Phenotype-specific medication-related changes in nonlinear Poincaré descriptors of postural sway in Parkinson’s disease

Postural instability is a major source of disability and fall risk in Parkinson’s disease (PD), yet the effect of dopaminergic medication on balance remains inconsistent. One possible explanation is clinical heterogeneity, particularly between tremor-dominant (TD) and postural-instability/gait-difficulty (PIGD) phenotypes. In this study, nonlinear Poincaré descriptors were calculated from postural sway data to examine whether they changed with medication state differently in these two phenotypes. In a public dataset of 32 individuals with idiopathic PD, 31 participants were classified as PIGD (n=12) or TD (n=19). Quiet-standing center-of-pressure (CoP) signals were examined in OFF and ON medication states under four stance conditions combining rigid or unstable surface with eyes open or closed. Pre-filtered CoP signals were downsampled to 25 Hz and analyzed separately in the anteroposterior and mediolateral directions using lag-1 Poincaré plots to obtain SD1, SD2, and SD1/SD2, which were also compared with conventional, entropy-based, and fractal measures. Medication-related changes tended to be larger and more often positive in the PIGD group, whereas TD patients showed little systematic change. The clearest between-phenotype differences (nominal p<.05) were observed for mediolateral SD1 on an unstable surface with eyes closed, mediolateral SD2 on a rigid surface with eyes closed, and the anteroposterior SD1/SD2 ratio on an unstable surface with eyes closed; given the number of comparisons and the modest sample, these findings should be regarded as exploratory. These findings suggest that motor phenotype should be considered when interpreting medication effects on posturography in PD, and that the SD1/SD2 ratio, which closely tracked entropy- and fractal-based measures, offers a simple, parameter-free index of medication-related changes in the temporal structure of sway.