«wili ch tool balbum mal laut hoerenzn» («I would like to listen to the Tool album ‘Loud’»). This seemingly banal sentence was uttered by a 36-year-old man with Amyotrophic Lateral Sclerosis (ALS) who had lost control of all his muscles because of the progression of the disease. In his state, termed complete locked-in syndrome, he had also lost all possibilities of communication; The exchange was made possible by using an experimental brain-machine interface. These interfaces measure, process and decode brain activity. This way, users can use communication software, prosthetic devices, or games by controlling their brain activity.

Advances in these technologies, long-time confined to research laboratories, are now being translated onto possible applications in the medical, industrial, and consumer-oriented domains. In April 2021, Elon Musk’s company Neuralink demonstrated with great pomp a monkey playing a simple video game using their experimental implant. Thus, reproducing results obtained several years ago in research laboratories with other technologies. Similarly, multiple less publicized clinical trials for similar devices pave the way for novel assistive and therapeutical approaches for people with mental disorders or cognitive disabilities. In turn, several companies currently commercialize direct-to-consumer headsets that record of brain activity; promoting their use as means for measuring levels of attention, guide meditation, or as gaming interfaces. Although many of these products are still in development and their efficacy is yet to be fully assessed, great expectations are put on the sector of neural technologies. The global neural technology market is expected to grow from USD 9.3 billion in 2020 to USD 21.9 billion by 2026.

Switzerland plays a leading role in the development of these technologies. World known research and clinical centers like the Center for Neuroprosthetics at EPFL, the Institute of Neuroinformatics at University of Zurich/ETHZ, the Wyss center, cantonal Hospitals in Vaud and Geneva and the SUVA Neurorehabilitation centers are responsible for key breakthroughs in the field. This is complemented by a thriving innovation ecosystem combining established actors and startups in the MedTech, digital health and data science sectors.

Brain-Machine Interfacing and Emerging Technologies

Brain-machine interfaces rely on several emerging technologies. A crucial component is the use of advanced sensors to measure brain activity. These sensors can be invasive, implanted within the body for medical or research purposes, or non-invasively placed outside of the body by means of headsets or electrodes on the skin. Significant efforts are being deployed to improve the safety, usability, and reliability of these sensors. Research in advanced materials, low-power electronics, optics, and nanotechnology are enabling a new generation of sensing possibilities.

Another major contributing technology is artificial intelligence, or more precisely machine learning; powerful algorithms that leverage data to create models for classifying patterns of brain activity onto commands. Typical challenges in applying machine learning include how to calibrate the models and generalize across users or environments. The lack of large-scale data sets limits the use of data-demanding approaches like deep neural networks, hence machine learning for brain machine interfacing is currently dominated for less data-demanding methods. Efforts to overcome this situation include the collection of large-scale data through research collaborations, use of consumer-oriented systems, or methods for transferring learning across models.

How Brain-Machine Interfacing Can Change Us

The confluence of artificial intelligence and brain-machine interfacing questions on how these technologies may change us as humans. For the first time in the history, we have means for interacting without using our physical body by having a direct pathway to our brain. Even though these means are rudimentary at the time being, it is worth considering their potential implications at the ethical legal and societal levels.

For example, the responsibility of actions or communications made through a brain-machine interface will be shared between the user and the interface. In case of an injury due to a brain-controlled prosthesis, we will face the challenge of identifying whether the user originally intended the actions that led to the injury, or if it was the result of erroneous decoding made by an algorithm. Technological, social, and legal considerations will be required to address these situations, attribute responsibilies and be able to resolve questions of accountability and liability of using these systems.

Decoding of brain activity makes it theoretically possible to obtain access to personal information without knowledge or consent of the user. Known threats to privacy brought by digital technologies are now enlarged by this new backchannel where malicious use of technology or cybersecurity weaknesses give access to private data. Specific considerations are thus required for protecting our mental privacy by explicitly addressing this specific threat.

Finally, these interfaces open a novel window to expand our understanding of the brain, and gain insights on how our mind and personality is defined by its function. Alteration of memories or behaviors by directly changing brain activity – for instance, through electrical stimulation – challenge our conception of where the mind resides and whether it is decoupled from mechanistic properties of our body organs. Philosophical questions about the existence of free will and the definition of self can now be further explored by the use of technology.

The Path to Robust, Responsible Brain-Machine Interfaces

Brain interfaces provide considerable opportunities to tackle mental health problems; nurture cognitive capabilities and increase our knowledge about brain and mind. It is of outmost importance to deploy the necessary means to steer these technologies towards solutions that are at once reliable, relevant, economically viable and respectful of the human dignity.
This requires coordinated efforts for supporting research and innovation, as well as sound regulatory and legal frameworks appropriate for emerging technologies. Several initiatives are currently trying to promote these necessary requirements through collaborative action of multiple stakeholders.

The International Brain Initiative gathers large-scale research efforts across the globe to advance research and development collaborations and promote adequate data-sharing governance mechanisms. Technical organizations like IEEE Brain Initiative and the International Neuroinformatics Coordinating Facility links groups of interests, academic and industry actors with regulatory agencies to produce technical standards for more robust development. 

At a policy level, the Organization for Economic Co-Operation and Development performed a large-scale multi-stakeholder effort to produce in 2019 the first international OECD Recommendation on Responsible Innovation in Neurotechnology. Non-governmental organizations like the European CLAIRE network in Artificial Intelligence and the Geneva Science and Diplomacy Anticipator are active for empowering researchers, policymakers and society at large for creating an innovation landscape for emerging technologies that will positively impact society.

We enter a new era where humans and machines will have seamless interactions; where privacy is less and less determined by physical constraints; and the way we experience and influence our environment will not require the involvement of our body. We have a unique opportunity for the advancement of ourselves and society, while challenging our conception of humanhood and self. It is a shared responsibility to make the best of the future brain-machine interfaces can provide us.

Dr. Chavarriaga acknowledges the support of the Digitalization Initiative of the Zurich Higher Education Institutions  (DIZH)