1. Neuroscience

Theta beta ratio in attention deficit hyperactivity disorder using a multiverse analysis

  1. Dawid Strzelczyk  Is a corresponding author
  2. Andrea Vetsch
  3. Nicolas Langer
  1. Methods of Plasticity Research, Department of Psychology, University of Zurich, Switzerland
  2. Neuroscience Center Zurich (ZNZ), Switzerland
https://doi.org/10.7554/eLife.111114.3
Share this article
Cite this article
  1. Dawid Strzelczyk
  2. Andrea Vetsch
  3. Nicolas Langer
(2026)
Theta beta ratio in attention deficit hyperactivity disorder using a multiverse analysis
eLife 15:RP111114.
https://doi.org/10.7554/eLife.111114.3
  2. Download BibTeX
  3. Download .RIS
15 figures, 12 tables and 1 additional file

Figures

Figure 1
Download asset Open asset
Final analysis tree illustrating all included analytical specifications.

Each node represents a distinct analytical parameter in the multiverse analysis. Note that due to space limitations, not all parameter nodes are displayed in this visualization. Information on comorbid diagnoses and medication status was available only in the healthy brain network (HBN) sample, and these factors were, therefore, investigated exclusively in this dataset. HC = healthy control, IAF = individual alpha frequency.

Figure 2
Download asset Open asset
Overview of sample characteristics and variable distributions.

Density plots show the distribution of (A) age, (B) IQ, (C) IAF, (D–F), and SWAN scores (inattention, hyperactivity, and total) across the three groups. (G–I) Scatter plots display the relationship between (G) IAF and age, (H) between TBR and age, (I) and TBR and IAF across all participants. The regression line is shown with standard error bands. Note. IAF = individual alpha frequency. TBR = theta beta ratio. HC = healthy control.

Figure 3 with 1 supplement
Download asset Open asset
Neurophysiological data from the healthy brain network sample.

(A) Scalp topographies, (B) aperiodic-adjusted power spectra, and (C) aperiodic signal for HC, attention deficit hyperactivity disorder (ADHD)-combined, and ADHD-inattentive groups. The figure displays aperiodic-adjusted power during eyes open (EO) condition, computed using a canonical frequency range (theta: 4–8  Hz; beta: 13–30  Hz) and TBR. The aperiodic-adjusted spectra and the reconstructed aperiodic signal are shown on a log-log scale. Electrodes labels highlighted on the topographies correspond to the six regions of interest derived from literature used across different branches of multiverse analysis. The power spectra and aperiodic signal were computed by averaging across all electrodes within each respective region of interest. Note. TBR = theta beta ratio. HC = healthy controls. Comb = combined. In = inattentive.

Figure 3—figure supplement 1
Download asset Open asset
Neurophysiological data from the healthy brain network sample.

(A) Scalp topographies, (B) 1/f-uncorrected power spectra, and (C) aperiodic signal for HC, attention deficit hyperactivity disorder (ADHD)-combined, and ADHD-inattentive groups. The figure displays 1/f-uncorrected power during EO condition, computed using a fixed frequency range (theta: 4–8  Hz; beta: 13–30  Hz) and theta-beta ratio (TBR). The 1/f-uncorrected spectra and the reconstructed aperiodic signal are shown on a log-log scale. Electrodes highlighted on the topographies correspond to the six regions of interest derived from literature used across different branches of multiverse analysis. The power spectra and aperiodic signal were computed by averaging across all electrodes within each respective region of interest. Note. HC = healthy controls. Comb = combined. In = inattentive.

Figure 4
Download asset Open asset
Specification curve representing all universes for the healthy control attention deficit hyperactivity disorder (ADHD)-Inattentive contrast.

The top panel shows regression estimates for each specification, sorted by regression estimate, with 95% confidence intervals. Statistically significant positive estimates are shown in green, negative estimates in red, and non-significant estimates in gray. The bottom panel indicates which analytical choices were associated with significantly positive (i.e. higher theta-beta ratio (TBR) for ADHD compared to healthy control, HC) or negative (i.e. lower TBR for ADHD compared to HC) effects across specifications. Note. CI = confidence interval. f=female. EC = eyes closed. EO = eyes open. AVG = average reference. LM = linked mastoid reference. canonical = canonical frequency range. relative to IAF = bandwidth relative to individual alpha frequency.

Figure 5
Download asset Open asset
Possibility space of regression estimates from the subset of analytical choices yielding the highest number of significant results.

(A) Attention deficit hyperactivity disorder (ADHD)-Inattentive * individual alpha frequency (IAF) interaction: Significant effects (46 significant out of 48 possibilities) were predominantly observed when using frequency bands relative to IAF, the aperiodic signal, and including children with comorbid diagnoses. (B) ADHD-inattentive * Age * IAF interaction: This positive effect emerged almost exclusively when TBR was computed using canonical frequency bands, aperiodic-adjusted power, and analyses included participants with comorbidities (37 out of 48 significant). (C) ADHD-inattentive * Gender * IAF interaction: The effect was found exclusively under frequency bands relative to IAF, primarily in the aperiodic signal, and when children with comorbid diagnoses were included (28 out of 48 significant). Note. The significant positive estimates are highlighted green, significant negative in red, and non-significant in gray.

