Comparative Spatiotemporal Dynamics of Access Resistance, Polarization Impedance, and Nested Bipolar Impedance in Standard and Dexamethasone-Eluting Electrode Arrays

Abstract

Introduction:

Intracochlear injury during surgical placement of a cochlear implant (CI) electrode array (EA) and the immune response to a foreign body can result in fibrotic scar formation and neo-ossification that may compromise device performance. Dexamethasone-eluting electrode arrays (Dex-EAs) have been developed to reduce the inflammatory response. Electrode impedance measurements provide a noninvasive biomarker of the intracochlear environment. Nested bipolar (BP) impedance, which uses only intracochlear electrodes for both stimulation and recording, may be particularly sensitive to local tissue changes. This study compared the spatial and temporal profiles of nested bipolar impedances with the profiles of access resistance (Ra) and polarization impedance (Zp), which are derived from measurements that use intra- and extracochlear electrodes, in recipients of standard EAs or Dex-EAs.

Methods:

Thirty-six adult CI recipients were included (17 control 632 EAs; 19 experimental 632D Dex-EAs). Impedance measurements were obtained at initial activation (IA) and at 2 weeks, 1 month, 3 months, 6 months and 12 months after initial activation. Ra and Zp were derived from transimpedance matrix measurements using a monopolar configuration, while nested bipolar impedance was obtained using a four-electrode intracochlear configuration. Impedance profiles were analyzed across electrodes and time to assess spatial and temporal trends.

Results:

Across all 3 impedance measures, all 22 electrodes, and all 6 post-operative time points, Dex-EAs consistently exhibit lower impedance values compared to controls. Even so, Zp decreases over time in both groups (though to a lesser extent for Dex-EAs) as does Ra for apical sites. When focusing on the region of highest interest - the base - different patterns are evident. In control subjects, Ra and nested bipolar impedance progressively increase over the 12 months, first noticeable at the most extreme basal sites and involving an expanding number of more apical sites with time. The changes are significantly more pronounced for nested BP impedance than Ra (Paired T-test: p = 0.001 at 6 months, p < 0.001 at 12 months). This particular pattern is specific to control ears.

Conclusion:

Dex-EAs are associated with lower and more stable impedance profiles over time. Nested bipolar impedance demonstrates greater sensitivity to intracochlear changes compared to Ra or Zp, supporting its use as a focused biomarker of the intracochlear environment in future studies.

1. Introduction:

Cochlear implantation is the standard-of-care treatment for individuals with severe-to-profound sensorineural hearing loss. Since the initial adoption of cochlear implants (CIs), substantial efforts have focused on improving patient outcomes through advances in CI design and signal processing, improved preoperative planning using imaging technology, refinement of surgical techniques including robotic-assisted electrode array (EA) insertion, and by improving understanding of the biological response of cochlear tissues to implanted biomaterials^1–12. Despite these advances, placement of an EA can result in mechanical trauma and elicit a foreign body response, leading to fibrotic tissue formation and, in some cases, new bone growth within the cochlea^13–15.

Such intracochlear tissue remodeling can compromise residual acoustic hearing and reduce the effectiveness of electrical stimulation.

Glucocorticoids have therefore been widely used in the perioperative period, administered systemically or locally, to attenuate inflammatory responses associated with EA insertion and subsequent fibrotic scar formation^16–19. Building on this approach, dexamethasone-eluting electrode arrays (Dex-EAs) have emerged as a promising strategy for prolonged intracochlear dexamethasone delivery^3,20–25. Compared with systemic or local approaches to steroid administration, Dex-EAs provide several potential advantages: they enable drug release in close proximity to the electrodes, extend the time period of steroid exposure beyond the immediate perioperative window, and offer a more consistent delivery profile with respect to timing and location across recipients. Importantly, the attenuated inflammatory response observed with Dex-EA persists, in contrast to local administration during surgery²⁰.

Electrode impedance measurements reflect the opposition to current flow between electrodes and are commonly used as a biomarker to assess intracochlear tissue changes, including fibrosis^22,26–33. Prior studies have demonstrated reduced electrode impedance values and intra-day impedance fluctuations in recipients implanted with Dex-EAs relative to standard EAs³⁴. These differences have been observed as early as one week following surgery and persist across multiple postoperative intervals, including initial activation (approximately two weeks postoperatively) as well as one, three, six, 12, and 24 months after initial activation (IA). More recently, our group demonstrated that Dex-EAs are associated not only with lower impedance values, but also more stability across repeated measurements during the first 90 days following IA compared with standard EAs²⁵. During an impedance measurement, a known current is delivered between specified elements of the device—intracochlear or extracochlear contacts—and the resulting voltage response is recorded across two points. Impedance is then calculated as the ratio of measured voltage to stimulation current. The interpretation of electrode impedance with respect to the intracochlear environment depends on multiple factors, two key factors being (1) the temporal portion of the voltage response used for impedance calculation^30,35,36 and (2) the specific cochlear implant components included in the measurement³³.

