1. The classical view

Pre-specified TBR pipeline N = 25 (after artefact rejection)

The theta/beta ratio is the closest thing pediatric EEG has to a household name. The original idea is straightforward. Frontal theta power, around 4 to 8 Hz, tracks under-arousal of cortical control circuits. Frontal beta power, around 13 to 30 Hz, tracks active engagement. The ratio of the two, computed from the resting-state spectrum at midline electrodes Fz, Cz, and Pz, is supposed to be a stable trait marker for ADHD and for executive dysfunction more broadly. Lubar’s 1991 monograph laid the foundation (Lubar, 1991). The neurofeedback community built clinical protocols on top.

The published evidence is more complicated than the protocol literature lets on. Snyder and Hall’s 2006 meta-analysis (Snyder & Hall, 2006) reported substantial between-study variance in the diagnostic accuracy of TBR. Arns, Conners, and Kraemer’s follow-up in 2013 (Arns et al., 2013) looked at a decade of studies and concluded the effect exists, but it shrinks as study quality and sample size increase. The paediatric sub-literature carries the same caveat: the ratio has predictive power on average, but the individual-level variance is large enough that the effect sometimes inverts at the cohort level.

I knew that going in. The cohort I was working with had already produced one related publication, an ERP study correlating P300 latency with digit-span performance in a similarly aged sample (Dewi et al., 2025), so the developmental-EEG pipeline was already running. When I designed the Biomarker_IIUM extension, the explicit hypothesis was that TBR at frontal midline would correlate with the executive-function task performance the children completed alongside the EEG. The pre-specified statistical plan listed eight correlation hypotheses linking conventional QEEG features (frontal TBR, frontal-asymmetry-of-alpha, alpha reactivity) to behavioural outcomes (a Bahasa-Indonesia executive-function battery called AUFEI plus Flanker and Digit Span). All eight were Spearman correlations, alpha = 0.05, FDR-corrected with Benjamini-Hochberg (Benjamini & Hochberg, 1995) across the eight hypotheses.

The pipeline was straightforward: HAPPE for artefact rejection (Gabard-Durnam et al., 2018), independent-component analysis for ocular and muscular artefacts, average-reference re-referencing, Welch power spectral density at 2-second windows with 50% overlap, theta and beta band power summed in the standard windows, ratio computed per channel. After artefact rejection, three children had insufficient clean frontal-midline data, leaving N = 25 for the TBR analyses (the full N = 28 cohort survived for the other features).

Figure 1. The theta/beta ratio formula and the cohort’s individual TBR values plotted against a published expected mean for paediatric typically-developing samples. The plot shows that per-child variance is large enough that several children at the top of the executive-function task fall in the TBR range that the published literature associates with elevated risk.

The figure above shows what the cohort looks like before the statistical test. There is a TBR effect in the raw means. The next section shows what happens to it under correction.

→ Continue to Section 2: The pivot

References

Arns, M., Conners, C. K., & Kraemer, H. C. (2013). A decade of EEG theta/beta ratio research in ADHD: A meta-analysis. Journal of Attention Disorders, 17(5), 374–383. https://doi.org/10.1177/1087054712460087
Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B, 57(1), 289–300.
Dewi, S. Y., Kusuma, N. K., Wahyuningsih, S., Habiburahman, Y., Listianto, H., Idhsa, Y. A., Rahmawati, G. D., & Rahman, A. W. A. (2025). Correlation between P300 latency and digit span performance in primary school children: A cross-sectional ERP study. 2025 13th International Conference on Orange Technology (ICOT), 1–8. https://doi.org/10.1109/ICOT68409.2025.11425349
Gabard-Durnam, L. J., Méndez Leal, A. S., Wilkinson, C. L., & Levin, A. R. (2018). The Harvard Automated Processing Pipeline for EEG (HAPPE): Standardized processing software for developmental and high-artifact data. Frontiers in Neuroscience, 12, 97. https://doi.org/10.3389/fnins.2018.00097
Lubar, J. F. (1991). Discourse on the development of EEG diagnostics and biofeedback for attention-deficit/hyperactivity disorders. Biofeedback and Self-Regulation, 16(3), 201–225. https://doi.org/10.1007/BF01000016
Snyder, S. M., & Hall, J. R. (2006). A meta-analysis of quantitative EEG power associated with attention-deficit hyperactivity disorder. Journal of Clinical Neurophysiology, 23(5), 441–456. https://doi.org/10.1097/01.wnp.0000221363.12503.78