Figure 6
Download asset Open asset
Specification curve representing all universes for the healthy control attention deficit hyperactivity disorder (ADHD)-combined contrast.

The top panel shows regression estimates for each specification, sorted by regression estimate, with 95% confidence intervals. Statistically significant positive estimates are shown in green, negative estimates in red, and non-significant estimates in gray. The bottom panel indicates which analytical choices were associated with significantly positive (i.e. higher theta-beta ratio, TBR, for ADHD compared to healthy control, HC) or negative (i.e. lower TBR for ADHD compared to HC) effects across specifications. Note. CI = confidence interval. f=female. EC = eyes closed. EO = eyes open. AVG = average reference. LM = linked mastoid reference. canonical = canonical frequency range. relative to IAF = bandwidth relative to individual alpha frequency.

Figure 7 with 2 supplements
Download asset Open asset
Possibility space of regression estimates from the subset of analytical choices yielding the highest number of significant results in the attention deficit hyperactivity disorder (ADHD)-combined vs. healthy control comparisons.

(A) ADHD-combined * age interaction: Significant effects (24 out of 96) were primarily observed when using frequency bands relative to individual alpha frequency (IAF) and the aperiodic signal. (B) ADHD-combined * IAF interaction: A negative effect emerged almost exclusively under frequency bands relative to IAF, predominantly in the aperiodic signal, and in analyses, including children on medication (47 out of 48 significant). (C) ADHD-combined * Gender interaction: This positive effect appeared almost exclusively in the eyes closed (EC) condition, with frequency bands relative to IAF and 1/f-uncorrected power (27 out of 48 significant). (D) ADHD-combined * Age * Gender, (E) ADHD-combined * Gender * IAF and (F) ADHD-combined * Age * Gender * IAF interactions: These effects appeared mainly in the EC condition, almost exclusively when comorbidities were included and medication excluded (26, 35 and 31 out of 72 significant, respectively). Note. Significant positive estimates are highlighted green, significant negative in red, and non-significant estimates are shown in grey.

Figure 7—figure supplement 1
Download asset Open asset
Proportion of specifications showing a significantly positive effect by regression term and attention deficit hyperactivity disorder (ADHD) subtype across all analytical specifications.

Each level of condition, reference, frequency definition, comorbid disorder, and medication includes 288 analytical universes. The Aperiodic Signal factor includes 192 universes per level, and the region of interest includes 96 universes per level. The red line in the plot indicates the exact proportion of significant results required for the binomial test to reach statistical significance (p<0.05) for each subset size.

Figure 7—figure supplement 2
Download asset Open asset
Proportion of specifications showing a significantly negative effect by regression term and attention deficit hyperactivity disorder (ADHD) subtype across all analytical specifications.

Each level of condition, reference, frequency definition, comorbid disorder, and medication includes 288 analytical universes. The Aperiodic Signal factor includes 192 universes per level, and the region of interest includes 96 universes per level. The red line in the plot indicates the exact proportion of significant results required for the binomial test to reach statistical significance (p<0.05) for each subset size.

Figure 8
Download asset Open asset
Distribution of effect sizes across analytical subspaces.

Violin plots display the distribution of Cohen’s d values across analytical paths for the healthy control (HC) versus attention deficit hyperactivity disorder (ADHD)-inattentive comparison and the HC versus ADHD-combined comparison. Subspaces were defined by crossing frequency band definition (IAF-relative vs canonical frequency bands) with spectral representation (aperiodic signal, 1/f-uncorrected power, and aperiodic-adjusted power), resulting in a total of 24 subspaces. Individual dots represent effect sizes from single analytical paths, while the error bars indicate mean effect sizes and 95% confidence intervals within each subspace. Despite frequent significant interaction effects observed in some multiverse subspaces, the corresponding effect sizes remained small across analyses. Positive values indicate higher values in healthy controls relative to ADHD groups.

Appendix 1—figure 1
Download asset Open asset
Overview of sample characteristics and variable distributions.

Density plots show the distribution of (A) IAF, (B) age, and (C) IQ across the three groups. (D–F) Scatter plots display the relationship between (D) IAF and age, (E) TBR and age, (F) and TBR and IAF across all participants. The regression line is shown with standard error bands. Note. IAF = individual alpha frequency. TBR = theta beta ratio. HC = healthy control.

Appendix 1—figure 2 with 1 supplement
Download asset Open asset
Neurophysiological data from the validation sample.

(A) Scalp topographies, (B) aperiodic-adjusted power spectra, and (C) aperiodic signal for HC, attention deficit hyperactivity disorder (ADHD)-combined, and ADHD-inattentive groups. The figure displays aperiodic-adjusted power during eyes-open (EO) condition, computed using a canonical frequency range (theta: 4–8  Hz; beta: 13–30  Hz) and theta-beta ratio (TBR). Electrodes labels highlighted on the topographies correspond to the six regions of interest derived from literature used across different branches of multiverse analysis. The power spectra and aperiodic signal were computed by averaging across all electrodes within each respective region of interest. Note. HC = healthy controls. Comb = combined. In = inattentive.

Appendix 1—figure 2—figure supplement 1
Download asset Open asset
Neurophysiological data from the validation sample.