Clinical impedance measures, which are predominantly used to identify device deficiencies (open and short circuits) and to determine the ceiling compliance limits for programming current levels, have historically used an extracochlear reference point for stimulation and recording. This monopolar montage encompasses a relatively broad area, which includes the cochlea and also surrounding conductive tissue (Fig. 1). A bipolar montage minimizes the influence of extracochlear tissue by narrowing the current field and recording region within the cochlea. In a nested bipolar configuration, two adjacent intracochlear electrodes are used to record voltage, while flanking intracochlear electrodes on each side are used for current delivery, resulting in a measurement that utilizes four intracochlear electrodes^23,24,33(Fig. 1). Since both stimulation and recording occur entirely within the cochlea, nested bipolar impedance is less affected by extracochlear factors and is thought to be more sensitive to local intracochlear tissue changes along the EA.³³ Several studies have investigated these measurements in both animal²³ and human^24,33 subjects. As expected, nested BP impedance is lower in recipients of Dex-EAs compared to standard EAs^23,24. Although not directly correlated to date, the profiles of nested BP impedance across electrodes for standard EAs^24,33 are reminiscent of the profiles of fibrous tissue/bone along the cochlear duct seen in post-CI cadaveric temporal bone studies³⁷. Direct correlation between nested bipolar impedance profiles and histopathologic findings should be addressed in future work.

Figure 1.

Three-dimensional reconstruction of a cochlear implant from a post-operative computed tomography scan to illustrate the smaller distance between active and reference electrodes in a nested bipolar impedance configuration compared to a monopolar impedance configuration. (S_AC: stimulating active, S_RE: stimulating reference, R_AC: recording active, R_RE: recording reference). The four most basal electrodes (1–4) are labelled. EC: extracochlear case.

In this study, we compared nested bipolar impedance measurements between subjects implanted with either a standard Cochlear 632 EA or the experimental 632D Dex-EA. In addition to replicating prior studies,^23,24,33 this report expands our knowledge by the presentation of nested bipolar impedance with high spatial resolution—at the level of individual electrodes in individual subjects—and temporal resolution (six measurements from IA to 12 months). This individual electrode-level analysis is critical, as impedance values are not uniform along the electrode array and despite some commonly observed profile patterns, can also vary widely across individuals. Likewise, impedance values vary across time, reflecting the evolving inflammatory fibrotic response.

Our laboratory has previously performed a subcomponent analysis of monopolar impedance using the time domain to separate access resistance (Ra) from polarization impedance (Zp)^30,35,36. The initial, rapid rise in voltage following the onset of a current pulse primarily represents the resistance to current flow from the environment surrounding the electrode (access resistance; Ra)³⁰. In contrast, the subsequent, slower change in voltage across the duration of the pulse reflects the capacitive and electrochemical processes at the electrode–tissue interface (polarization impedance; Zp)³⁰. This subcomponent analysis (Ra and Zp) has been used in multiple studies to probe the electrode–environment interface. For example, in a study evaluating impedance changes in electro-acoustic stimulation CI users with delayed loss of acoustic hearing, patients who experienced hearing loss demonstrated increases in access resistance over time, while polarization impedance remained stable³⁸. In contrast, subjects with stable hearing showed stable access resistance, accompanied by a decline in polarization impedance. Collectively, these findings demonstrate that the two impedance components likely reflect distinct physiological processes. Due to the availability of the Ra and Zp data for this data set, these impedance profiles and their evolution will also be compared with nested bipolar impedance.

Here we show that all three impedance measurements—nested bipolar impedance, Ra, and Zp—are lower in ears with Dex-EAs compared to ears with standard EAs. We further demonstrate that nested bipolar impedance is more sensitive to spatial and temporal impedance dynamics than Ra and Zp given the spatially confined stimulation/recording region of the nested bipolar measurements.

2. Materials & Methods:

2.1. Study population

The study was conducted according to guidelines for the protection of human subjects as set forth and approved by the Institutional Review Board (IRB) at the University of Iowa (IRB: 201805740, 202210440). Adult CI candidates were recruited from the University of Iowa Healthcare Department of Otolaryngology—Head and Neck Surgery clinics. Seventeen subjects were included in the control group and 19 in the experimental group. Eight total subjects (4 control 632 EA and 4 experimental 632D Dex-EA) were part of a Cochlear^® multicenter clinical trial (NCT04750642) and were randomized to their groups. The remaining 15 subjects in the experimental group were enrolled in a single-center clinical trial conducted at Iowa that investigated the same device. These 15 experimental subjects and the remaining 13 independently recruited control subjects were assigned to groups in a nonrandomized manner, based on the availability and sterilization schedule of the investigational implants. Exclusion criteria were as follows: auditory neuropathy, anatomic malformations or involvement of the cochlea or nerve, history of bacterial meningitis, large vestibular aqueduct syndrome, otosclerosis, active middle ear infection, current use of tympanostomy tubes, unhealed tympanic membrane perforation, and known allergy to dexamethasone or other corticosteroids.

