engleski

Automated estimation of peak-to-peak amplitude and latency of motor evoked potentials in transcranial magnetic stimulation studies

The aim of the present study was to propose an algorithm for automatic analysis of amplitude and latency of single MEP responses recorded continuously in a file. Recorded MEPs often have characteristic shapes and are well defined however, due to intra- and inter-individual MEP variability, any proposed algorithm intended for such signal processing needs to be robust enough to handle signal aberrations. Hence, in order to produce a robust tool for MEP analysis our goal was to deliver an algorithmic solution that can be defined by the following operational modules: 1) extraction from European Data Format (EDF) files ; 2) pre- processing (artefacts removal, de-noising) ; 3) fusion including linking of channel containing MEP responses with the triggering channel data (onset of magnetic stimulation) ; 4) graphical representation of individual channels ; 5) calculation of peak-to-peak amplitudes and latencies of individual MEPs ; 6) selection of individual MEP responses for later statistical analysis ; 7) graphical presentation of selected MEP signals with basic statistical data ; 8) preparing the given data set for future data learning, feature extraction and automatic classification and clustering. The methodology was successfully tested on signals continuously stored as EDF files by means of navigated brain stimulation (NBS) system of the manufacturer Nexstim (Nexstim NBS System 4, Nexstim Oy, Helsinki, Finland). Tests were performed on ten healthy subjects, in a study investigating short afferent cortical inhibition (SAI) by paired-pulse paradigm where conditioning electrical stimuli are applied to the median nerve followed by a single cortical TMS pulse over the M1 at different inter- stimulus intervals (ISIs). The algorithm can serve as additional armamentarium for the analysis of MEPs in research and clinical settings.

MEP ; latency ; TMS ; algorithm

nije evidentirano