(A) Scalp topographies, (B) 1/f-uncorrected power spectra, and (C) aperiodic signal for HC, attention deficit hyperactivity disorder (ADHD)-combined, and ADHD-inattentive groups. The figure displays 1/f-uncorrected power during eyes-open (EO) condition, computed using a fixed frequency range (theta: 4–8 Hz; beta: 13–30 Hz) and theta-beta ratio (TBR). Electrodes labels highlighted on the topographies correspond to the six regions of interest derived from literature used across different branches of multiverse analysis. The power spectra and aperiodic signal were computed by averaging across all electrodes within each respective region of interest. Note. HC = healthy controls. Comb = combined. In = inattentive.

Appendix 1—figure 3
Download asset Open asset
Specification curve representing all universes for the healthy control vs. attention deficit hyperactivity disorder (ADHD)-inattentive comparison in the validation sample.

The top panel shows regression estimates sorted by effect size with 95% confidence intervals. Statistically significant positive estimates are shown in green, negative in red, and non-significant in gray. The bottom panel maps the analytical choices that produced each estimate. Note. CI = confidence interval. f=female. EC = eyes closed. EO eyes open. AVG = average reference. LM = linked mastoid reference. canonical = canonical frequency range. relative to IAF = bandwidth relative to individual alpha frequency.

Appendix 1—figure 4
Download asset Open asset
Possibility space for the attention deficit hyperactivity disorder (ADHD)-inattentive * Age and ADHD-inattentive * Age * individual alpha frequency (IAF) interactions.

(A) Negative ADHD-inattentive * Age effect: 8 out of 24 significant universes with IAF-relative bands and aperiodic signal. (B) Positive ADHD-inattentive * Age * IAF effect: 15/24 significant under aperiodic signal with IAF-relative bands (C) Positive ADHD-inattentive * Age * IAF effect: 15/24 significant with IAF-relative bands and uncorrected power. Note. The significant positive estimates are highlighted green, significant negative in red, and non-significant in gray.

Appendix 1—figure 5
Download asset Open asset
Specification curve representing all universes for the healthy control vs. attention deficit hyperactivity disorder (ADHD)-Combined comparison in the validation sample.

The top panel shows regression estimates sorted by effect size with 95% confidence intervals. Statistically significant positive estimates are shown in green, negative in red, and non-significant in gray. The bottom panel maps the analytical choices that produced each estimate. Note. CI = confidence interval. F=female. EC = eyes closed. EO = eyes open. AVG = average reference. LM = linked mastoid reference. Canonical = canonical frequency range. Relative to IAF = bandwidth relative to individual alpha frequency.

Appendix 1—figure 6 with 2 supplements
Download asset Open asset
Possibility space for the attention deficit hyperactivity disorder (ADHD)-Combined * Age, ADHD-combined * Age * individual alpha frequency (IAF) and ADHD-combined * Age * Gender * IAF interactions.

(A) Negative ADHD-combined * Age effect: 12 out of 24 significant universes with IAF-relative bands and uncorrected power. (B) Positive ADHD-combined * Age * IAF effect: 14/24 significant under aperiodic signal with IAF-relative bands (C) Positive ADHD-combined * Age *Gender * IAF effect: 13/24 significant with IAF-relative bands and uncorrected power. Note. The significant positive estimates are highlighted green, significant negative in red and non-significant in gray.

Appendix 1—figure 6—figure supplement 1
Download asset Open asset
Proportion of specifications showing a significantly positive effect by regression term and attention deficit hyperactivity disorder (ADHD) subtype across all analytical specifications.

Each level of condition, reference, frequency definition includes 72 analytical universes. The Aperiodic Signal factor includes 48 universes per level, and the Region of Interest includes 24 universes per level. The red line in the plot indicates the exact proportion of significant results required for the binomial test to reach statistical significance (p<0.05) for each subset size.

Appendix 1—figure 6—figure supplement 2
Download asset Open asset
Proportion of specifications showing a significantly negative effect by regression term and attention deficit hyperactivity disorder (ADHD) subtype across all analytical specifications.

Each level of condition, reference, frequency definition includes 72 analytical universes. The Aperiodic Signal factor includes 48 universes per level, and the Region of Interest includes 24 universes per level. The red line in the plot indicates the exact proportion of significant results required for the binomial test to reach statistical significance (p<0.05) for each subset size.

Appendix 3—figure 1 with 2 supplements
Download asset Open asset
Specification curve representing all universes for the dimensional SWAN score analysis.

The top panel shows regression estimates for each specification, sorted by regression estimate, with 95% confidence intervals. Statistically significant positive estimates are shown in green, negative estimates in red, and non-significant estimates in gray. The bottom panel indicates which analytical choices were associated with significantly positive (i.e. higher theta-beta ratio (TBR) for attention deficit hyperactivity disorder (ADHD) compared to healthy control, HC) or negative (i.e. lower TBR for ADHD compared to HC) effects across specifications. Note. CI = confidence interval. f=female. EC = eyes closed. EO = eyes open. AVG = average reference. LM = linked mastoid reference. Canonical = canonical frequency range. Relative to IAF = bandwidth relative to individual alpha frequency.

