Secondary exploratory discrete wavelet analysis of the electroretinogram in adults with ADHD
Marie-Soleil Vallière1, Marc-André Dubois2, Marc Hébert3
1CERVO Brain Research Centre, Centre intégré universitaire de santé et de services sociaux de la Capitale-Nationale, Quebec, QC, Canada; School of Psychology, Faculty of Social Sciences, Université Laval, Quebec, QC, Canada.
Background:
Conventional electroretinographic (ERG) analyses identified sex-specific retinal differences in adults with attention-deficit/hyperactivity disorder (ADHD). We applied the discrete wavelet transform (DWT) to the same recordings to characterize multiscale time-frequency features.
Methods:
ERGs from 25 adults with ADHD and 25 age- and sex-matched controls (16 female and 9 male pairs) were analyzed at seven flash intensities. Six DWT descriptors were extracted, and global scores averaged across intensities. Paired comparisons used false-discovery-rate correction. Secondary analyses examined intensity, sex, eyes, inter-eye agreement, within-session repeatability, current medication status, and exploratory classification.
Results:
Global 40a and 20b values were lower in ADHD (Cohen's dz. = -0.63 and - 0.60; FDR q = 0.020 for both). Differences increased with flash intensity and had the same direction in female and male pairs and both eyes, without significant group × sex or group × eye interactions. Findings were stable across participant- and intensity based sensitivity analyses. Inter-eye ICC(A,1) was 0.750 for 40a and 0.865 for 20b; bilateral within-session ICC(A,1) values ranged from 0.839 to 0.981. Reductions were also present among the eight ADHD participants reporting no current ADHD medication use (dz = -0.74 and - 0.84). Exploratory leave-one-pair-out AUCs ranged from 0.653 to 0.731.
Conclusions:
DWT analysis identified lower 40a and 20b values as candidate retinal time-frequency features of ADHD. Their consistency across sensitivity analyses, sexes, eyes, medication strata, and repeated measurements supports prospective evaluation in independent, clinically characterized cohorts.


