Optically pumped magnetometers for a brain-computer interface based on event-related desynchronization

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Background Stroke is one of the leading causes of disability worldwide and often responsible for impairments of hand motor function. Rehabilitation and restoration of motor functions can be significantly improved using devices controlled by brain signals, so called brain-computer interfaces (BCIs). Most current BCI systems are based on electroencephalography (EEG), which provides only limited spatial resolution and thus limited versatility of control commands. Compared to EEG, optically pumped magnetometers (OPMs) measure cortical magnetic fields without contact to the scalp and provide a higher spatial resolution and bandwidth. In contrast to superconducting quantum interference device (SQUID)-based magnetoencephalography (MEG), OPMs have low maintenance cost and allow movement in the scanner, making them more applicable in clinical contexts. Methods We quantify the signal characteristics of a commercially available OPM system (FieldLine Inc., USA) in terms of noise floor, dynamic range and bandwidth to verify its suitability for cortical measurements. We then develop an experiment contrasting resting and right hand grasping imagery to measure modulations of the sensorimotor rhythm (SMR) with 17 OPMs over the left motor cortex of 18 healthy participants. The BCI capabilities of the OPM acquisition system are evaluated with a modular near real-time classification pipeline, which provides visual feedback to the user. Results The sensor characterization revealed a system noise floor of about 27 fT/ √ Hz at 10 Hz, a bandwidth of 400 Hz and a dynamic range of ±15 nT, fulfilling the minimum requirements for cortical measurements. In 10 of 16 eligible participants, a difference in SMR power between resting and grasping condition could be identified. We show that OPMs are suitable to measure SMR modulations in near real-time and that the classification performance of our pipeline significantly exceeds chance level. Discussion OPMs allow for the online quantification of voluntary modulations of the sensorimotor rhythm on single-trial basis, a central requirement for many BCI systems used in the rehabilitation of stroke survivors. With their higher spatial resolution compared to EEG, OPMs could be used for more complex classification paradigms and ultimately facilitate a development towards more versatile BCI applications. The increasing availability and sensitivity of commercialized OPM systems allows for the exploration of MEG in new research areas. OPMs are projected to become an important tool in the field of cognitive neuroscience within the next few years. Keywords: brain-computer interface, optically pumped
Faculteit der Sociale Wetenschappen