Assessing brain-behaviour coupling in non-invasive brain stimulation: Insights from reliable change indices and Monte...

Non-invasive brain stimulation (NiBS) studies frequently report exploratory brain-behaviour correlations following the application of NiBS. However, these analyses are typically underpowered and rely on ratio-based change scores with known statistical limitations. We addressed these limitations by pooling individual data from three independent studies (N = 69) and combining empirical analysis with a Monte Carlo simulation framework. All studies applied 20 Hz transcranial alternating current stimulation (tACS) targeting the pre-supplementary motor area (preSMA) and right inferior frontal gyrus (rIFG) during a stop signal task. Changes in preSMA-rIFG EEG imaginary coherency (ImCoh) and stop-signal reaction time (SSRT) were quantified using reliable change indices (RCI), z-standardised within studies for pooled mixed-effects regression. We found no meaningful association between tACS-induced ImCoh change and SSRT change (r = .013, marginal R² = .004), with project-wise correlations being small, non-significant, and directionally inconsistent. While sensitivity analysis using ratio-based change scores converged on the same null result (r = .014), the simulation framework revealed that ratio-based scores show near-complete failure to recover true brain-behaviour associations across all plausible effect sizes. Conversely, RCI-based scores preserved directional accuracy and provided well-calibrated Type I error control, despite moderate attenuation under realistic noise conditions. These results highlight the severe statistical limits of detecting individual-level brain-behaviour coupling following 20 Hz tACS. Rather than ruling out a true biological association, our findings demonstrate that standard sample sizes and typical measurement noise create substantial 'blindness' to realistic effect sizes, underscoring a broader need for more reliable metrics and adequately powered designs in NiBS research.