Towards a mechanistic approach for the development of non‐invasive brain‐computer interfaces for motor rehabilitation

Abstract

Brain-computer interfaces (BCIs) designed for motor rehabilitation use brain signals associated with motor-processing states to guide neuroplastic changes in a state-dependent manner. These technologies are uniquely positioned to induce targeted and functionally relevant plastic changes in the human motor nervous system. However, while several studies have shown

Challenges and Opportunities for the Future of Brain-Computer Interface in Neurorehabilitation

“…By closing the disrupted sensorimotor loop and providing tangible feedback, the patient learns to control the effector by motor imagery or movement intentions. Restoring relevant sensory feedback in relation to volitional movement attempts is believed to mobilise the fundamental mechanisms of motor learning (Mrachacz-Kersting et al, 2021). As such, they can engage in mental practise of movement and keep their motor neural circuitry active, warding off the detrimental effects of limb non-use (Buxbaum et al, 2020), its associated white matter degeneration (Egorova et al, 2020), and promote usedependant neuroplastic processes (Xing and Bai, 2020).…”

“…For the BCI participant, learning to control the effector requires multiple practise sessions, viewing continuous feedback and learning by reward (Chavarriaga et al, 2017;Mrachacz-Kersting et al, 2021). While passive/implicit learning is known to play a role in BCI control (Othmer, 2009), most human participants report developing and fine-tuning mental strategies throughout the course of training, usually involving imagination of movement (Majid et al, 2015;Ruddy et al, 2018;, or in the case of brain injured patients, attempts to make movement with the paretic limb (Blokland et al, 2012;Balasubramanian et al, 2018;Bai et al, 2020).…”

“…Of these, candidate neural changes that co-occur with improvement in motor function include enhanced desynchronisation of sensorimotor rhythms (i.e., neural oscillations in the alpha 8-12 Hz and beta 13-30 Hz frequency range) over scalp locations corresponding to motor cortex (Buch et al, 2008;Prasad et al, 2010;Li et al, 2014;Ono et al, 2015), changes in functional connectivity (Varkuti et al, 2013;Pichiorri et al, 2015;Biasiucci et al, 2018;Wu et al, 2020), lateralisation of neural activity (Ramos-Murguialday et al, 2013) or changes to white matter microstructure (Song et al, 2015;Hong et al, 2017). Others have speculated that BCI works by mobilising the brain's intrinsic learning mechanisms, adapting behaviour using classical and/or operant conditioning giving rise to neural adaptations (Mrachacz-Kersting et al, 2021). To date, there has not been a comprehensive account that successfully resolves the aforementioned different perspectives into a holistic mechanistic model encompassing the electrophysiological, haemodynamic, and neurochemical components.…”

Brain‐computer interfaces and plasticity of the human nervous system

“…The overview points to the emergent concepts of combining classic NMES methods with brain interfacing. The core concept is that of classical conditioning via peripheral stimulation combined with centrally initiated neural activity, as reviewed by Mrachacz-Kersting et al (2021). The effects induced by this combination are based on associative plasticity, as first demonstrated by Mrachacz-Kersting et al (2012).…”

“…In this paradigm, a BCI decodes the instantaneous brain state to trigger direct brain stimulation or peripheral sensory feedback. These approaches, as well as the operant conditioning approaches, have been classified and reviewed by Mrachacz-Kersting et al (2021) in this special issue. Core questions in these applications are to what extent repetitive pairing of stimuli with different brain states would lead to lasting changes of corticospinal excitability and the extent of the functional consequences of enhanced cortical excitability.…”