Appendix 3—figure 1—figure supplement 1
Download asset Open asset
Proportion of specifications showing a significantly positive effect by regression term across all analytical specifications.

Each level of condition, reference, frequency definition, comorbid disorder, and medication includes 288 analytical universes. The Aperiodic Signal factor includes 192 universes per level, and the Region of Interest includes 96 universes per level. The red line in the plot indicates the exact proportion of significant results required for the binomial test to reach statistical significance (p<0.05) for each subset size.

Appendix 3—figure 1—figure supplement 2
Download asset Open asset
Proportion of specifications showing a significantly negative effect by regression term across all analytical specifications.

Each level of condition, reference, frequency definition, comorbid disorder, and medication includes 288 analytical universes. The Aperiodic Signal factor includes 192 universes per level, and the Region of Interest includes 96 universes per level. The red line in the plot indicates the exact proportion of significant results required for the binomial test to reach statistical significance (p<0.05) for each subset size.

Tables

Table 1
Demographics information of the healthy brain network sample.
HCCombinedInattentiveΧ²p-value
N228429465
 Male125328311
 Female10310115432.608.33e-8***
Comorbidities
 YesNA338334
 No228911315.420.019*
Medication
 YesNA13165
 No22526938338.555.35e-10***
MeanSDMeanSDMeanSDF-valuep-value
Age (years)10.063.529.522.8910.992.9925.831.1e-11***
SWAN
 IN–0.311.141.250.831.130.84237.612.3e-86***
 HY–0.461.131.200.740.250.86279.051.4e-98***
 Total–0.391.051.230.660.690.69313.253.3e-108***
WISC FSIQ106.1014.9898.3216.1097.9715.3518.948.6e-9***
IAF9.881.029.741.089.961.005.230.006**

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. HC = healthy control. SD = standard deviation. IN = inattentive. IAF = individual alpha frequency.

Table 2
Results of the multiverse analysis in the healthy brain network (HBN) sample showing the proportion of significant effects across 576 universes.
Significantly
positive		Significantly
negative		Not significantTotal
ADHD[IN]0 (0.00%)0 (0.00%)576 (100.00%)576
ADHD[IN] * Age0 (0.00%)9 (1.56%)567 (98.44%)576
ADHD[IN] * Gender[Female]13 (2.26%)1 (0.17%)562 (97.57%)576
ADHD[IN] * IAF15 (2.60%)76 (13.19%)485 (84.20%)576
ADHD[IN] * Age * Gender[Female]0 (0.00%)4 (0.69%)572 (99.31%)576
ADHD[IN] * Age * IAF62 (10.76%)2 (0.35%)512 (88.89%)576
ADHD[IN] * Gender[Female] * IAF40 (6.94%)5 (0.87%)531 (92.19%)576
ADHD[IN] * Age * Gender[Female] * IAF5 (0.87%)10 (1.73%)561 (97.40%)576
ADHD[Comb]11 (1.91%)0 (0.00%)565 (98.09%)576
ADHD[Comb] * Age64 (11.11%)0 (0.00%)512 (88.89%)576
ADHD[Comb] * Gender[Female]61 (10.59%)3 (0.52%)512 (88.89%)576
ADHD[Comb] * IAF4 (0.69%)62 (10.76%)510 (88.54%)576
ADHD[Comb] * Age * Gender[Female]42 (7.29%)25 (4.34%)509 (88.36%)576
ADHD[Comb] * Age * IAF43 (7.46%)1 (0.17%)532 (92.36%)576
ADHD[Comb] * Gender[Female] * IAF10 (1.73%)67 (11.63%)499 (86.63%)576
ADHD[Comb] * Age * Gender[Female] *
IAF		9 (1.56%)85 (14.76%)482 (83.68%)576

  1. Note. Positive universe = higher TBR in ADHD compared to HC. Negative universe = lower TBR in ADHD compared to HC. IN = Inattentive. Comb = Combined. IAF = Individual alpha frequency.

Appendix 1—table 1
Demographics information of the validation sample.
HCCombinedInattentiveΧ²p-value
N1074684
Male453472
Female62121241.361e-9***
MeanSDMeanSDMeanSDF-valuep-value
Age (years)14.494.6411.203.8612.963.0711.351.96e-5***
IQ111.0711.82103.7714.50101.9015.5311.421.87e-5***
IAF9.911.159.571.049.601.132.470.087

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. HC = healthy control. SD = standard deviation. IN = inattentive. HY = hyperactive.

