Intraoperative neural monitoring method utilizing wavelet-based event detection

Inventors

Wybo, Christopher • Nay, David S. • Scarfe, Darren P. • Scarfe, Lukas T. • O'Neil, Samantha J.

Assignees

Neuralytix LLC

Abstract

A method of alerting a user to an artificially induced neuromuscular response in a subject includes generating a mechanomyography (MMG) output signal corresponding to a mechanical motion of a muscle of a subject, applying a wavelet transform to the MMG output signal to determine a convolution coefficient for each of a plurality of daughter wavelets, summing the convolution coefficients determined across the plurality of daughter wavelets at each timestep across a plurality of timesteps to generate a net-convolution coefficient (NCC), identifying one or more peaks in the NCC via a peak finding algorithm, and alerting a user of an artificially induced neuromuscular response following the identification of one or more peaks in the NCC. Each daughter wavelet of the plurality of daughter wavelets is a time-scaled variant of a common mother wavelet, and the convolution coefficient is indicative of a similarity between the daughter wavelet and the MMG output signal.

Core Innovation

The invention provides an intraoperative neural monitoring system and methods for alerting a user to the existence of an artificially induced neuromuscular response in a subject. A mechanomyography (MMG) output signal is generated corresponding to a mechanical motion of a muscle of the subject using a mechanical sensor in physical communication with the muscle. A wavelet transform is applied to the MMG output signal to determine a convolution coefficient for each of a plurality of daughter wavelets as time-scaled variants of a common mother wavelet, where the convolution coefficient is indicative of similarity between the daughter wavelet and the MMG output signal.

The convolution coefficients determined across the plurality of daughter wavelets at each timestep across a plurality of timesteps are summed to generate a net-convolution coefficient (NCC) that varies across the plurality of timesteps. One or more peaks in the NCC are identified via a peak finding algorithm, and a plurality of peaks in the NCC corresponding to a transmitted electrical stimulus are identified, where the plurality of peaks have a periodicity that is about equal to a periodicity of the plurality of discrete electrical pulses.

Following identification of the plurality of peaks, an electrical stimulus comprising a second plurality of discrete electrical pulses is transmitted, where each discrete electrical pulse has a different current magnitude selected to identify a current sub-range that contains the minimum stimulus current. The minimum stimulus current is determined and displayed, and a user is alerted following identification of peaks in the NCC.

Claims Coverage

The provided excerpt contains three independent methods. The claims are built around MMG generation, a wavelet-based NCC peak-detection framework using time-scaled daughter wavelets from a common mother wavelet, and determination of a minimum stimulus current using discrete electrical pulses and periodicity matching.

Across the independent claims, the core inventive coverage is the use of a wavelet-transform framework that produces an NCC from MMG using time-scaled daughter wavelets from a common mother wavelet, followed by NCC peak identification and determination of a minimum stimulus current using discrete electrical pulses from a selectively electrifiable nerve stimulator with NCC peak periodicity about equal to pulse periodicity and with a current sub-range bracketing the minimum.

Stated Advantages

Documented Applications