Motor Imagery (MI) BCI is a brain-computer interface paradigm that enables users to control external devices through the mental simulation of movements, without actually performing any physical movement. This approach leverages the fact that the brain's motor networks activate similarly whether performing a movement or simply imagining it[1][2].
Motor Imagery BCI is particularly valuable for neurodegenerative disease applications because it provides a non-invasive communication channel for patients with severe motor impairments, enables neuroplasticity-based rehabilitation, and can be used for long-term monitoring of motor cortex function.
When a person imagines performing a movement, the motor cortex activates in patterns similar to actual movement execution[1:1]:
Motor imagery engages synaptic plasticity through BDNF-mediated long-term potentiation, enhancing the effectiveness of BCI-based rehabilitation.
Motor imagery primarily modulates two EEG frequency bands[2:1]:
The strength of mu/beta modulation correlates with the vividness of motor imagery and is highly variable between individuals, making calibration essential.
The most common approach uses scalp EEG electrodes placed over the sensorimotor cortex[2:2]:
Standard Electrode Positions:
Advantages:
Limitations:
For higher precision, electrocorticography (ECoG) arrays can be implanted on the brain surface[3]:
Advantages:
Applications:
Functional near-infrared spectroscopy can complement EEG for motor imagery[4]:
Raw EEG signals require several preprocessing steps[5]:
Signal Acquisition -> Bandpass Filtering (8-30 Hz) -> Artifact Rejection -> Spatial Filtering
Bandpass Filtering: Removes DC offset and high-frequency noise
Artifact Rejection: Removes eye movements, muscle artifacts
Common Spatial Patterns (CSP): Enhances discriminability
Key features extracted from motor imagery signals[5:1]:
| Feature Type | Description | Application |
|---|---|---|
| Band Power | Power in mu/beta bands | Primary feature for MI |
| ERD/ERS | Event-related desynchronization/synchronization | Movement prediction |
| Phase Locking Value | Phase synchronization | Connectivity analysis |
| Coherence | Frequency-specific connectivity | Network analysis |
Common classifiers for motor imagery[5:2][6]:
Motor imagery BCI has emerged as a powerful tool for stroke rehabilitation[7][8]:
Mechanism:
Evidence:
For ALS patients, motor imagery BCI provides[9]:
Considerations:
Motor imagery in PD serves multiple purposes[10]:
Research Applications:
Motor imagery applications in AD are emerging[11]:
Motor imagery BCI applications in FTD are emerging as a research frontier[12]:
Research Considerations:
Motor imagery in Huntington's disease (HD) serves both diagnostic and therapeutic purposes[13]:
Clinical Applications:
Evidence:
Standard motor imagery tasks include[1:2]:
Session Structure:
Calibration:
| User Group | Typical Accuracy | Factors |
|---|---|---|
| Healthy Adults | 70-90% | Training, user ability |
| Stroke Patients | 60-85% | Residual function |
| ALS Patients | 55-80% | Disease stage |
| Novice Users | 50-70% | Initial session |
| System | Channels | Specialization | Cost |
|---|---|---|---|
| g.tec g.tecUSBAmp | 16-64 | Research grade | High |
| BCI2000 | Variable | Flexible | Free |
| OpenBCI Cyton | 8-32 | Open source | Medium |
| Emotiv EPOC+ | 14 | Consumer/research | Medium |
| Muse | 4 | Consumer | Low |
Motor imagery BCI may not be suitable for[14]:
Adaptive Algorithms: Machine learning that adapts to neural changes[6:1]
Hybrid Systems: Combining motor imagery with other paradigms[15]
Closed-Loop Rehabilitation: Real-time feedback integration[8:1]
Tele-BCI: Remote rehabilitation and monitoring
Pfurtscheller G, Neuper C., Motor imagery and direct brain-computer communication. Proceedings of the IEEE 2001. 2001. ↩︎ ↩︎ ↩︎
Wolpaw JR et al., Brain-computer interfaces for communication and control. [Clinical Neurophysiology 2002](https://doi.org/10.1016/S1388-2457(02). 2002. ↩︎ ↩︎ ↩︎
Leuthardt EC et al., The emerging world of motor neuroprosthetics. Neurosurgical Focus 2009. 2009. ↩︎
Naito M et al., A hybrid EEG-NIRS BCI system. International Journal of Bioelectromagnetism 2007. 2007. ↩︎
Lotte F et al., Review of BCI signal processing. IEEE Transactions on Biomedical Engineering 2018. 2018. ↩︎ ↩︎ ↩︎
Schirrmeister RT et al., Deep learning with convolutional neural networks for EEG decoding. Human Brain Mapping 2017. 2017. ↩︎ ↩︎
Pichiorri F et al., Motor imagery-based brain-computer interface robot rehabilitation in stroke. Brain Stimulation 2015. 2015. ↩︎
Cervera MA et al., Brain-computer interfaces for post-stroke rehabilitation. Journal of Neural Engineering 2018. 2018. ↩︎ ↩︎
Wolpaw JR, McFarland DJ., Control of a two-dimensional movement signal by a non-invasive brain-computer interface. PNAS 2002. 2002. ↩︎
Brunner P et al., A review on BCI in stroke rehabilitation. Clinical Neurophsyiology 2015. 2015. ↩︎
Ray LV et al., Motor imagery in Alzheimer's disease. Journal of Alzheimer's Disease 2021. 2021. ↩︎
Rascună C et al., Motor imagery in frontotemporal dementia. Frontiers in Neurology 2022. 2022. ↩︎
Squarcini L et al., Motor imagery deficits in Huntington's disease. Journal of Neurology 2021. 2021. ↩︎
Blankertz B et al., The Berlin Brain-Computer Interface. IEEE Transactions on Biomedical Engineering 2007. 2007. ↩︎
Pfurtscheller G et al., Hybrid BCI approaches. Frontiers in Neuroscience 2010. 2010. ↩︎