7The Twitch: Facial-Nerve Stimulation
An electrode array sits a few millimetres from the labyrinthine segment of the facial nerve, separated only by the thin otic-capsule wall. When that wall lets current leak, the recipient gets a twitch instead of a tone. Facial-nerve stimulation is the textbook example of current going where it was never meant to go, and the most instructive complication to troubleshoot.
FWhy the nerve twitches: current spread across the otic capsule
The labyrinthine (and to a lesser extent tympanic) segment of the facial nerve runs close to the basal turn of the cochlea, separated by a thin shell of dense otic-capsule bone; intracochlear current that escapes this shell can depolarise the nerve and produce facial-muscle twitching (aberrant facial-nerve stimulation, FNS). FNS is a non-auditory side effect: the patient feels or sees a twitch of the eyelid, cheek or corner of the mouth synchronous with sound or with specific electrodes, rather than hearing more loudly. Healthy otic-capsule bone is a good electrical insulator; FNS arises when that insulation fails or is bypassed, so anything that lowers bone impedance or shortens the current path to the nerve raises the risk. Electrodes nearest the basal turn and nearest the modiolus deliver current closest to the labyrinthine facial nerve, which is why the apical-to-basal pattern of FNS so often points at the basal end of the array.[2013][2020]
CWho is at risk: otosclerosis, demineralisation and malformation
Otosclerosis is the classic risk factor: spongiotic otic-capsule bone shows resorption, vascular proliferation and new bone formation that lower electrical impedance, letting current shunt to the facial nerve. Histology shows FNS occurs when disease erodes the endosteum of the upper basal turn and the fallopian canal and all the bone between them. In a pooled meta-analysis the overall aberrant-FNS rate was about 5.6% (range 0.68-43%), but otosclerosis carried an odds ratio of roughly 13.7 with rates near 26%, and cochlear malformations the highest rates at about 28%. In far-advanced otosclerosis the risk accumulates over time: one cohort reported a 15-year cumulative FNS incidence of about 33%, so a recipient can be twitch-free at switch-on and develop FNS years later as disease progresses. Lateral-wall (straight) arrays carried a higher odds of FNS than perimodiolar arrays (OR ~3.92); perimodiolar designs direct current toward the modiolus and away from the lateral otic-capsule wall, and in otosclerotic temporal bones avoided FNS where straight arrays produced it in ~40%. Other contributors include cochlear malformation (thin or dehiscent walls), prior otic-capsule demineralisation, high stimulation levels, and electrodes positioned closest to the basal turn.[2013][2020][2023]
CDetecting and confirming FNS
Detection usually begins with the patient (or a family member) reporting a twitch, blink, or pulling at the eye or mouth that is time-locked to listening, often loudest in noisy or loud environments. The clinician confirms FNS at the programming session by stimulating individual electrodes while watching the face: the offending electrodes are identified one at a time, frequently those nearest the modiolus and the basal turn. FNS may appear only at higher current levels, so it can hide just above the comfort level and emerge as the map is made louder; conversely it may worsen as impedances drift over months. Imaging (CT of the bone between the fallopian canal and the upper basal turn) can predict and explain FNS in otosclerosis and helps separate it from a device fault.[2013][2023]
TThe management ladder and its cost in channels
Step 1 is to reduce the charge delivered by the offending electrode: lower the current level, then lengthen the pulse width to keep loudness while dropping peak current, and reduce the stimulation rate. These steps often abolish FNS without losing an electrode. Step 2 is to focus the current: switching from broad monopolar to focused configurations (bipolar, tripolar or partial tripolar) confines current near the modiolus and away from the facial nerve, and changing pulse shape/polarity (triphasic or anodic-leading pulses) reduces FNS while preserving speech performance. Step 3, when reprogramming fails, is to deactivate the offending electrode(s); reprogramming with or without deactivation eliminates FNS in the large majority, with explant a rare last resort (only 4 explants across 5694 patients in one meta-analysis). Every deactivated electrode is a lost spectral channel and focused modes cost battery and dynamic range, so management is a balance: silence the twitch while sacrificing as little frequency resolution and speech benefit as possible. Revision cochlear implantation (reinsertion, sometimes with a different array type or position) is reserved for intractable FNS that defeats the programming ladder.[2019][2020][2023][2025]
What is the most appropriate first management step?
Why does far-advanced otosclerosis predispose to facial-nerve stimulation from a cochlear implant?
Which array characteristic is associated with a higher risk of aberrant facial-nerve stimulation?
What is the principal clinical cost of deactivating electrodes to abolish FNS?