Skip to main content
[Preprint]. 2023 Aug 8:2023.08.03.23293450. [Version 1] doi: 10.1101/2023.08.03.23293450

Figure 4. Data-driven biomarker identification for all hemispheres.

Figure 4.

a-b, Results of the within-subject nonparametric cluster-based permutation analysis for in-clinic recordings. a, Graphs illustrate the effect size (Cohen’s d) of the main effect of medication on power as a function of frequency in the STN (left), cortical montage 1 (middle), and cortical montage 2 (right) for all patients. Red and blue colors represent positive effects (medication on > medication off) and negative effects (medication off > medication on), respectively. For all patients, we found finely-tuned gamma (FTG) oscillations in the STN (pat-2, both hemispheres) and cortex (pat-1, left hemisphere and pat-3, right hemisphere) to be the optimal biomarker for medication-related fluctuations during active stimulation (Extended Data Fig. 2). We did not find any significant effects in the left hemisphere of pat-3. b, The effect sizes for cortical and STN FTG oscillations (right) were superior to those for STN beta oscillations (left) for all patients (mean±standard error of the mean across permutations). c-e, Results of the within-subject linear discriminant analysis for at-home recordings using power spectral density at the three brain sites to predict the occurrence of the most bothersome symptom. c, The three graphs illustrate the initial area under the curve (AUC) prior to bandwidth optimization as a function of frequency for the STN (left), cortical montage 1 (middle), and cortical montage 2 (right) for all patients. Across patients, we show that FTG oscillations in the STN (pat-2, both hemispheres) and cortex (pat-1, left hemisphere and pat-3, both hemispheres) were the best predictors of the occurrence of the most bothersome symptom and superior to beta oscillations (d, mean±standard error of the mean across permutations, Extended Data Fig. 3). e, The combined use of STN/cortical gamma and STN beta bands provided minimal improvement in the AUC of at-home symptom prediction using linear discriminant analysis (mean±standard error of the mean across permutations).