An automated magnetoencephalographic data cleaning algorithm

Abstract

The problem of cleaning magnetoencephalographic data is addressed in this manuscript. At present, several denoising procedures have been proposed in the literature, nevertheless their adoption is limited due to the difficulty in implementing and properly tuning the algorithms. Therefore, as of today, the gold standard remains manual cleaning. We propose an approach developed with the aim of automating each step of the manual cleaning. Its peculiarities are the ease of implementation and using and the remarkable reproducibility of the results. Interestingly, the algorithm has been designed to imitate the reasoning behind the manual procedure, carried out by trained experts. Our statistical analysis shows that no significant differences can be found between the two approaches.

Keywords:

• Magnetoencephalography (MEG)
• bad data
• noise reduction
• automated artifact removal

Notes

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 CJS, version 09.152.1.23, available at http://home.kpn.nl/stam7883/brainwave.html.

2 The code is available at http://audition.ens.fr/adc/NoiseTools/.

Additional information

Funding

This work has been partly supported by the University of Napoli Parthenope under the ‘Bando di ricerca competitiva di ateneo per il triennio 2016–2018 progetto: implementazione di studi di connettività cerebrale mediante magnetoencefalografia in ambito multidisciplinare’.