This randomized controlled study investigated whether the combined effect of tDCS and KI-BCI on upper limb function in participants with subacute stroke is more effective than the effect of tDCS or KI-BCI alone. Our results showed that tDCS combined with KI-BCI training significantly improved upper extremity function (include fingers) compared with tDCS alone. There was no statistically difference between the KI-BCI combined with tDCS or use KI-BCI alone. In contrast, participants with subacute stroke who received KI-BCI training alone showed better performance in upper limb function (including and excluding fingers) and ADL than those who received tDCS alone. In terms of neurophysiological indicators, there were no significant differences among the three groups after intervention, but the participants who received tDCS treatment showed better quantitative electroencephalography (QEEG) indicators compared to before treatment.
Physical rehabilitation programs, especially when delivered as soon as possible after the onset of stroke, can be highly efficacious34. Notwithstanding, the rate of improvement in functional ability regained through physical rehabilitation tended to peak after a few months post-stroke and eventually tapered.35 Therefore, this study focused on participants with first-ever subacute stroke to minimize the influence of disease duration on rehabilitation outcomes. Recovery after stroke implies reorganization of the cortex to compensate for the lesioned area. This is possible through neuroplasticity mechanisms, whereby the brain learns and reorganizes itself to compensate for lost functions36. In this context, interventions that increase or prolong neuroplasticity have been the target of recent investigations.
Neuroplasticity refers to the brain’s ability to undergo functional and structural changes in response to external or internal stimuli from the environment or organs in the body37. Synaptic plasticity is recognized as a fundamental aspect of neuroplasticity. Long-term potentiation (LTP) is a primary manifestation of synaptic plasticity38. LTP changes promote structural changes and are likely to lead to alterations in the interconnectivity between neurons by activating formerly silent synapses or creating new ones39. This structural plasticity, represented by modifications in connectivity patterns, synaptic numbers, as well as changes in branching patterns of axons and dendrites40. tDCS induces mechanisms similar to LTP, which is believed to result from modifications at the dendritic level involving glutamatergic receptors such as the NMDA receptor, as well as an increased release of BDNF20. Furthermore, the normal brain achieves and maintains functional matching and balance between the two hemispheres through reciprocal interhemispheric inhibition. After stroke, the affected hemisphere shows reduced excitability, while the unaffected hemisphere exhibits excessive inhibition towards the affected hemisphere41. Anodal stimulation in tDCS typically increases neuronal excitability at the site of stimulation42. In this study, anodal stimulation was specifically targeted at the M1 region of the affected hemisphere, leading to a significant decrease in QEEG values compared to the pre-stimulation measurements in the tDCS group. Rizzo et al43. employed anodal tDCS stimulation on the affected primary motor cortex, effectively rectifying pathological interhemispheric inhibition and leading to improvements in upper limb motor function in participants. However, the cortical plasticity outcomes of non-invasive brain stimulation vary extensively both between and within individuals44. Despite the tDCS group showing less improvement compared to the other two groups, there was still a significant enhancement in upper limb function compared to the pre-treatment condition.
Links between induced cortical excitability changes and alterations of motor behavior or learning are hard to establish and there is a need for clear mechanistic bridges between the pre-conditioning of different circuits through brain stimulation and motor function and learning45. BCI can establish an alternative non-muscular channel between the participant’s brain activity and a computer, providing neurofeedback in a closed-loop. Non-invasive approaches (e.g., EEG, MEG, fMRI) have been commonly used in BCIs for stroke rehabilitation. These approaches involve both recording brain activity and controlling an external actuator, such as a robotic arm, for motor rehabilitation46. In essence, BCIs either help participants to learn to volitionally produce specific movement-related brain patterns, or they target brain structures and pathways thought to play an essential role in motor learning and motor control47.
BCIs for the induction of plasticity may be designed in two ways. The first encompasses an instruction to the user on how to produce specific types of brain signals. The second involves pairing specific, naturally occurring brain states with external stimuli45. Our study selected the latter paradigm in which subjects need to produce MI in order to activate M1 circuits in a task-related manner in order to enhance or promote sensorimotor rhythms coordinating the firing activity of large populations of neurons within a certain area48. According to João D et al.49, the activation of brain areas involved in action preparation during MI aligns with the involvement of the premotor cortex in neuroplastic changes associated with the recovery of function after a stroke50. Specifically, the KI we used in this study could induce a cognitive substitution wherein subjects perceive the virtual body's movement as their own. Notably, more evidence suggested that KI shared a common neural substrate for brain activity ae real movement51. Miyawaki et al.'s findings52 suggested that the impact of KI on upper limb motor function was mediated indirectly through its influence on spasticity. In this study, the attention scores of participants during motor imagery tasks were displayed in real-time on a screen, providing them with visual feedback of their brain activity. This neurofeedback intervention aimed to enhance the participants’ attention, potentially leading to improvements in their motor abilities53. Additionally, motor imagery can be considered as a cognitive task, while training involving active movements or active-passive movements (driven by a robotic arm) can be viewed as motor tasks. Therefore, the KI-BCI treatment protocol can also be seen as a dual-task paradigm of cognitive-motor integration. Studies have confirmed that cognitive-motor dual-task training is effective in promoting the recovery of both cognitive and motor functions54,55.
Based on the potential mechanism of tDCS and KI-BCI, we hypothesized that combining tDCS-induced synaptic plasticity with the activation of the motor cortex through KI-BCI could enhance the impact of each individual treatment on neuralplasticity. However, contrary to our hypothesis, participants who received both tDCS and KI-BCI treatments did not show significantly better upper limb function, which aligns with the findings of previous studies56. We speculate that this may be attributed to the need for a longer duration of treatment to achieve functional plasticity following tDCS-induced synaptic plasticity, which was not achieved in this study.
There are a few limitations in this study. First, this study had a relatively small sample size. Second, the number of EEG channels collected in this study was limited, which may restrict the precise localization of brain activity, signal interpretation, and detailed analysis, potentially reducing the statistical power of the study. Third, this study did not extract time-domain features from the EEG signals nor analyze the specific frequency band changes, which may have resulted in some changes going unnoticed.