About this Research Topic
The brain-machine interface (BMI) is an emerging technology that enables a direct communication pathway between the brain, external devices, or robots. BMI systems translate the activity of the brain to conduct action (active interfacing) or may serve to enable the detection of humans’ factors or cognitive states by the machines interacting with humans (passive interfacing). However, a single brain data can contain multiple kinds of artifacts, electrical noise, and environmental noise. Such a high variability and non-stationarity in the brain signals make its viability difficult for applications requiring their use for long periods. Some neuroimage modalities such as EEG can be highly impacted by artifacts and the quality of their signal can degrade across time. There are other novel neuroimage modalities such as functional near-infrared spectroscopy (fNIRS), which can also be made portable and are less affected by artifacts, however, their temporal resolution is slow and that makes them unsuitable for several BMI applications requiring a rapid rate of response. There is also a trade-off between the portability and level of comfort of neuroimage devices and the quality of their signals. That makes clinically grade EEG with many electrodes and wired set-up impractical for real-world BMI applications. In addition to the challenges related to the neuroimaging modalities, there are also issues in brain recordings that are related to human factors and ergonomics of human conditioning, such as tiredness, lack of concentration, lapses of attention. Patterns of activation in an individual can frequently change across sessions and experiments. Moreover, it has been suggested that the level of performance of a subject in a BMI task correlates to a specific cognitive profile, making many individuals unable to use BMI powered devices as they do not fit that prototypal profile. Several of these issues seems unsolvable without better neuroimaging modalities, however, the scientific community has demonstrated that by cleverly using novel artificial intelligence and machine learning methods, it is possible to overcome some of these issues. Some areas have greatly benefited from these advanced AI methods in recent years such as multi-modal BMI, multi-paradigm BMI, user-independent interfaces, long-term use brain-computer interfaces, human-in-the-loop brain interfaces to name a few.
This advanced BMI system can effectively empower new direct brain interfaces with assistive robots, rehabilitation robots, personal and companion robots, robotics swarms, and intelligent environments and ecosystems.
This Research Topic aims to solicit original research papers as well as review articles focusing on recent advances in AI to explore, understand, model, and develop BMIs from brain signals and related neuroimaging data. The main topics include, but are not limited to, the following:
• Application of AI in the analysis of brain signals from any functional or structural neuroimaging modalities (fMRI /MRI, PET/SPECT, EEG, MEG, fNIRS, DOI, EROS, etc.) to power BMI applications.
• AI to explore and understand the brain processes in cognitive neuroscience and computational neuroscience to improve BMI.
• Application of AI for brain-machine interfaces (BMI)
• Application of AI for Neurofeedback in BMI.
• Application of AI for Neural Rehabilitation and Rehabilitation Robotics.
• Application of AI in Clinical BMI and Assistive Neuro-robotics.
• AI for Humanoid Brain Sciences.
• AI for Intuitive Human-Robot BMI
• AI in Social Neuro-robotics.
• AI-based closed-loop BMI.
• AI to support Passive Machine Interfaces for the decoding of Cognitive States and Human Factors.
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.