Appendix 1—table 2
Results of the multiverse analysis in the validation sample showing the proportion of significant effects across 144 universes.
Significantly
Positive		Significantly
Negative		Not SignificantTotal
ADHD[IN]2 (1.39%)11 (7.64%)131 (91.00%)144
ADHD[IN] * Age1 (0.69%)16 (11.10%)127 (88.20%)144
ADHD[IN] * Gender[Female]4 (2.78%)1 (0.69%)139 (96.50%)144
ADHD[IN] * IAF0 (0.00%)10 (6.94%)134 (94.10%)144
ADHD[IN] * Age * Gender[Female]1 (0.69%)0 (0.00%)143 (99.30%)144
ADHD[IN] * Age * IAF30 (20.80%)0 (0.00%)114 (79.20%)144
ADHD[IN] * Gender[Female] * IAF7 (4.86%)0 (0.00%)137 (95.10%)144
ADHD[IN] * Age * Gender[Female] * IAF0 (0.87%)5 (3.47%)139 (96.50%)144
ADHD[Comb]1 (0.69%)2 (1.39%)141 (97.90%)144
ADHD[Comb] * Age4 (2.78%)19 (13.20%)121 (84.00%)144
ADHD[Comb] * Gender[Female]0 (0.00%)0 (0.00%)144 (100.00%)144
ADHD[Comb] * IAF0 (0.00%)1 (0.69%)143 (99.30%)144
ADHD[Comb] * Age * Gender[Female]0 (0.00%)8 (5.56%)136 (94.40%)144
ADHD[Comb] * Age * IAF14 (9.72%)6 (4.17%)124 (86.10%)144
ADHD[Comb] * Gender[Female] * IAF2 (1.39%)0 (0.00%)142 (98.60%)144
ADHD[Comb] * Age * Gender[Female] * IAF0 (0.00%)27 (18.80%)117 (81.20%)144

  1. Note. Positive universe = higher TBR in ADHD compared to HC. Negative universe = lower TBR in ADHD compared to HC. IN = inattentive. Comb = combined. IAF = individual alpha frequency.

Appendix 2—table 1
Results of the bootstrap robustness analysis across 1000 resampled multiverse iterations.
Significantly
positive		Significantly
negative		Not significantTotal
ADHD[IN]13.28±3.5 (2.31%)15.86±3.58 (2.75%)546.86±5.13
(94.94%)		576
ADHD[IN] * Age10.17±3.15 (1.77%)17.15±3.77
(2.98%)		548.68±4.92
(95.26%)		576
ADHD[IN] * Gender[Female]20.68±3.33 (3.59%)6.68±2.43
(1.16%)		548.64±4.17 (95.25%)576
ADHD[IN] * IAF23.90±4.75 (4.15%)61.40±6.42 (10.66%)490.70±7.78 (85.19%)576
ADHD[IN] * Age * Gender[Female]4.20±1.99
(0.73%)		6.16±2.25
(1.07%)		565.64±2.96 (98.20%)576
ADHD[IN] * Age * IAF62.80±6.63 (10.90%)30.16±5.03 (5.24%)483.04±8.01 (83.86%)576
ADHD[IN] * Gender[Female] * IAF40.34±4.75 (7.00 %)18.59±4.03 (3.23 %)517.07±6.12 (89.77%)576
ADHD[IN] * Age * Gender[Female] * IAF22.90±4.47 (3.98%)27.34±4.67 (4.75%)525.76±6.46 (91.28%)576
ADHD[Comb]28.03±4.76 (4.87%)4.06±2.00
(0.70%)		543.92±5.13 (94.43%)576
ADHD[Comb] * Age51.01±6.32 (8.86%)6.30±2.44
(1.09%)		518.70±6.80 (90.05%)576
ADHD[Comb] * Gender[Female]56.06±5.82 (9.73%)4.25±1.97
(0.74%)		515.69±6.10 (89.53%)576
ADHD[Comb] * IAF12.01±3.42 (2.09%)46.67±6.07 (8.10%)517.32±6.87 (89.81%)576
ADHD[Comb] * Age * Gender[Female]44.40±4.94 (7.71%)15.09±3.30 (2.62%)516.51±5.95 (89.67%)576
ADHD[Comb] * Age * IAF44.38±5.76 (7.70%)27.80±4.87 (4.83%)503.83±7.10 (87.47%)576
ADHD[Comb] * Gender[Female] * IAF12.76±3.25 (2.22%)64.78±5.87 (11.25%)498.46±6.68 (86.54%)576
ADHD[Comb] * Age *Gender[Female] * IAF9.99±2.80
(1.74%)		87.31±7.41 (15.16%)478.70±7.85 (83.11%)576

  1. Note. Values reported represent the mean across 1000 iterations ± standard deviation, as well as the percentage reflecting the mean value. Positive universe = higher TBR in ADHD compared to HC. Negative universe = lower TBR in ADHD compared to HC. IN = inattentive. Comb = combined. IAF = individual alpha frequency.

Appendix 3—table 1
Results of the multiverse analysis using dimensional SWAN scores instead of categorical attention deficit hyperactivity disorder (ADHD) diagnoses.
Significantly
positive		Significantly
negative		Not significantTotal
SWAN0 (0.00%)21 (3.65%)555 (96.35%)576
SWAN * Age7 (1.21%)20 (3.47%)549 (95.31%)576
SWAN * Gender[Female]65 (11.28%)0 (0.00%)562 (88.72%)576
SWAN * IAF14 (2.43%)134 (23.26%)428 (74.31%)576
SWAN * Age * Gender[Female]9 (1.56%)0 (0.00%)567 (98.44%)576
SWAN * Age * IAF43 (7.47%)0 (0.00%)533 (92.53%)576
SWAN * Gender[Female] * IAF79 (13.71%)5 (0.87%)492 (85.42%)576
SWAN * Age * Gender[Female] * IAF3 (0.52%)67 (11.63%)506 (87.85%)576

  1. Note. Positive universe = higher TBR in ADHD compared to HC. Negative universe = lower TBR in ADHD compared to HC. IAF = individual alpha frequency.

