4ECAP threshold & functions

A single ECAP waveform is a curiosity; the clinical value lives in how the ECAP changes as you vary the stimulus. Sweep the level and you get the amplitude growth function, whose intercept is the ECAP threshold and whose slope speaks to neural survival. Vary the masker–probe interval and you get the recovery function, a window onto refractoriness. Move the masker along the array and you get the spread of excitation, a map of channel interaction. This module is about turning the waveform into numbers that inform the MAP.

TThe ECAP threshold (tNRT / tECAP)

The ECAP threshold — written tNRT, tECAP, or T-NRT depending on the manufacturer — is the lowest stimulus level that produces a detectable neural response. It is not read off a single trace; it is found by recording an amplitude growth function and finding where the fitted line crosses zero amplitude (or the lowest level with a visible N1–P2, in a visual-detection approach).[1999, 2005]

The ECAP threshold also varies with the stimulating electrode configuration (monopolar vs more focused modes), and early work establishing the measure compared these configurations directly against psychophysical responses — a reminder that an ECAP threshold is always read in the context of how it was elicited.[1996]

The reason the ECAP threshold matters clinically is that it correlates — loosely but usefully — with the behavioural levels that define the MAP. It tends to sit within the behavioural dynamic range, generally between T- and C-level, which makes it a starting anchor when no behavioural data exist. Module 9 develops how that anchor is actually used (and how loose the correlation really is).[2000, 2010]

TCThe amplitude growth function

The amplitude growth function (AGF) plots N1–P2 amplitude against stimulus current level. Two features are read from it:

• The x-intercept — the extrapolated zero-amplitude level, i.e. the ECAP threshold.
• The slope — how steeply amplitude grows with level. A steeper slope generally indicates a larger or healthier population of excitable neurons responding to each increment of current.

AGF slope is one of the most studied candidate proxies for local neural survival. Steeper slopes have been associated with better neural populations in animal models and with some outcome measures in humans, though the relationship is noisy and confounded by electrode position and distance from the modiolus. Treat slope as a soft indicator, not a survival assay.[2017]

CThe recovery (refractory) function

Using the same forward-masking machinery introduced in Module 3, but now varying the masker–probe interval, you can chart how ECAP amplitude recovers as the nerve emerges from refractoriness. At very short intervals the probe response is suppressed; as the interval lengthens, amplitude recovers along an exponential time course.

The recovery time constant indexes how quickly the local nerve population can fire again, which bears on temporal processing — the ability to follow rapid stimulation rates. Abnormal (often slowed) recovery has been reported in auditory neuropathy and in ears with poor neural status, and is of interest for tailoring stimulation rate to the individual nerve.[2017]

CSpread of excitation

The spread of excitation (SOE) — sometimes called the spatial-spread or channel-interaction function — measures how broadly the neural population is activated when you stimulate one electrode. It uses a forward-masking paradigm in which the masker electrode is stepped along the array while the probe stays fixed: the more a distant masker still suppresses the probe response, the wider the overlap of their neural populations.[2003, 2008]

The result is a roughly bell-shaped function centred on the probe electrode. Two readings matter:

• Width — a broad function means heavy channel interaction: adjacent electrodes excite overlapping neurons, blurring spectral information. Narrower is generally better for spectral resolution.
• Peak position — if the function peaks away from the stimulating electrode, current is being steered toward neurons elsewhere, which can flag electrode-position anomalies or dead regions.

SOE has been used to identify electrodes worth deactivating in site-selection strategies and to study why some recipients do poorly despite normal-looking thresholds. Newer panoramic ECAP methods estimate patient-specific current spread and neural-health patterns across the whole array, and have been used to compare excitation between lateral-wall and perimodiolar arrays.[2009, 2011, 2021, 2025]

CReading neural health — promise and caution

Taken together, AGF slope, recovery kinetics, spread of excitation, and the polarity sensitivity of the response form an emerging “objective profile” of the cochlear neural substrate. The hope is to move from setting levels to characterising the nerve — identifying poor-survival regions, individualising rate and electrode selection, and predicting outcome.[2017]

The honest caution: every one of these measures is confounded by electrode-to-modiolus distance, array type, and recording noise, and none is a validated standalone survival assay in humans. They are research-grade indicators finding their way, cautiously, into practice — powerful for understanding a difficult ear, not yet a substitute for behavioural optimisation.