2.2. Surgeries

Subjects in the control group were implanted with the commercially available slim perimodiolar 632 cochlear implant arrays (Cochlear Limited, Australia). Subjects in the experimental group received the investigational dexamethasone-eluting 632D array (Cochlear Limited), which is based on the standard 632 design and incorporates drug-eluting silicone wells distributed along the apical 16 electrodes (electrodes 7–22). All subjects underwent unilateral cochlear implantation performed by one of three faculty neurotologists using a standard facial recess approach, with array insertion via the round window, extended round window, or cochleostomy. Approach was not chosen according to array type but rather by surgeon preference for the individual case and anatomy. Approaches involving additional exposure (cochleostomy and extended round window) were used for recipients of both 632 and 632D arrays. One subject (Dex-EA) had ossification around the round window that required additional drilling for access. Two subjects (control: 1, Dex-EA: 1) had tip fold over identified and corrected at time of surgery. No other notable intraoperative issues were encountered with either the standard 632 or investigational 632D arrays. Subjects in both groups received perioperative intravenous dexamethasone (10 mg). No additional corticosteroids were administered perioperatively or during the immediate postoperative period.

2.3. Impedance – Access resistance and Polarization Impedance

To measure impedance, a biphasic pulse with a known current amplitude is delivered to an electrode, and the resulting voltage response is recorded. Total impedance, Ra and Zp were derived from the Transimpedance Matrix (TIM) measurement module within Custom Sound EP 6.0 software (Cochlear Corp). TIM measurements were obtained in the 3-pt, monopolar configuration, in which the stimulating and recording electrodes are intracochlear and the single reference (case, or MP2, ground) electrode is extracochlear (Fig 1).

When a single intracochlear electrode is stimulated, voltage recordings are obtained from that electrode as well as from all other electrodes along the cochlear implant array. In a 22-electrode array, stimulation of one electrode results in recordings from all 22 electrodes, producing a 1 × 22 array. This is repeated for each electrode in the array resulting in a 22 × 22 matrix i.e., TIM. Only the voltages measured from the on-electrode recordings (i.e. the stimulation electrode = recording electrode) were used.

The TIM was measured at two time points (6 μSec and 50 μSec post stimulus onset) during a 50 μSec per phase, cathodic-leading, symmetrical biphasic pulse with a 9 μSec interphase gap. The current level (CL) was 90, which is equivalent to 88.9 μA. Total impedance was calculated from the measured voltage at the 50-μSec time point, which is a longer pulse duration than used for clinical impedance measurements. Access resistance was calculated from the measured voltage (V) at the 6-μs time point, while polarization impedance was calculated as the difference between the measured V at 50 μs and 6 μs (V_{50 μs} − V_{6 μs})³⁰.

2.4. Impedance – Nested Bipolar Impedance

Nested BP impedance was obtained within Custom Sound EP 6.0 software using a research key. The stimulation parameters used for each measurement were identical to that used for the TIM measurement within the same testing series. For each measurement, two adjacent electrodes serve as the recording pair, while the immediately flanking electrodes on either side are used for stimulation (Fig 1). In a 22-electrode array, such as the 632 and 632D, this configuration yields a total of 19 unique measurement configurations.

Impedance data were collected at initial activation (approximately two weeks postoperatively), two weeks, one month, three months, six months, and 12 months post-activation. Of note, electrodes deactivated by the clinical team from the patients' stored programs were removed from our analysis. Clinical reasons for deactivating electrodes include open circuits (electrode 10 for one individual), non-auditory percepts or aversive sound quality, limited loudness growth, and/or restricted compliance limits. The rationale for excluding deactivated electrodes was to avoid elevated impedances due to the absence of current flow. There were a disproportionate number of deactivated electrodes for 632 arrays compared to 632D arrays (Table 1).

Table 1.

Number of subjects with deactivated electrodes at each time point.

	Electrode 1		Electrode 2		Electrode 3		Electrode 10
	Control	Dex	Control	Dex	Control	Dex	Control	Dex
IA	6	4	1	0	0	0	0	1
2wk	9	5	3	2	1	0	0	1
1m	9	5	4	2	2	0	0	1
3m	10	5	4	2	2	0	0	1
6m	11	5	8	2	4	0	0	1
12m	11	5	8	2	4	0	0	1

2.5. Analysis

This study focuses primarily on descriptive analyses (means, standard deviations). To assess whether nested bipolar impedance was more sensitive to the observed increases for basal electrodes compared to access resistance, proportional changes relative to the IA value were calculated, and paired t-tests were performed for the control group at 6 months and 12 months.