Appendix 4—table 1
Effects of group, age, gender, and IAF on theta-beta ratio (TBR) calculated from 1/f-uncorrected power using canonical frequency bands.
VariableβSECIt-valuep-value
Intercept0.000.10–0.2–0.20.00p=1.000
Age–0.100.09–0.29–0.08–1.09p=0.277
Gender[Female]–0.090.15–0.38–0.2–0.59p=0.557
IAF–0.130.11–0.36–0.09–1.16p=0.246
Age * Gender[Female]0.000.14–0.27–0.270.00p=1.000
Age * IAF0.000.10–0.2–0.20.02p=0.986
Gender[Female] * IAF0.080.16–0.23–0.390.51p=0.611
Age * Gender[Female] * IAF–0.010.13–0.28–0.25–0.11p=0.912
ADHD[IN]0.000.12–0.23–0.23–0.01p=0.995
ADHD[IN] * Age–0.070.11–0.29–0.16–0.58p=0.563
ADHD[IN] * Gender[Female]–0.120.18–0.47–0.24–0.65p=0.516
ADHD[IN] * IAF–0.020.13–0.27–0.24–0.13p=0.898
ADHD[IN] * Age * Gender[Female]0.010.18–0.34–0.360.04p=0.965
ADHD[IN] * Age * IAF0.090.12–0.15–0.320.72p=0.474
ADHD[IN] * Gender[Female] * IAF–0.010.19–0.38–0.37–0.03p=0.972
ADHD[IN] * Age * Gender[Female] * IAF0.010.18–0.35–0.360.04p=0.964
ADHD[Comb]0.140.12–0.09–0.371.23p=0.220
ADHD[Comb] * Age0.130.11–0.09–0.361.18p=0.238
ADHD[Comb] * Gender[Female]–0.110.20–0.49–0.28–0.54p=0.588
ADHD[Comb] * IAF–0.010.13–0.26–0.25–0.05p=0.959
ADHD[Comb] * Age * Gender[Female]–0.130.19–0.51–0.25–0.67p=0.502
ADHD[Comb] * Age * IAF–0.010.12–0.25–0.23–0.09p=0.931
ADHD[Comb] * Gender[Female] * IAF–0.060.19–0.44–0.31–0.33p=0.741
ADHD[Comb] * Age * Gender[Female] * IAF0.060.17–0.28–0.40.37p=0.709

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. In = inattentive. Comb = combined. IAF = individual alpha frequency. SE = standard error. CI = 95% confidence intervals.

Appendix 4—table 2
Effects of group, age, gender, and IAF on theta-beta ratio (TBR) calculated from 1/f-uncorrected power using frequency bands relative to IAF.
VariableβSECIt-valuep-value
Intercept–0.090.07–0.22–0.05–1.22p=0.222
Age–0.240.07–0.37 to –0.11–3.59p=3.43e-04***
Gender[Female]–0.050.10–0.25–0.15–0.49p=0.623
IAF–0.530.08–0.69 to –0.37–6.66p=4.37e-11***
Age * Gender[Female]0.060.10–0.13–0.240.58p=0.559
Age * IAF0.290.070.15–0.434.09p=4.69e-05***
Gender[Female] * IAF–0.030.11–0.25–0.18–0.32p=0.751
Age * Gender[Female] * IAF–0.140.09–0.32–0.04–1.50p=0.135
ADHD[IN]0.040.08–0.12–0.210.54p=0.591
ADHD[IN] * Age0.020.08–0.13–0.180.26p=0.793
ADHD[IN] * Gender[Female]–0.040.13–0.28–0.21–0.28p=0.779
ADHD[IN] * IAF–0.100.09–0.28–0.08–1.14p=0.256
ADHD[IN] * Age * Gender[Female]0.010.13–0.24–0.250.04p=0.965
ADHD[IN] * Age * IAF–0.110.08–0.27–0.05–1.31p=0.189
ADHD[IN] * Gender[Female] * IAF0.070.13–0.2–0.330.49p=0.624
ADHD[IN] * Age * Gender[Female] * IAF0.160.13–0.09–0.411.25p=0.213
ADHD[Comb]0.080.08–0.08–0.241.00p=0.316
ADHD[Comb] * Age0.170.080.02–0.332.18p=0.030*
ADHD[Comb] * Gender[Female]0.020.14–0.25–0.290.18p=0.860
ADHD[Comb] * IAF–0.120.09–0.3–0.06–1.35p=0.177
ADHD[Comb] * Age * Gender[Female]–0.050.14–0.32–0.21–0.39p=0.696
ADHD[Comb] * Age * IAF–0.170.08–0.34 to –0.01–2.06p=0.040*
ADHD[Comb] * Gender[Female] * IAF0.020.14–0.25–0.280.14p=0.888
ADHD[Comb] * Age * Gender[Female] * IAF0.150.12–0.09–0.391.26p=0.209

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. In = inattentive. Comb = combined. IAF = individual alpha frequency. SE = standard error. CI = 95% confidence intervals.

