Article Figures & Data
Figures
- Figure 1.
Schematic summarizing waveform shape calculations. A, First rising and falling zerocrossings were identified from the β-filtered signal. Then, peaks and troughs that lie between adjacent zerocrossings were found in the raw signal. B, Peak sharpness was calculated as the mean of the difference between the peak (purple circle) and the voltage points three samples before and after the peak (purple triangles). The trough sharpness was calculated in a similar fashion, indicated with blue circles and triangles similar to the peaks analysis. The yellow area, between a trough and subsequent peak, indicates an area where rise steepness is determined, while the green area, between a peak and subsequent trough, indicates an area where decay steepness was found. The maximum slope (shown with arrows) for each of these shaded regions (red circles) was derived as rise and decay steepness and then the ratio was taken to find steepness ratio, as described in Materials and Methods. Figure modeled after Cole et al. (2017), but here is shown with the present EEG data. C, Average power spectral density plots for each group. Plots confirm that there is a β oscillation as well as a prominent α/μ oscillation (each indicated with black arrow).
- Figure 2.
Sharpness ratio, steepness ratio, and PAC all decreased with medication in PD patients. A, Box plots of each measure for patients off medication, on medication, and control participants. Asterisk represents FDR corrected significance at p < 0.05. Note that the on versus off medication comparison was a paired comparison (Wilcoxon sign rank) whereas the off medication versus control comparison was unpaired (Wilcoxon rank sum). B, Individual data for each measure and each patient on and off medications. Each color corresponds to a different participant. Diagonal lines represent unity. C, Scalp topography for on versus off medication for each measure. For each electrode, test statistics below the critical value, indicating significance, were set to zero. Test statistics above the critical value were rescaled and normalized to the percentage of the most significant test statistic, such that 100 would reflect the most significant test statistic.
- Figure 3.
Both sharpness ratio and steepness ratio over sensorimotor cortex correlate with PAC in both medication states.
- Figure 4.
Sensorimotor cortex recordings have a canonical shape. A, Peak-to-trough ratio versus rise-to-fall ratio over sensorimotor electrodes (C3 and C4) shown with individual participants/electrodes. Circles represent off medication, while squares represent on medication. The color map corresponds to PAC values. Only patient data are shown, but control data followed a similar pattern. B, Peak-to-trough versus rise-to-fall ratio in electrodes closest to temporalis muscles (F7 and F8) shown separately. Note the scale of the PAC color map is different from in A. C, Representative waveform shape for each quadrant. The sensorimotor data falls mainly in Q4 (blue). This corresponds to sharper peaks and steeper decays.
Tables
Frequency range Metric Test statistic (signed rank) p value (FDR corrected) Effect size (d) μ/α (8–12 Hz) Sharpness ratio 20 0.023 (0.0278) 0.70 Steepness ratio 12 0.006 (0.0121) 0.60 PAC 30 0.088 (0.0881) 0.51 β (13–30 Hz) Sharpness ratio 12 0.006 (0.0121) 0.86 Steepness ratio 9 0.004 (0.0121) 0.68 PAC 15 0.011 (0.0166) 0.83 Values are shown for both μ/α and β based on a priori hypotheses (β) and the average power spectral density plots (μ/α and β; Fig. 1C). Multiple comparisons correction (FDR) is applied including correction for all comparisons in the table.
Frequency range Metric Test statistic (rank sum) p value (FDR corrected) Effect size (d) μ/α (8–12 Hz) Sharpness ratio 1.03 0.304 (1) 0.28 Steepness ratio 2.53 0.011 (0.0331) 0.87 PAC 2.17 0.030 (0.0602) 0.76 β (13–30 Hz) Sharpness ratio 0.91 0.3631 (1) 0.19 Steepness ratio 2.53 0.011 (0.0331) 0.88 PAC 1.19 0.236 (1) 0.27 Values are shown for both μ/α and β, as in Table 1. Multiple comparisons correction (FDR) is applied including all comparisons in the table.
- Table 3.
Effect size (Cohen’s d) for all patients and for Q4 patients, on versus off medication
Metric All Q4 Sharpness ratio 0.86 0.95 Steepness ratio 0.68 0.70 PAC 0.75 0.83 All metrics were log scaled.
Extended Data 1
Code used for this manuscript. Also available at https://github.com/SwannLab/Jackson_2019. Download Extended Data 1, ZIP file.
In this issue