Separately, Welch’s t-tests were performed to compare impedance measures between control and Dex-EA groups at the base (average of 5 most basal electrodes or electrode pairs) and apex (average of 5 most apical electrodes or electrode pairs) at IA and 12 months.

3. Results:

3.1. Baseline characteristics

As shown in Table 2, there were 17 subjects (15 male, 2 female) in the control group with a mean age at surgery of 63.24 years and 19 subjects (12 male, 7 female) in the experimental group with mean age at surgery of 68.37 years. For the 17 subjects in the control group, the most common etiology of hearing loss was noise exposure (n=6), followed by idiopathic (n=5), genetic/familial (n=5), and Meniere’s disease (n=1). For the 19 subjects in the experimental group, the most common etiology of hearing loss was noise exposure (n=8), followed by idiopathic (n=7), genetic/familial (n=3), and Meniere’s disease (n=1). There was no significant difference in preoperative low-frequency pure-tone average (LFPTA) between the control and Dex-EA groups (unpaired t-test, p = 0.45) (Table 3).

Table 2.

Baseline characteristics of study participants

	Control (632) (n = 17) – (%)	Experimental (632D) (n = 19) – (%)
Age at surgery (mean ± Std) yrs	63.24 ± 16.08	68.37 ± 9.21
Sex
Male	15 (88)	12 (63)
Female	2 (12)	7 (37)
Etiology of hearing loss
Idiopathic	5 (29.4)	7 (36.8)
Noise exposure	6 (35.3)	8 (42.1)
Genetic/Familial	5 (29.4)	3 (15.8)
Meniere’s disease	1 (5.9)	1 (5.3)

Table 3.

Preoperative low-frequency pure-tone averages (LFPTA) for control and Dex-EA groups.

	LFPTA-mean	LFPTA-median	LFPTA-min	LFPTA-max
Control (n=17)	70 dB HL	73 dB HL	33 dB HL	102 dB HL
Dex-EA (n=19)	66 dB HL	67 dB HL	37 dB HL	97 dB HL

Although the total number of subjects was 17 in the control group and 19 in the experimental group, the varying numbers (N) across different figures is due to missed appointments or missed measurements at those instances. Reasons for missed appointments included illness, weather, or scheduling conflicts.

As displayed in Figure 2A, there were no noticeable differences in the perioperative total monopolar impedance between the control (black) and experimental (green) groups across all 22 electrodes. Figures 2B–G illustrate evolving spatiotemporal dynamics of total monopolar impedance for each participant across all 22 electrodes at initial activation (B), 2 weeks (C), 1 month (D), 3 months (E), 6 months (F), and 12 months (G).

Figure 2.

Total monopolar impedance for each participant across all 22 electrodes (individual line) perioperatively (A), and at initial activation (B), 2 weeks (C), 1 month (D), 3 months (E), 6 months (F), and 12 months (G). Color denotes type of electrode array (black: 632; green: 632D). There was one data point (23.3 kilo-ohms (kΩ)) in the control group that exceeded the upper y-axis limit (22 kΩ).

3.2. Individual data – access resistance, polarization impedance, and nested bipolar impedance

Figure 3 displays the Ra for each participant across all 22 electrodes across six time points: initial activation (A), 2 weeks (B), 1 month (C), 3 months (D), 6 months (E), and 12 months (F). In the control group (black), Ra increased over time, starting from the most basal (electrode 1) electrodes and progressing apically with time. As expected, Dex-EAs (green) were associated with reduced impedances across all electrodes at all observed time points.

Figure 3.

Access resistance for each participant across all 22 electrodes (individual line) across six time points: (A) initial activation, (B) 2 weeks, (C) 1 month, (D) 3 months, (E) 6 months, and (F) 12 months. Color denotes type of electrode array (black: 632; green: 632D).

Figure 4 displays the Zp for each participant across all 22 electrodes across six time points: initial activation (A), 2 weeks (B), 1 month (C), 3 months (D), 6 months (E), and 12 months (F). In the control group (black), Zp decreases after IA with more stability between three months and 12 months. In comparison, Dex-EAs (green) were associated with reduced impedances across all electrodes at all observed time points.

Figure 4.

Polarization impedance for each participant across all 22 electrodes (individual line) across six time points: (A) initial activation, (B) 2 weeks, (C) 1 month, (D) 3 months, (E) 6 months, and (F) 12 months. Color denotes type of electrode array (black: 632; green: 632D).