Appendix 4—table 3
Effects of group, age, gender, and IAF on theta-beta ratio (TBR) calculated from aperiodic-adjusted power using canonical frequency bands.
VariableβSECIt-valuep-value
Intercept–0.010.10–0.21–0.18–0.15p=0.882
Age–0.050.10–0.23–0.14–0.50p=0.618
Gender[Female]–0.080.15–0.37–0.21–0.54p=0.591
IAF–0.120.11–0.34–0.1–1.05p=0.292
Age * Gender[Female]–0.020.14–0.28–0.25–0.13p=0.899
Age * IAF–0.060.10–0.26–0.14–0.58p=0.559
Gender[Female] * IAF0.030.16–0.27–0.340.22p=0.825
Age * Gender[Female] * IAF0.100.13–0.16–0.360.73p=0.466
ADHD[IN]0.000.12–0.23–0.230.01p=0.994
ADHD[IN] * Age–0.060.11–0.28–0.16–0.55p=0.585
ADHD[IN] * Gender[Female]–0.050.18–0.41–0.3–0.28p=0.780
ADHD[IN] * IAF–0.090.13–0.35–0.16–0.73p=0.467
ADHD[IN] * Age * Gender[Female]0.070.18–0.28–0.420.40p=0.691
ADHD[IN] * Age * IAF0.150.12–0.08–0.381.26p=0.208
ADHD[IN] * Gender[Female] * IAF0.110.19–0.26–0.490.59p=0.552
ADHD[IN] * Age * Gender[Female] * IAF–0.150.18–0.51–0.2–0.84p=0.402
ADHD[Comb]0.100.12–0.13–0.330.84p=0.400
ADHD[Comb] * Age0.080.11–0.15–0.30.67p=0.503
ADHD[Comb] * Gender[Female]0.030.20–0.35–0.420.18p=0.859
ADHD[Comb] * IAF–0.020.13–0.28–0.23–0.16p=0.873
ADHD[Comb] * Age * Gender[Female]–0.060.19–0.44–0.32–0.32p=0.751
ADHD[Comb] * Age * IAF0.090.12–0.14–0.330.78p=0.434
ADHD[Comb] * Gender[Female] * IAF–0.150.19–0.52–0.23–0.75p=0.452
ADHD[Comb] * Age * Gender[Female] * IAF–0.110.17–0.45–0.24–0.61p=0.544

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. In = inattentive. Comb = combined. IAF = individual alpha frequency. SE = standard error. CI = 95% confidence intervals.

Appendix 4—table 4
Effects of group, age, gender, and IAF on theta-beta ratio (TBR) calculated from aperiodic-adjusted power using frequency bands relative to IAF.
VariableβSECIt-valuep-value
Intercept–0.060.10–0.26–0.13–0.63p=0.528
Age–0.230.09–0.42 to –0.04–2.43p=0.015*
Gender[Female]0.090.15–0.19–0.380.64p=0.522
IAF0.250.110.03–0.472.18p=0.029*
Age * Gender[Female]0.080.14–0.19–0.340.55p=0.580
Age * IAF–0.030.10–0.23–0.16–0.34p=0.736
Gender[Female] * IAF–0.230.16–0.54–0.08–1.47p=0.142
Age * Gender[Female] * IAF0.090.13–0.17–0.350.66p=0.508
ADHD[IN]0.030.12–0.2–0.260.26p=0.793
ADHD[IN] * Age0.010.11–0.22–0.230.05p=0.960
ADHD[IN] * Gender[Female]0.050.18–0.3–0.410.29p=0.773
ADHD[IN] * IAF–0.090.13–0.34–0.17–0.67p=0.502
ADHD[IN] * Age * Gender[Female]0.070.18–0.27–0.420.42p=0.678
ADHD[IN] * Age * IAF0.020.12–0.22–0.250.13p=0.896
ADHD[IN] * Gender[Female] * IAF0.050.19–0.32–0.420.26p=0.797
ADHD[IN] * Age * Gender[Female] * IAF–0.040.18–0.39–0.32–0.21p=0.837
ADHD[Comb]0.120.12–0.11–0.351.00p=0.319
ADHD[Comb] * Age0.210.11–0.02–0.431.82p=0.069
ADHD[Comb] * Gender[Female]–0.120.20–0.5–0.27–0.60p=0.548
ADHD[Comb] * IAF0.040.13–0.22–0.290.29p=0.774
ADHD[Comb] * Age * Gender[Female]–0.150.19–0.52–0.23–0.76p=0.449
ADHD[Comb] * Age * IAF0.030.12–0.2–0.270.27p=0.785
ADHD[Comb] * Gender[Female] * IAF0.050.19–0.32–0.430.29p=0.775
ADHD[Comb] * Age * Gender[Female] * IAF–0.130.17–0.47–0.21–0.75p=0.451

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. In = inattentive. Comb = combined. IAF = individual alpha frequency. SE = standard error. CI = 95% confidence intervals.

