Limitations in dosing of electrical nerve stimulation – assessment with compound potential measurements

Most therapeutic applications of peripheral nerve electrical stimulation do not measure resulting nerve activation. A fixed axon population is often assumed to stably respond to electrical stimulation, with nerve firing rates increasing proportionally to stimulation frequency. We investigated this assumption in peripheral somatosensory fibres of 10 healthy human participants, measuring evoked compound action potentials (CAP) as a representation of nerve population firing rate. One-second stimulus trains at frequencies 25 to 200 Hz were repeated with inter-stimulus intervals of 0.1 to 6 seconds. Over duration of a 1s stimulus train, CAP amplitudes showed mean declines of nearly 40% for 200 Hz stimulation frequency and 20% for 100 Hz, compared to less than 10% for 25 and 50 Hz, with majority of these decreases happening within 0.2s of stimulation onset. CAP amplitude declines were mainly associated with stimulation frequency and short timescales, compared to longer timescales across dozens of repeated stimulus trains (mean decline of 4.9%, range 1.03 to 11.6% across parameter combinations). These findings show that the assumption of axon populations responding stably to electrical stimulation is untrue. CAP amplitude declines showed large variation between different participants, emphasising the need to tailor dosing of electrical stimulation for individual users using objective physiological measures.