Figure 5 displays the nested bipolar impedance for each participant across all 19 recording electrode pairs across six time points: initial activation (A), 2 weeks (B), 1 month (C), 3 months (D), 6 months (E), and 12 months (F). In the control group (black), nested bipolar impedance increased over time, starting from the most basal (electrode 1) electrodes and progressing apically with time. In comparison, Dex-EAs (green) were associated with reduced impedances across all electrodes at all observed time points. Group differences are largest for basal electrodes and the latest measurement time point.

Figure 5.

Nested bipolar impedance for each participant across all 19 recording electrodes pairs (individual line) across six time points: (A) initial activation, (B) 2 weeks, (C) 1 month, (D) 3 months, (E) 6 months, and (F) 12 months. Color denotes type of electrode array (black: 632; green: 632D). In panels E and F there were one (2049 Ohms) and two (3108 and 2241 Ohms) data points, respectively, exceeded the upper y-axis limit; all three out-of-range values belonged to the same control subject.

3.3. Group trends – access resistance, polarization impedance, and nested bipolar impedance

Group means for each measurement time point and electrode are provided in Figure 6 to highlight some systematic changes across cochlear space and time. This arrangement of the data also assists with comparing across types of impedance measurements and highlights group differences.

Figure 6.

Access resistance (panels A and D), polarization impedance (panels B and E), and nested bipolar impedance (panels C and F) averaged across all subjects (individual line) at one time point: initial activation (black), 2 weeks (red),1 month (orange), 3 months (green), 6 months (blue), and 12 months (purple). Top and bottom rows show data for control and Dex-EA subjects, respectively.

For both control and Dex-EAs, polarization impedance (middle panel) tends to be highest at initial activation. Although impedance decrements across time are observed across the array, the largest decreases are observed at apical electrode sites. For control ears, the largest change is observed between initial activation and 2 weeks than for Dex-EAs, followed by smaller changes. For Dex-EAs, after 3 months, decrements are no longer observed.

Group differences are more obvious in the dynamic spatial profiles of access resistance and nested bipolar impedance (left and right panels). For control ears, the access resistance and nested bipolar impedance profiles are similar (though more pronounced for nested bipolar impedance) in that impedance increases at basal sites over time; these subsequently progress toward more apical sites. Although nested bipolar impedance for the most apical electrodes is generally more stable through 6 months, access resistance continues to decline through 12 months for these sites. This reduction in access resistance is also visible for the Dex-EAs. With this one exception, the access resistance and nested bipolar impedance profiles for Dex-EAs differ dramatically from control EAs: besides a small decline in nested bipolar impedance post activation, the profiles are otherwise stable remaining lower than the nested bipolar impedance in control EAs.

Figure 7 provides a three-dimensional depiction of mean nested bipolar impedance across time in the control group (left; corresponding to Figure 6C) and Dex-EA group (right; corresponding to Figure 6F).

Figure 7.

Three-dimensional rendering of average nested bipolar impedances over time for control (A; same data as in 6C) and Dex-EA (B; same data as in 6F) subjects.

3.4. Statistical tests – access resistance versus nested bipolar impedance

To account for multiple electrodes per individual, basal electrode measurements were averaged within each subject (electrodes 3–6 for Ra and electrode pairs 2–5 for nested bipolar impedance). This analysis focused on the basal region because it demonstrated the largest impedance increases in the control group, the most profound effect of the Dex-EA, and the clearest difference between nested bipolar impedance and Ra. Electrodes 1–2 were excluded because frequent inactivation at later time points resulted in missing longitudinal data.

As shown in the Table 4, the paired t-test comparing the average (across electrodes 3–6) proportion change from IA to 6 months in the control group measured with Ra to the average (electrode pairs 2–5) proportion change measured with nested bipolar impedance yielded a statistically significant difference (p = 0.001), with a mean factor change of 1.68 for Ra compared to 4.25 for nested bipolar impedance.

Table 4.

Statistical tests comparing proportional change (relative to IA) in Ra and nested bipolar impedance (BPZ) at 6 months and 12 months.

Treatment Group	Time	Mean factor change - Ra	Mean factor change – BPZ	p-value
Control	IA - 6 months	1.68	4.25	0.001
Control	IA - 12 months	1.79	7.20	<0.001

The paired t-test comparing the average (across electrodes 3–6) proportion change from IA to 12 months in the control group measured with Ra to the average (electrode pairs 2–5) proportion change measured with nested bipolar impedance also yielded a statistically significant difference (p < 0.001), with a mean factor change of 1.79 for Ra compared to 7.20 for nested bipolar impedance.

