docs.google.com
BCI Definitions - Request for Input
Overall Purpose
The idea of a brain-computer interface (BCI) is becoming far more common in the public consciousness with the ever-increasing depictions of BCIs in popular news media, entertainment, and even governmental regulations. Further, BCIs have moved out of the purely academic world and towards commercial translation as an increasing number of private companies are developing BCIs for a variety of purposes. These BCIs include external devices that may be considered as exempt devices from a regulatory perspective, as well as surgically implanted systems that are regulated by national or international governmental agencies.
As the International Brain-Computer Interface Society, we believe it is important to put forward a clear and comprehensive definition of a brain-computer interface that reflects both the history and current state of the field, that simply captures the essential aspects of what a BCI is and is not, and that is useful for the Society in both internal and external communications. There have been numerous definitions of BCIs created by individual scientists in the field, or by other societies and organizations. Here, we intended to adopt or develop a definition that has broad acceptance within the academic community of scientists, engineers, clinicians that have an interest in BCIs.
Specific Goals and Objectives
1) Adopt or develop a definition of a brain-computer interface that is supported by the majority of the membership of the International BCI Society.
2) Provide a reference resource for BCI definitions that can be used by the Society and other stakeholders in publications, materials, interactions with media outlets, etc.
Existing Definitions
A thoughtful, well-supported, and well-motivated definition of a brain-computer interface was published by John and Elizabeth Wolpaw in 2012. It has been updated or presented in several forms, but the most referenced definition is:
A BCI is a system that measures CNS activity and converts it into artificial output that replaces, restores, enhances, supplements, or improves natural CNS output and thereby changes the ongoing interactions between the CNS and its external or internal environment.(Wolpaw and Wolpaw, 2012)
Other definitions exist within the field, but to a considerable extent, they overlap with the concepts and ideas described above. Further, there may be additional terms for different types or flavors of BCIs that could exist within this framework, with active and passive BCIs being one specific example.
Several key concepts can be drawn from the definition of a BCI provided above:
BCIs are output devices. The directional flow of information is from the brain to the external world.
Many potential use cases exist. Information acquired can be used for more than simply replacing a lost control channel.
Using a BCI changes the interaction between the user’s central nervous system (CNS) and their environment.
Current Landscape
There has been significant development in academic and commercial neurotechnology since this definition was proposed. This includes the creation of numerous companies developing BCIs. There have also been developments in the demonstrated capabilities of what are generally categorized as BCIs, including real-time speech synthesis, high-bandwidth communication devices, high-degree of freedom robotic arm control, somatosensory and visual feedback systems, and others.
There has also been a general acceleration of what could be considered the larger field of neural interfaces, in which engineered systems are being created that interface with the body throughout the nervous system. This has the potential to raise into question what is or is not a BCI.
Potential Questions or Limitations of the existing definition
The definition proposed by Wolpaw and Wolpaw raises several questions and presents a few limitations.
First, the CNS consists of both the brain and spinal cord, and most definitions of the CNS include the first and second cranial nerves. Further the retina itself is classified as being part of the CNS. However, neural interfaces, neuroprosthetics or neuromodulation devices that are deployed in the spinal cord, cranial nerves, or retina would not often be considered as a BCI. On the other hand, changing the interface target to be the brain may represent a significant change from the Wolpaw and Wolpaw definition.
Question 1: Should a definition of a BCI define the brain or the entire CNS as the interface target?
Second, the Wolpaw and Wolpaw definition recognized that ‘interaction’ is an important component of BCIs. According to that definition, a device that simply monitors signals from the brain is not a BCI. A BCI “changes the ongoing interactions between the CNS and its external or internal environment.” However, interactions between the CNS and the external or internal environment are always (or very nearly always) bidirectional. Sensation drives motor action, which drives new sensations. Monitoring brain activity and providing neurofeedback to the user elicits a reaction from the CNS. If the purpose of the CNS is to acquire and process sensory information and then respond to that information with appropriate outputs, then a definition of a BCI that addresses only the role of the CNS in generating movement is perhaps unnecessarily limiting. Indeed, ‘direct input’ systems were briefly discussed in Wolpaw and Wolpaw 2012, but at the time, this was a nearly non-existent field.
Question 2: Should a definition of a BCI include a description of a system that (only, or in addition to reading and decoding brain signals) modulates or modifies brain function?
Revised Definition Proposal
Below are revised definitions of a BCI that minimally modify the Wolpaw and Wolpaw 2012 definition, but that address one or both questions above.
A definition that focuses specifically on the brain as compared to the entire CNS.
A BCI is a system that measures brain activity and converts it online into artificial output that replaces, restores, enhances, supplements, or improves the natural output of the brain and thereby changes the ongoing interactions between the brain and its external or internal environment.
A definition that incorporates bidirectional interfaces.
A BCI is a system that measures CNS activity and converts it online into artificial outputs, and/or modifies CNS activity using artificial inputs, to replace, restore, enhance, supplement, and/or improve natural CNS outputs of the brain, or to perceptually alter the inputs of the brain, thereby changing the ongoing interactions between the CNS and its external or internal environments.
A definition that incorporates both modifications.
A BCI is a system that measures brain activity and converts it online into artificial outputs, and/or modifies brain activity using artificial inputs, to replace, restore, enhance, supplement, and/or improve the natural outputs of the brain, or to perceptually alter the inputs of the brain, thereby changing the ongoing interactions between the brain and its external or internal environments.
Other General Comments
It may not be necessary to develop a new term to describe input systems, such as a computer-brain interface (CBI). While the term brain-computer interface can (and often has) been associated with information flow from the brain to a computer, there may be no reason to insist upon a directional interpretation of the term, apart from historical precedent.
Definition of the brain: The brain includes the cerebrum, the diencephalon, the brainstem, and the cerebellum. It does not include the cranial nerves. Furthermore, it does not include the measurement of intention or information (that could have originated in the brain) obtained from downstream areas of the nervous system (e.g. peripheral nerves or muscles).
There is no modality limitation to what type of technology or signal sources might form the basis of a BCI system, as long as they interact with the brain itself. For example, a BCI system could be based on one or more of the following (not exhaustive) existing technologies or methodologies: TMS, tDCS, EEG, ECOG, FUS, MEG, fMRI, SEEG, microelectrode arrays, optical imaging, optogenetics.
The definition of Wolpaw and Wolpaw, as well as the suggested alternative definitions, do not require the measured brain activity to contain information about the ongoing intention of the user, because this would exclude technologies referred to as passive BCIs from the definition. The consequence of this is that technologies such as closed-loop neuromodulation (e.g., responsive DBS) may fall under the definition of a BCI. Suggestions on the need to distinguish between these types of technologies, and how to differentiate between them, are welcomed.
The term ‘online’ in the proposed definitions is meant to reflect ‘(nearly) real-time’, not ‘on the internet’.
Other Relevant Resources
https://www.gao.gov/products/gao-22-106118
References
Wolpaw, J.R., Wolpaw, E.W., 2012. Brain–Computer Interfaces Principles and Practice 3–12. https://doi.org/10.1093/acprof:oso/9780195388855.003.0001
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------
At this stage, the board of the BCI Society is seeking input from BCI Society members, as well as other BCI stakeholders, on the above considerations and alternative definitions of a BCI. We invite all of you to answer the questions below.
NOTE: We are asking for your email and name as we may wish to follow up with you for clarification or to discuss these ideas further. Responses will only be seen by board members and will not be communicated to or shared with other parties.
