Optimal selection of multipolar electrode configurations for nerve burst detection.

Nerve cuff electrodes are commonly used for neural stimulation and recording applications. Usually, these electrodes are composed of a limited set of metal rings, disposed around the nerve. Although widely used, this technology may be insufficient to record and stimulate in a more selective manner. Higher resolution electrodes, usually composed of a matrix of independent contact points, have been proposed in this sense. These electrodes allow for the exploration of a wide variety of bipolar or multipolar setups, for selective recording and stimulation. In this study, we propose a method to optimally select such multipolar setups and to quantitatively evaluate the performance of a multi-contact neural organic electrode (OE) in recording burst discharges from the rat's phrenic nerve. A 16-channel OE was wrapped around the phrenic nerve (studied electrode) and a suction electrode was applied to the cut-end of the same nerve (gold standard electrode). Analysis of all possible combinations of bipoles and tripoles from the OE were carried out to assess the improvement in the recording performance, measured as the signal-to-noise ratio, compared to the gold standard. The results showed that the bipolar and tripolar configuration significantly increased the overall recording performance. Such configurations are therefore essential to improve nerve burst detection.

Assessment of the Use of Multi-Channel Organic Electrodes to Record ENG on Small Nerves: Application to Phrenic Nerve Burst Detection.

Effective closed-loop neuromodulation relies on the acquisition of appropriate physiological control variables and the delivery of an appropriate stimulation signal. In particular, electroneurogram (ENG) data acquired from a set of electrodes applied at the surface of the nerve may be used as a potential control variable in this field. Improved electrode technologies and data processing methods are clearly needed in this context. In this work, we evaluated a new electrode technology based on multichannel organic electrodes (OE) and applied a signal processing chain in order to detect respiratory-related bursts from the phrenic nerve. Phrenic ENG (pENG) were acquired from nine Long Evans rats in situ preparations. For each preparation, a 16-channel OE was applied around the phrenic nerve's surface and a suction electrode was applied to the cut end of the same nerve. The former electrode provided input multivariate pENG signals while the latter electrode provided the gold standard for data analysis. Correlations between OE signals and that from the gold standard were estimated. Signal to noise ratio (SNR) and ROC curves were built to quantify phrenic bursts detection performance. Correlation score showed the ability of the OE to record high-quality pENG. Our methods allowed good phrenic bursts detection. However, we failed to demonstrate a spatial selectivity from the multiple pENG recorded with our OE matrix. Altogether, our results suggest that highly flexible and biocompatible multi-channel electrode may represent an interesting alternative to metallic cuff electrodes to perform nerve bursts detection and/or closed-loop neuromodulation.