3.5. Statistical tests – control versus Dex-EA

Ra was significantly lower in the Dex-EA group at both the base and apex at IA (both p < 0.0001) and at 12 months (both p < 0.0001). Similarly, Zp was significantly lower in the Dex-EA group at both the base and apex at IA (both p < 0.0001) and at 12 months (both p < 0.0001).

In contrast, nested bipolar impedance did not differ significantly between the control and Dex-EA groups at IA at either the base (p = 0.5154) or apex (p=0.6154). However, by 12 months, nested bipolar impedance was significantly lower in the Dex-EA group compared to controls at both the base (p < 0.0001) and apex (p = 0.0035).

3.6. Selected patient

Figure 8 displays Ra (A), Zp (B), and nested bipolar impedance (C) across the electrode array at multiple postoperative time points for a selected patient from the experimental group who subjectively reported a “silicone allergy” at a post-operative clinical appointment. This patient received an experimental Dex-EA. As displayed in Figure 5A and 8C, this patient exhibited unusually elevated nested bipolar impedance at initial activation that decreased at the 2-week and 1-month visits and fell within the group range by 3 months (Fig 8C). Although Ra and Zp data were not available for this patient at initial activation, both parameters demonstrated a similar decreasing trend between the 2-week and 3-month time points (Fig. 8A, B).

Figure 8.

Access resistance (A), polarization impedance (B), and nested bipolar impedance (C) across the electrode array for a selected patient with “silicone allergy.” Within each panel, longitudinal data is shown at initial activation (black, panel C only), 2 weeks (red),1 month (orange), 3 months (green), 6 months (blue), and 12 months (purple).

4. Discussion:

This study characterizes and compares the spatial profiles and temporal dynamics of Ra, Zp, and nested bipolar impedance over 12 months following cochlear implantation with standard 632 EAs and evaluates the effects of Dex-EAs on these dynamics. In control subjects, Ra and nested bipolar impedance increase over the 12 months, starting from the basal electrodes and progressing apically with time. In control subjects, Zp decreases after IA with more stability between three months and 12 months. In comparison, across all three impedance measures, all 22 electrodes, and all six postoperative time points, Dex-EAs consistently exhibit lower impedance values compared with controls. Impedances are also generally more stable with a few exceptions: Ra and Zp decline through 3 months, particularly for apical sites; nested bipolar impedance declines across the array after initial activation only.

When comparing the three impedance measures, nested bipolar impedance appears more sensitive to ongoing intracochlear changes between 6 and 12 months compared to Ra or Zp. Because both nested bipolar impedance and Ra are theoretically sensitive to the bulk tissue surrounding of the EA, comparing these two measures may be particularly informative for assessing intracochlear tissue changes over time. Specifically, comparison of the 6-month (Fig. 5E) and 12-month (Fig. 5F) time points demonstrates a clear basal-to-apical increase in nested bipolar impedance among control subjects. This progressive rise likely reflects evolving tissue changes within the intracochlear environment occurring during this interval. In contrast, Ra exhibits less dramatic changes between 6 (Fig. 3E) and 12 (Fig. 3F) months in control subjects. Analysis of basal electrodes in control ears demonstrated that the proportional increase in impedance from IA to both 6 months (mean factor change 4.25 vs 1.68) and 12 months (7.20 vs 1.79) was significantly greater for nested bipolar impedance compared to Ra. Taken together, these findings suggest that nested bipolar impedance is a more sensitive metric for detecting the spatiotemporal dynamics of CI impedance over time.

The higher sensitivity of nested bipolar impedance to ongoing tissue changes are likely explained by the inherent design of this measurement, which is obtained using a bipolar configuration. In this configuration, the stimulating electrodes and both recording electrodes are intracochlear. As a result, current is delivered between two closely spaced stimulating electrodes, which are separated only by the two adjacent recording electrodes positioned between them. The resulting voltage change is measured across these two adjacent recording electrodes. This highly localized four-electrode arrangement enables interrogation of impedance over a spatially confined region of the intracochlear electrode–tissue interface. Notably, when comparing the typical magnitude of these different impedance measurements in the experimental cohort, Ra values cluster around ~1.5 kOhms, Zp around ~3 kOhms, and nested bipolar impedance only around ~0.15 kOhms. Additionally, in this configuration, the sensing electrodes do not carry current, so impedance contributions from the electrode-tissue interface are effectively excluded and the measured voltage reflects only the resistance of nearby tissue and perilymph³³. Thus, the greater proportional change observed with nested bipolar impedance may reflect both its lower baseline magnitude and biologically meaningful sensitivity to localized intracochlear changes.

