Article Figures & Data
Figures
- Figure 1.
TOVA experiment protocol design. Above, Number of targets and nontargets in each quarter and half, with target-to-nontarget frequency and total time. Below, Individual trial structure, millisecond timings, and stimuli appearance.
- Figure 2.
Behavioral results on standard scores (i.e., <100 implies worse performance than normative database). A, B, Commission errors (lower values = more errors compared with norm) and d′, from TOVA H1. C, RTV from TOVA H2 (points in panels A–C shown with horizontal jitter for visibility). D–F, Rug-and-density plots show commission errors, d′, omission errors, across both groups and TOVA halves. Red vertical lines show the data mean, 2.5%, and 97.5% quantiles; *p < 0.05, **p < 0.01, ***p < 0.001.
- Figure 3.
ERP image plots illustrate group-wise differences in target-locked and response-locked amplitudes. A, ERP images at frontal and parietal ROIs during correct response trials during TOVA H1. B, ERP images during TOVA H2. ERP images are matrices of activation data where each row is a single trial with color-coded amplitude from −8 to 8 μV (cool and warm colors, respectively). On the left are target-locked trials from −200 to +800 ms of stimulus onset and time of response shown by the black sigmoidal curve. On the right are response locked trials from −600 to +400 ms of RT, and stimulus onset shown by the black sigmoidal curve. Cumulative ERP waves are shown below each ERP image, split by the median RT. First and second row (both panels), control, and ADHD groups: these ERP images show that stimulus-locked early waves (i.e., amplitudes at a fixed lag from 0), and the response-locked waves are both much stronger for control than ADHD group. Control group’s P3 also clearly begins before the response, in contrast to ADHD group. Third row (both panels), ERP waves (solid lines) for each ROI. Vertical gray areas are test windows (aligned to N2, P3 in target-locked trials), group differences are highly significant. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
- Figure 4.
Whole-head scalp maps of broad-band amplitudes averaged within the time windows used for tests in Figure 3, for each condition: response versus inhibition × H1 versus H2. Color scales are matched to the range of the data shared between groups within each condition and time window. ROIs are shown as black and green circles.
- Figure 5.
α phase-locked ERP images in both ROIs, and parietal ITC. A, Phase-locked ERP images in the parietal ROI. Stacked correct-trial amplitudes are sorted by the prestimulus α (8–12 Hz) phase at −80 ms. Panels show the first and second halves of TOVA (H1 and H2), for targets (response) and nontargets (inhibition), and both groups. Amplitude from −11 to 11 μV is color-coded from blue to red, respectively. The control group shows higher amplitudes in stimulus-locked waves (starting at 100 ms), i.e., phase-resetting reaction is enhanced compared with ADHD group. B, Phase-locked ERPs in all conditions in the frontal ROI. All ERP images have adjacent ERPs split at median phase, with the post-stimulus phase resetting point of alignment marked by narrow black vertical lines. C, ITC calculated for both groups and both halves of TOVA, all four ITC curves show a peak around 200 ms and smaller harmonic peaks, especially at 400. Wider lines show when ITC was significantly above chance level (the level of ITC which is significant depends on the sample, so no horizontal indicator is drawn). Top left, Control group condition H1 versus H2, almost no difference between conditions is seen. Bottom left, ADHD group condition H1 versus H2, a small reduction from H1 to H2 is seen. Top right, Control H1 versus ADHD H1, substantial peak differences are seen at 200 (∼40%) and 400 (∼120%) ms. Bottom right, Control H2 versus ADHD H2, large peak differences are seen at 200 (∼90%) and 400 (∼112%) ms. Control versus ADHD comparisons also show that ADHD ITC is more dispersed, i.e., having weaker phase-locking to targets.
- Figure 6.
Log spectral power density calculated during periods of baseline (BL; solid lines) versus poststimulus (PS; dashed lines), for control (red) versus ADHD (blue). Panels show each condition by ROI combination. Frontally and parietally there is a higher-frequency to lower-frequency shift from baseline to poststimulus: parietal baseline peaks are α, whereas frontal peaks are at 8 Hz; in both ROIs the poststimulus shift is toward 4 Hz.
- Figure 7.
Contrasting groups for ERSPs, induced power, and evoked power. A, TFRs plotted within parietal ROI. B, TFRs plotted in frontal ROI. Each TFR is averaged across all electrodes in the ROI; TFRs are locked to stimulus onset at time 0 (dashed vertical lines); frequencies are plotted on a log scale. Within each panel, first row shows ERSP, second shows evoked power, third shows induced power, and fourth row shows the mean across frequencies of the upper three (with evoked power divided by 10 to illustrate). Left and middle columns show ERS/ERD for control and ADHD groups, respectively; the right column shows the subtracted difference between groups for the ROI-averaged TFR (or its mean in row four).
- Figure 8.
ERSP plots for condition: correct inhibition versus response trials, in H1 versus H2. A, ERSPs plotted in the parietal ROI. B, Frontal ROI ERSPs. The time-frequency data are averaged across all electrodes in the ROI; the scale of power perturbations goes from −2 to 2 dB; ERSPs are locked to stimulus onset at time 0 (dashed lines); frequencies are plotted on a log scale. Within each panel, each condition is plotted column-wise. Top and middle rows show ERS/ERD for control and ADHD groups, respectively. Bottom row shows the log-mean difference between groups for the ROI-averaged time-frequency data; these plots are masked by a permutation-based significance test: gray is non-significant (n.s.), lighter-toned blobs are different at p < 0.05, and full-color blobs are different at p < 0.0005 (all tests uncorrected).
In this issue
