The ADHD and Control groups did not differ in terms of age, sex, educational level, ethnic background or IQ (see ). The ADHD group had significantly higher rates of mood, other disruptive behaviors, and substance abuse disorders when compared to controls.
Patterns of cortical activation across baseline and cognitive activation conditions
Presented in are the power spectra for the ADHD and control groups across the EO, EC and CPT conditions. The repeated-measures ANOVAs indicated significant main effects of condition for all frequency bins (4-20 Hz) across conditions (F(2,104)= 3.6 -69.0), p-values range from <.05 to <.0001), which suggests that there are widespread cortical changes when going from EC to EO to CPT. Significant interaction effects of condition and ADHD diagnostic status emerged in the lower alpha range (8-10 Hz) in frontal (F(2,104)=5.9, p = .008, FDR-adjusted p= .01) and parietal (F(2,104)=5.0, p=.01, FDR adjusted p= .03) regions. Significant effects consistently emerged for the 8-9-Hz and 9-10-Hz frequency bins, therefore, we grouped them together for all analyses. Results of the analyses indicate that the ADHD group exhibited greater attenuation of 8-10 Hz power than controls suggesting increased cortical activation in the ADHD group. In addition, the ADHD group exhibited reduced power in the 8-10Hz range in the frontal (but not the parietal) region (Frontal: F(1,52)=6.1, p=.02, FDR adjusted p= .05); Parietal: F(1,52)=1.2, p=.27, FDR adjusted p=.41). This group difference was due primarily to significant differences in alpha power during the EC (F(1,52)=6.3, p=.01, FDR adjusted p=.03) and CPT conditions (F(1,52)=7.1, p=.01, FDR adjusted p=.03 ) where the ADHD group displayed significantly reduced alpha power. The group difference in alpha power during the EO conditions was not statistically significant (F(1,52)=2.5, p=.11, FDR adjusted p=.17). Because alpha activity is inversely related to cortical arousal, these results suggest that the ADHD group exhibited greater cortical arousal during the eyes closed condition.
Figure 1 Spectral power showing brain activity across conditions and by diagnosis. Lines indicate significant condition by diagnosis interaction effect; p-value is noted between the lines. Frequency bins are in 1 hertz (Hz) each. Alpha power is inversely correlated (more ...)
In the beta range, another interaction effect emerged according to ADHD status in the 17-18 Hz range indicating that controls show a slight attenuation in frontal beta activity during the EO and CPT conditions that ADHD group does not (Frontal: F(2,104)=3.91, p=.03, FDR adjusted p=.05). In the parietal region, another effect of diagnostic status on cortical activation was found in the 13-14 Hz range (Parietal: F(2,104)=3.6, p=.04; FDR adjusted p=.07). This suggests that controls again had greater attenuation of beta power during EO and CPT conditions than the ADHD group While ADHD adults exhibited slightly increased beta activity compared to controls, the between-group difference (i.e., main effect of diagnosis) was not statistically significant (Frontal 17-18Hz: F(1,52)=<1, n.s.; Parietal 13-14Hz: F(1,52)=<1, n.s.). Beta activity is thought to represent cortical activation and suggests that the ADHD group exhibited greater cortical activation than the controls during the CPT condition.
Behavioral performance on the CPT was compared across diagnostic groups and is presented in . Overall, CPT performance between ADHD and controls did not differ significantly, with the exception of beta response style (F(1,77) = 8.34, p=.005). Beta in this context refers to a person's response style or approach to responding on the task. Low beta values indicate that a person has a more impulsive response style that emphasizes commission errors over omission errors. High beta values indicate a more cautious response style that emphasizes errors of omission over errors of commission. Controls had significantly higher scores on beta when compared to the ADHD group, suggesting that controls were more cautious or careful in responding, while the ADHD group had a more impulsive response style.
Continuous Performance Test performance by ADHD diagnosis
Relationship between cortical activation and behavioral performance
In order to assess the functional relationship of cortical activation patterns with CPT performance, Pearson correlations were run between spectral power and CPT performance in the whole sample. All of the correlations were modest (none larger than .23) and statistically non-significant. We then ran separate correlations for the control and ADHD groups; results are summarized in . Differential patterns of correlations emerged according to diagnostic group suggesting that cortical activation patterns are associated with different cognitive processes among ADHD and control groups. Within the control group, increased frontal and parietal power in the 8-10 Hz range was significantly associated with higher rate of commission errors (frontal r=.43, p<.01; parietal r=.45, p<.01), faster reaction time (frontal r=-.49, p<.005; parietal r=-.48, p<.005) and higher scores on d-prime (frontal r=.46, p<.01; parietal r=.45, p<.01). These correlations suggest that increased alpha power is associated with a fast, impulsive response style and reduced stimulus discrimination among the controls. Only one performance variable (task variability) was significantly associated with cortical activity for the ADHD group, indicating that frontal alpha (8-10 Hz, r=-.45, p<.01), frontal beta (17-18 Hz, r=-.52, p<.005) and parietal beta power (13-14 Hz, r=-.47, p<.01) were associated with reduced task variability.
Correlations between cortical activation and task performance
A Fisher r to z transformation was used to determine whether the strength of the correlations between cortical activation and CPT performance differed according to ADHD status. As seen in , several statistically significant differences in correlation emerged (as indicated by the boxes) according to diagnostic status, the majority of which (5 out of 7) were in the frontal region. This suggests that cortical activity in the frontal regions is significantly different among ADHD and control participants.
Patterns of cortical activation during sustained attention task across time
Presented in is the EEG power recorded during the first third, middle third, and last third of the CPT task (each section ~5 minutes) for the ADHD and control groups. Repeated measures ANOVAs indicate main effects of time for the 7-12 Hz range that trended toward significance in the frontal region (F(1,62)=3.5-5.3, p=.03-.05, FDR adjusted p=.09) and the 12-13 Hz range in the parietal region (F(1,62)=4.2, p=.04, FDR adjusted p=.08). This suggests that both groups showed decreases in alpha and beta power over time on the CPT task. Significant interaction effects (time by diagnosis) emerged in the 8-10 Hz frequency range in the frontal region (F(1,62)=7.0, p=.01, FDR adjusted p=.05) and trended toward significance in the 9-10 Hz frequency range in the parietal region (F(1,62)=5.8, p=.02, FDR adjusted p=.07). This suggests that control and ADHD groups exhibited significantly different patterns of cognitive activation across time with both groups exhibiting similar levels of activation in the initial 5-minutes of the task. The control group then exhibited a large increase in alpha power in the next 10 minutes, whereas the ADHD group maintained the same level of alpha power during the same time period. This suggests that the control group adapted to the sustained attention task as time went on and required less cortical activation, while the ADHD group maintained a high level of neural activation throughout the task.
Figure 2 Alpha power (8-10 Hz) during sustained attention over time in frontal and parietal regions. Lines indicate significant condition by diagnosis interaction effect; p-value is noted between the lines. Note that alpha power is inversely related to cortical (more ...)