This measurement configuration may also explain why nested bipolar impedance did not show a significant group difference at IA, despite significant differences in Ra and Zp. Ra and Zp are derived from monopolar measurements and include contributions from the stimulating electrode–tissue interface. In contrast, nested bipolar impedance uses a spatially focused four-electrode configuration in which the recording electrodes do not carry current, eliminating electrode–tissue interface contributions and emphasizing local intracochlear tissue/perilymph resistance. Thus, at IA, the local intracochlear environment evaluated by nested bipolar impedance may not yet differ substantially between groups. As fibrotic tissue formation or other local remodeling progresses over time, however, nested bipolar impedance may become increasingly sensitive to these localized changes, accounting for the emergence of significant group differences by 12 months.

The findings of this study are consistent with prior reports demonstrating that Dex-EAs are associated with reduced electrode impedance over time^3,23,24. Needham et al., demonstrated lower nested bipolar impedance measurements in guinea pigs with Dex-EAs while Briggs et al., demonstrated lower nested bipolar impedance measurements in human subjects with Dex-EAs^23,24. However, to our knowledge, this is the first study to characterize the spatiotemporal dynamics of nested bipolar impedance at the level of individual electrodes within individual subjects and to compare these patterns between control and Dex-EA recipients. Furthermore, this is the first study to compare the spatiotemporal dynamics of Ra, Zp, and nested bipolar impedance at such fine spatial resolution (individual electrodes in individual subjects) in both standard and Dex-EAs, and to demonstrate the greater sensitivity of nested bipolar impedance for detecting impedance changes over time. Further strengths of this study include its prospective design and standardized data collection, as well as a relatively large and well-matched sample size in both groups.

Although the dexamethasone-eluting silicone wells on the 632D array are located along electrodes 7–22, lower impedance values were also observed at more basal electrodes. The mechanism underlying this finding is not fully established. One possibility is that dexamethasone released into the cochlear fluid space diffuses beyond the immediate region of the drug-eluting wells and modulates inflammatory or fibrotic responses more broadly, including the basal cochlear region. In addition, local suppression of inflammation around the apical and mid-array portions of the electrode may influence the overall intracochlear foreign body response surrounding the array. Systemic steroid exposure is unlikely to explain the group differences, as both groups received the same perioperative intravenous dexamethasone dose. Overall, the precise mechanism responsible for basal impedance reduction remains uncertain.

Impedance can be impacted by the surface area of the electrode (Ra and Zp)³⁰, by the cross-sectional area of the cochlear duct (nested BP impedance)³⁹, and by the distances between the electrode contact and scalar boundary (nested BP)³⁹. Specifically, smaller electrode surface areas, smaller cross-sectional duct areas and smaller distances between the electrode and ductal wall are associated with higher impedances. Scalar tympani size tends to be larger in the base than the apex⁴⁰. For the perimodiolar arrays used in this data set, although contact surface areas between basal and apical electrodes differ minimally (i.e. by 0.01mm² between 0.16mm² and 0.15mm², respectively), closer electrode-modiolar position is expected in general for apical electrodes compared to basal electrodes⁴⁰. These reasons, collectively, may explain why impedances are higher for apical electrodes compared to basal electrodes in some instances. In this data set, Ra and Zp are higher for apical sites for Dex-EAs across all time points and for standard EAs at the earliest measurements. The pattern is likely clearer in the Dex-EAs because changes in the intracochlear environment, which also impact impedance and are of greater interest, are minimized²⁰. It is not clear why the nested BP impedance profiles for the Dex-EAs do not reflect this same pattern.

Figure 5A highlights a patient in the Dex-EA group who exhibited unusually elevated nested bipolar impedance across all electrodes at initial activation. These elevated values decreased at the 2-week and 1-month visits and fell within the group range by 3 months (Fig. 8C). Although Ra and Zp data were not available at initial activation for this patient, both parameters demonstrated a similar decreasing trend between the 2-week and 3-month time points (Fig. 8A, B). Interestingly, this patient reported a history of “silicone allergy” at one of the early postoperative clinical visits. Review of the preoperative medical record revealed no documented history of silicone allergy and no objective testing supporting this diagnosis. Nonetheless, the observed impedance trajectory in this patient—particularly in the context of implantation with a Dex-EA—may provide further insights into the effects of dexamethasone on impedance behavior. One plausible explanation is that the elevated impedances at initial activation reflected an exaggerated inflammatory response to silicone components of the implant, which was subsequently attenuated by local dexamethasone release, allowing impedances to normalize by 3 months. Given that impedance measures tended to increase over time in control subjects, it is quite possible that this patient may have exhibited persistently elevated or progressively increasing impedances had a standard electrode array been implanted rather than a Dex-EA. Of note, this interpretation remains hypothetical due to lack of confirmatory allergy testing.