Appendix 4—table 5
Effects of group, age, gender, and IAF on theta-beta ratio (TBR) calculated from aperiodic signal power using canonical frequency bands.
VariableβSECIt-valuep-value
Intercept0.130.09–0.05–0.321.40p=0.162
Age–0.200.09–0.37 to –0.02–2.20p=0.028*
Gender[Female]–0.300.14–0.57 to –0.03–2.15p=0.032*
IAF–0.270.11–0.48 to –0.06–2.51p=0.012*
Age * Gender[Female]0.010.13–0.24–0.260.06p=0.953
Age * IAF0.100.10–0.09–0.281.02p=0.309
Gender[Female] * IAF0.240.15–0.04–0.531.66p=0.098
Age * Gender[Female] * IAF–0.250.12–0.49 to –0.01–2.01p=0.045*
ADHD[IN]–0.010.11–0.23–0.2–0.11p=0.911
ADHD[IN] * Age–0.080.11–0.29–0.13–0.75p=0.456
ADHD[IN] * Gender[Female]–0.330.17–0.66–0–1.95p=0.052
ADHD[IN] * IAF0.160.12–0.08–0.41.33p=0.185
ADHD[IN] * Age * Gender[Female]–0.080.17–0.41–0.25–0.49p=0.622
ADHD[IN] * Age * IAF–0.030.11–0.25–0.18–0.31p=0.757
ADHD[IN] * Gender[Female] * IAF–0.220.18–0.57–0.13–1.23p=0.220
ADHD[IN] * Age * Gender[Female] * IAF0.330.170–0.661.93p=0.053
ADHD[Comb]0.100.11–0.12–0.310.90p=0.367
ADHD[Comb] * Age0.020.11–0.19–0.230.15p=0.882
ADHD[Comb] * Gender[Female]–0.270.18–0.63–0.09–1.49p=0.136
ADHD[Comb] * IAF0.130.12–0.11–0.371.07p=0.283
ADHD[Comb] * Age * Gender[Female]–0.030.18–0.38–0.33–0.14p=0.886
ADHD[Comb] * Age * IAF–0.140.11–0.36–0.08–1.27p=0.205
ADHD[Comb] * Gender[Female] * IAF–0.130.18–0.48–0.23–0.71p=0.475
ADHD[Comb] * Age * Gender[Female] * IAF0.430.160.11–0.752.66p=0.008**

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. In = inattentive. Comb = combined. IAF = individual alpha frequency. SE = standard error. CI = 95% confidence intervals.

Appendix 4—table 6
Effects of group, age, gender, and IAF on theta-beta ratio (TBR) calculated from aperiodic signal power using frequency bands relative to IAF.
VariableβSECIt-valuep-value
Intercept–0.140.06–0.25 to –0.03–2.54p=0.011*
Age–0.140.05–0.24 to –0.04–2.65p=0.008**
Gender[Female]–0.030.08–0.19–0.13–0.34p=0.737
IAF–0.690.06–0.81 to –0.56–10.80p=6.23e-26***
Age * Gender[Female]0.040.08–0.11–0.190.56p=0.576
Age * IAF0.330.060.22–0.445.80p=8.66e-09***
Gender[Female] * IAF0.110.09–0.06–0.281.27p=0.205
Age * Gender[Female] * IAF–0.170.07–0.32 to –0.02–2.29p=0.022*
ADHD[IN]0.080.07–0.05–0.211.23p=0.220
ADHD[IN] * Age0.040.06–0.08–0.170.67p=0.505
ADHD[IN] * Gender[Female]–0.060.10–0.26–0.14–0.57p=0.567
ADHD[IN] * IAF–0.110.07–0.25–0.03–1.55p=0.121
ADHD[IN] * Age * Gender[Female]–0.020.10–0.22–0.18–0.20p=0.841
ADHD[IN] * Age * IAF–0.090.07–0.22–0.04–1.31p=0.190
ADHD[IN] * Gender[Female] * IAF0.080.11–0.13–0.290.76p=0.446
ADHD[IN] * Age * Gender[Female] * IAF0.060.10–0.14–0.260.61p=0.545
ADHD[Comb]0.100.07–0.03–0.231.50p=0.134
ADHD[Comb] * Age0.100.06–0.03–0.221.49p=0.136
ADHD[Comb] * Gender[Female]0.140.11–0.08–0.351.25p=0.213
ADHD[Comb] * IAF–0.160.07–0.3 to –0.01–2.14p=0.033*
ADHD[Comb] * Age * Gender[Female]0.070.11–0.15–0.280.61p=0.540
ADHD[Comb] * Age * IAF–0.130.07–0.26–0–1.96p=0.051
ADHD[Comb] * Gender[Female] * IAF–0.110.11–0.32–0.1–1.00p=0.317
ADHD[Comb] * Age * Gender[Female] * IAF0.060.10–0.13–0.250.62p=0.536

  1. *p<0.05. **p<0.01. ***p<0.001.

  2. Note. In = inattentive. Comb = combined. IAF = individual alpha frequency. SE = standard error. CI = 95% confidence intervals.

Additional files

MDAR checklist
https://cdn.elifesciences.org/articles/111114/elife-111114-mdarchecklist1-v1.docx

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Dawid Strzelczyk
  2. Andrea Vetsch
  3. Nicolas Langer
(2026)
Theta beta ratio in attention deficit hyperactivity disorder using a multiverse analysis
eLife 15:RP111114.
https://doi.org/10.7554/eLife.111114.3