Heatmaps of damage scores (0-3) across the five identified types of damage and across the four imaged lesioning arrays.

Electrodes are displayed using the orientation in Figure 1 (electrode tips facing viewer, wire bundle on the right). Second-to-rightmost column displays summed damage scores for each array across the five types of damage. Electrodes used for electrolytic lesioning are denoted with blue dots. Median summed scores for each radial section of the array are plotted in the bar charts to the right of the heatmaps. Ring layout and numbering information are available in Supplementary Figure 1. Unwired electrodes (electrodes not wire bonded at time of manufacture) and electrodes with shank fractures are ignored and displayed in black, as they are not scored.

Stacked histograms of damage scores (0-3) across the five identified types of damage and across the four imaged lesioning arrays.

Gray indicates normal electrodes, and black indicates lesioning electrodes. The rightmost column displays the average distribution of damage scores across the five types of damage. Electrodes with shank fractures (SF) are ignored, as they are not scored.

Correlation plot (Pearson’s R) across the five different damage types.

AD = abnormal debris, TB = tip breakages, CC = metal coating cracks, PC = parylene C cracks, PD = parylene C delamination. Results demonstrate overall low correlation (magnitude < 0.25) across different damage types, with the exception of coating cracks and tip breakage, with a correlation coefficient of 0.47. Test values with Bonferroni-corrected p < 0.05 are displayed with asterisks. Raw p-values are separately available in Supplementary Table 2. Electrodes with shank fractures are ignored, as they are not scored.

Numbering layout of the electrodes as imaged (pins facing outwards, towards the reader).

Wire bundle is arranged to the right. Each radial section of the array is also color-coded and labeled. R1 refers to the innermost core electrodes of the array, R2 refers to the next outer ring of the array, and so on.

Heatmaps of damage scores (0-3) across the five identified types of damage and across all eight intact, imaged NHP arrays.

Electrodes are displayed using the orientation in Figure 1 (electrode tips facing viewer, wire bundle on the right). Second-to-rightmost column displays summed damage scores for each array across the five types of damage. Electrodes used for electrolytic lesioning are denoted with blue dots. Median summed scores for each radial section of the array are plotted in the bar charts to the right of the heatmaps. Ring layout and numbering information is available in Supplementary Figure 1. Unwired electrodes (electrodes not wire bonded at time of manufacture) and electrodes with shank fractures are ignored and displayed in black, as they are not scored.

Stacked histograms of damage scores (0-3) across the five identified types of damage and across all eight intact, NHP imaged arrays.

Gray indicates normal electrodes, and black indicates lesioning electrodes. Rightmost column displays the average distribution of damage scores across the five types of damage. Electrodes with shank fractures (SF) are ignored, as they are not scored.

Correlation plots (Pearson’s R) for each of the five rings across all four imaged lesioning arrays.

Test values with Bonferroni-corrected p < 0.05 are displayed with asterisks. Raw r and p-values are separately available in Supplemental Table 5.

Scatterplots of the lesion electrodes’ damage scores over the target current for each lesion.

A best-fit line is also plotted along with regression results. In cases where only one target current was used (ie. subjects with only one lesion), linear regression was not performed. No best-fit lines were statistically significant at p = 0.05.

Scatterplots of the lesion electrodes’ damage scores over the target lesioning procedure duration for each lesion.

A best-fit line is also plotted along with regression results. In cases where only one duration was used (ie. subjects with only one lesion, subjects with the same duration across all lesions), linear regression was not performed. Scatterplots with statistically significant lines of best fit are starred in yellow.

Histograms of the lesion electrodes’ damage scores, separated by array material.

Blue bars are IrOx arrays and orange bars are Pt arrays.The first row shows scores across all eight scored NHP arrays (n=384 IrOx electrodes, n=384 Pt electrodes). The second row shows scores for only the four NHP arrays used for lesioning (n=192 IrOx electrodes, n=192 Pt electrodes). The third row shows scores for only the subset of electrodes used for NHP lesioning (n=10 IrOx electrodes, n=20 Pt electrodes). Notably, it appears that platinum arrays imaged in this work generally appear to report less damage than iridium oxide arrays when considering all arrays. Additionally, the imaged platinum arrays appear to always report less Parylene C damage. However, platinum arrays generally appear to report more other types of damage (abnormal debris, tip breakage, coating cracks) when used for lesioning. This supports previous literature that indicates IrOx may be more resistant to stimulation-related damage than Pt [26, 41–43]. It is important to note that many other factors, such as time in tissue, explant surgery, and immune response of the specific implant subject, all impact the conclusions of this analysis.

Details and characteristics for all imaged and analyzed devices in this work.

Shapiro-Wilkes p-values for different electrode populations.

Mann-Whitney U and Levene test p-values when comparing scores between lesioning electrodes and non-lesioning elec-trodes on the same array.

Mann-Whitney U and Levene test p-values when comparing scores between all electrodes on arrays used and not used for lesioning experiments when implanted in the same subject.

Bonferroni-corrected p-value is 0.005 (0.05/n=10 comparisons).

Lesion parameters for NHP lesions analyzed in this work.