Although the first 8 subjects were randomized as part of a multicenter RCT, the remaining subjects were enrolled non-randomly. The next 15 subjects were assigned to the Dex-EA group because the investigational devices were approaching expiration and were therefore used consecutively while available. After these devices were no longer available, subsequent subjects were continuously recruited into the control group. Thus, subjects were not selectively assigned based on demographic or clinical characteristics, and nonrandomized group assignment was determined only by investigational device availability. We do not have a priori hypotheses about adult age or sex on electrode impedance, and the consistency of group differences across impedance measures and longitudinal time points is reassuring. The age differences across group are noteworthy and are being carefully considered for audiological and electrophysiological outcome measures.

The statistical analyses were designed to support the descriptive goals of this study rather than to perform exhaustive electrode-by-electrode hypothesis testing. The primary objectives were to characterize broad spatiotemporal impedance patterns between control and Dex-EA subjects and to compare the relative sensitivity of nested bipolar impedance and access resistance to longitudinal impedance changes. Specifically, the analyses focused on two principal observations: clear longitudinal group differences between control and Dex-EA subjects, and greater sensitivity of nested bipolar impedance compared with access resistance for detecting these changes. Although more comprehensive approaches, including linear mixed-effects models, correction for multiple comparisons, and modeling of spatial autocorrelation among adjacent electrodes, may provide a more rigorous framework in larger datasets, these methods were not applied here because of the modest cohort size, missing data at later time points, and the large number of repeated measurements generated across electrodes, impedance measures, and longitudinal time points. Accordingly, the results should be interpreted as descriptive evidence of consistent spatiotemporal trends rather than exhaustive hypothesis testing across all electrodes, impedance measures, and time points.

Post-implantation foreign body response—and the resulting fibrosis and neo-ossification—remain critical challenges in cochlear implant surgery and represent important targets for research and therapeutic solutions.

The dramatic reduction in impedance values across all electrodes and time points observed in subjects implanted with Dex-EAs suggests meaningful potential for reducing foreign body response, fibrosis, and neo-ossification—processes that can limit the efficiency of electrical stimulation and compromise implant performance. Ultimately, however, the clinical value of Dex-EAs must be determined by whether these biophysical improvements translate into improved patient outcomes without significant adverse effects. Additionally, longer-term impedance trajectories will be important for comparison with clinical outcomes. The present study was limited to 12 months of follow-up. However, these patients will continue to be followed at 24 and 36 months, which will help determine how the observed impedance patterns evolve over time.

The control versus Dex-EA clinical trial provided a unique opportunity to evaluate multiple impedance metrics. Because dexamethasone was hypothesized to reduce impedance through its modulation of the foreign body response, this setting allowed us to determine which impedance measure was most sensitive to these expected differences. Comparison of the impedance metrics demonstrated that nested bipolar impedance is more sensitive to ongoing impedance changes than Ra and Zp. Although the Ra for basal electrodes incrementally increases, the proportional changes observed in the nested BP impedance measures and the greater inter-individual variability suggests that this measure more directly reflects the underlying mechanism of interest, with reduced contribution from confounding factors. When interpreting these impedance values in the context of expected increasing apical to basal gradient of fibrosis and neo-ossification (from other post-mortem data), the nested BP impedance appears more sensitive to these predicted changes in the intracochlear environment. Importantly, we demonstrate these impedance patterns at the finest spatial resolution yet reported—individual electrodes within individual subjects. Collectively, these findings support nested bipolar impedance as a focused biomarker of the intracochlear environment with enhanced sensitivity to ongoing biological changes around the EA. However, routine clinical implementation remains limited because access to the nested bipolar impedance module currently requires a research key. While the ultimate clinical value of Dex-EAs remains to be determined, future studies evaluating different steroid dosages, alternative steroid formulations, or other pharmacologic interventions should consider nested bipolar impedance as a primary tool for probing the intracochlear environment.

5. Conclusion:

This study demonstrates that Dex-EAs are associated with consistently lower and more stable impedance values compared to standard EAs across multiple postoperative time points up to one year. Additionally, nested bipolar impedance appears more sensitive to ongoing changes in the intracochlear environment. Because this measurement is obtained entirely within the cochlea using a localized four-electrode configuration, it likely reflects changes in the surrounding intracochlear tissue environment with reduced influence from extracochlear factors compared to Ra. Although the ultimate clinical benefits of Dex-EAs remain to be determined, future studies evaluating alternative steroid dosages, formulations or other approaches to modulate the foreign body response would benefit from incorporating nested bipolar impedance to more precisely characterize changes within the intracochlear environment.

HIGHLIGHT.

• There is a spatial and temporal evolution of impedance changes after cochlear implantation beginning in the base and progressing apically over time.

• Nested bipolar impedance appears more sensitive to these changes compared with monopolar measures such as access resistance.

• Dexamethasone elution is highly effective in reducing and stabilizing electrode impedance across the entire electrode array.