The resting electroencephalogram (EEG) reflects development and arousal, but whether it can support clinical diagnosis of attention-deficit/hyperactivity disorder (ADHD) remains controversial. Here we examined whether theta power and theta/beta ratio are consistently elevated in ADHD and younger age as proposed. Topographic 48-channel EEG from 32 children (8-16 years) and 22 adults (32-55 years) with ADHD and matched healthy controls (n = 30 children/21 adults) was compared. Following advanced artefact correction, resting EEG was tested for increased theta and theta/beta activity due to ADHD and due to normal immaturity. Discriminant analyses tested classification performance by ADHD and age using these EEG markers as well as EEG artefacts and deviant attentional event-related potentials (ERPs). No consistent theta or theta/beta increases were found with ADHD. Even multivariate analyses indicated only marginal EEG power increases in children with ADHD. Instead, consistent developmental theta decreases were observed, indicating that maturational lags of fewer than 3 years would have been detected in children. Discriminant analysis based on proposed simple spectral resting EEG markers was successful for age but not for ADHD (81 vs. 53 % accuracy). Including ERP markers and EEG artefacts improved discrimination, although not to diagnostically useful levels. The lack of consistent spectral resting EEG abnormalities in ADHD despite consistent developmental effects casts doubt upon conventional neurometric approaches towards EEG-based ADHD diagnosis, but is consistent with evidence that ADHD is a heterogeneous disorder, where the resting state is not consistently characterised by maturational lag.

No abstract

Frontal midline (fm-)theta activity has been related to working memory (WM) processes, as it typically increases with WM load. The robustness of this effect, however, varies across studies and subjects, putting limits to its interpretation. We hypothesized that variation in the fm-theta effect may reflect individual differences in task difficulty with increasing WM load as indicated by behavioural responses. We further tested whether effects in the alpha range are robust markers of WM load. We recorded 64-channel EEG from 24 healthy adults while they memorized either 2 or 4 unfamiliar symbols (low vs. high WM load) in a modified Sternberg task. The last 2 s of the retention phase were analyzed for WM load-related changes in the theta (5-7 Hz) and alpha range (lower: 8-10 Hz, upper: 10.5-12.5 Hz). Higher WM load led to less accurate and slower responses. The increase of fm-theta with WM load was most pronounced at fm electrodes, localized to anterior cingulate regions, and correlated with the participants' decrease in accuracy due to higher WM load. Alpha peak frequency increased in the high compared to the low WM load condition, corresponding to a decrease in lower alpha range across all channels. The results demonstrate that previously reported variation in fm-theta workload effects can partly be explained by variation in task difficulty indexed by individual task accuracy. Moreover, the results also demonstrate that alpha WM load effects are prominent when separating upper and lower alpha. AbstractFrontal midline (fm-)theta activity has been related to working memory (WM) processes, as it typically increases with WM load. The robustness of this effect, however, varies across studies and subjects, putting limits to its interpretation. We hypothesized that variation in the fm-theta effect may reflect individual differences in task difficulty with increasing WM load as indicated by behavioural responses. We further tested whether effects in the alpha range are more robust markers of WM load. We recorded 64-channel EEG from 24 healthy adults while they memorized either 2 or 4 unfamiliar symbols (low versus high WM load) in a modified Sternberg task. The last 2 seconds of the retention phase were analyzed for WM load-related changes in the theta (5-7 Hz) and alpha range (lower: 8-10 Hz, upper: 10.5-12.5 Hz). Higher WM load led to less accurate and slower responses. The increase of fm-theta with WM load was most pronounced at frontal midline electrodes, localized to anterior cingulate regions, and correlated with the participants' decrease in accuracy due to higher WM load. Alpha peak frequency increased in the high compared to the low WM load condition, corresponding to a decrease in lower alpha range across all channels. The results demonstrate that previously reported variation in fm-theta workload effects can partly be explained by variation in task difficulty indexed by task accuracy. Moreover, the results also demonstrate that alpha WM load effects are prominent when separating upper and lower alpha.