only one had a substantial negative correlation with SCL; two had no relationship, suggesting natural alpha components with different non-arousal functionality in EC. CONCLUSION: Some alpha components in both EC and EO reflect arousal, with other non-arousal components in EC. Our f-PCA approach offers insight into previously-noted alpha anomalies in disorders such as ADHD. SIGNIFICANCE: This proof of concept demonstration in typical participants may provide the basis for a new research effort in clinical disorders involving atypical arousal patterns.

The Theta/Beta Ratio as an Index of Cognitive Processing in Adults With the Combined Type of Attention Deficit Hyperactivity Disorder.

An elevated theta/beta ratio in the EEG has long been observed among individuals with Attention Deficit Hyperactivity Disorder (ADHD). The theta/beta ratio was previously hypothesised to be an index of arousal, but a number of studies failed to find any association between the ratio and indices of arousal, instead proposing that the theta/beta ratio may actually be indicative of cognitive processing. This hypothesis was tested by Clarke et al using a sample of healthy adults, with results indicating that the theta/beta ratio correlated with a marker of cognitive processing (P300 latency in an auditory oddball task), while P300 amplitude correlated with an arousal marker (alpha power). The aim of this study was to test whether similar results could be found in a sample of 41 adults with the combined type of ADHD. EEGs were recorded during an eyes-closed resting condition and an auditory oddball task. Results demonstrated that the theta/beta ratio correlated significantly with P300 latency. Absolute alpha power did not correlate significantly with P300 amplitude or P300 latency. These results support the hypotheses that the theta/beta ratio is a marker of cognitive processing capacity in both the general population and in participants with ADHD, and that the alpha/arousal linkage is anomalous in ADHD.

Dysfunctional response preparation and inhibition during a visual Go/No-go task in children with two subtypes of attention-deficit hyperactivity disorder.

While a response inhibition problem is well-established in children with attention-deficit/hyperactivity disorder of the combined subtype (AD/HDcom), the predominantly inattentive subtype (AD/HDin) has not been investigated previously. This study examined control versus subtype differences in visually evoked response inhibition using task performance and event-related potential (ERP) measures. Children with AD/HDcom (n=15) and AD/HDin (n=15) and age-matched controls (n=15) performed a cued visual Go/Nogo task requiring either activation or inhibition (30%) of a button-press response to the S2 (Go or Nogo stimulus) following the S1 (warning stimulus), presented 1380 ms earlier. Task performance and ERP indices of Warning, Go and Nogo stimulus processing, as well as preparation during the S1-S2 interval, were examined for group differences. Behavioural results indicated a response inhibition deficit in children with AD/HDcom and AD/HDin, with additional response activation problems in AD/HDcom. Topographic ERP differences between controls and both clinical groups suggested atypical (a) preparation for S2 as indexed by the late CNV, (b) early sensory/attentional processing of both S1 and S2, and (c) response inhibition as indexed by N2 and P3. In addition to replicating previous AD/HDcom findings, these results indicate that children with AD/HDin differ from controls in response preparation and inhibition during a cued visual Go/Nogo task.

EEG development in Attention Deficit Hyperactivity Disorder: From child to adult.

OBJECTIVE: Attention Deficit Hyperactivity Disorder (ADHD) is one of the most common psychiatric disorders found in children. While an extensive literature has documented the EEG in this clinical population, few studies have investigated EEG throughout the lifespan in ADHD. This study aimed to investigate EEG maturational changes, in subjects with ADHD combined type, that spanned from childhood into adulthood. METHOD: Twenty five male adults with ADHD were assessed between the ages of 8-12 years and again as adults. At both ages, an EEG was recorded during an eyes-closed resting period, and power estimates were calculated for relative delta, theta, alpha and beta. RESULTS: At the childhood assessment, the ADHD subjects had elevated posterior delta. Relative theta was elevated, with diminished alpha activity across all sites. Significant maturational changes were observed, with reductions in the delta and theta bands, and increases in the alpha and beta bands across all electrodes. In adulthood, relative to controls, diminished frontal delta and elevated global theta activity were apparent. CONCLUSIONS: Substantial developmental changes occurred in the EEG of these subjects. These results identify important issues when using EEG as part of the diagnosis for ADHD. SIGNIFICANCE: This study is the first to explore EEG changes from childhood to adulthood over an 11 year period in the same subjects with ADHD.

EEG power and coherence in autistic spectrum disorder.

OBJECTIVE: Autistic spectrum disorder (ASD) has been defined as a neurodevelopmental disorder with associated deficits in executive function, language, emotional, and social function. ASD has been associated with pathophysiology in cerebral organization. The current study investigated quantitative EEG findings in twenty children diagnosed with autistic disorders as compared to 20 controls matched for gender, age and IQ. METHODS: The EEG was recorded during an eyes-closed resting condition and topographical differences in cerebral functioning were examined using estimates of absolute, relative, and total power, as well as intrahemispheric and interhemispheric coherences. RESULTS: There were group differences in power, intrahemispheric and interhemispheric coherences. Findings included excessive theta, primarily in right posterior regions, in autistics. There was also a pattern of deficient delta over the frontal cortex and excessive midline beta. More significantly, there was a pattern of underconnectivity in autistics compared to controls. This included decreased intrahemispheric delta and theta coherences across short to medium and long inter-electrode distances. Interhemispherically, delta and theta coherences were low across the frontal region. Delta, theta and alpha hypocoherence was also evident over the temporal regions. Lastly, there were low delta, theta and beta coherence measurements across posterior regions. CONCLUSIONS: These results suggest dysfunctional integration of frontal and posterior brain regions in autistics along with a pattern of neural underconnectivity. This is consistent with other EEG, MRI and fMRI research suggesting that neural connectivity anomalies are a major deficit leading to autistic symptomatology. SIGNIFICANCE: This paper reports the largest integrated study of EEG power and coherence during a resting state in children suffering autism spectrum disorder.

Timing of caffeine's impact on autonomic and central nervous system measures: clarification of arousal effects.

The timing of caffeine effects on arousal levels was examined. From previous work in our laboratory, an increase in skin conductance level (SCL) was used as the marker of arousal increase, and we sought to identify the timing of this and related effects following caffeine ingestion. A single oral dose of caffeine (250 mg) was used in a randomised double-blind placebo-controlled repeated-measures cross-over study. Eyes-closed resting electroencephalogram (EEG) and autonomic data (SCL, heart rate, respiration rate, and systolic and diastolic blood pressure) during 2 min epochs that commenced every 4 min after ingestion, were analysed. The SCL placebo data were used to identify potential arousal measures prior to examining caffeine effects. Caffeine was associated with increased SCL, increased respiratory rate and a global reduction in alpha power. There were no significant cardiovascular effects of caffeine-induced arousal. These caffeine results are consistent with our recent electrodermal and EEG studies of arousal, and confirm the potential use of caffeine as a simple means of experimentally modifying arousal levels without task-related confounds.