Human enhancement through the lens of experimental and speculative neurotechnologies

Abstract Human enhancement deals with improving on and overcoming limitations of the human body and mind. Pharmaceutical compounds that alter consciousness and cognitive performance have been used and discussed for a long time. The prospect of neurotechnological applications such as brain‐steered devices or using invasive and noninvasive electromagnetic stimulations of the human brain, however, has received less attention—especially outside of therapeutic practices—and remains relatively unexplored. Reflection and debates about neurotechnology for human enhancement are limited and remain predominantly with neurotech engineers, science‐fiction enthusiasts and a small circle of academics in the field of neuroethics. It is well known, and described as the Collingridge dilemma, that at an early stage of development, changes can easily be enacted, but the need for changes can hardly be foreseen. Once the technology is entrenched, opportunities and risks start to materialize, and the need to adapt and change is clearly visible. However, carrying out these changes at such a late stage, in turn, becomes very difficult, tremendously expensive, and sometimes practically impossible. In this manuscript, we compile and categorize an overview of existing experimental and speculative applications of neurotechnologies, with the aim to find out, if these real or diegetic prototypes could be used to better understand the paths these applications are forging. In particular, we will investigate what kind of tools, motivations, and normative goals underpin experimental implementations by neurohackers, speculative designers and artists.

community or neurohackers. While EEG is just one example within a wide range of tools that can be used for neurotechnology, the decreasing costs and increasing availability of devices-including g.tec since 1999 (UnicornBI, 2019a), NeuroSky since 2004 (NeuroSky, 2019), Emotiv since 2011 (Emotiv, 2019), OpenBCI since 2014 (Crunchbase, 2019), and Muse since 2014(O'Rourke, 2015-as well as the increasing number of available applications (as shown in this paper), lead to questions about the motivations behind DIY use and projects. An overview is relevant for both policymakers who want to regulate the DIY use in order to control the risks at an early stage (see Collingridge dilemma [Worthington, 1982]), and for companies that sell neurotechnology and therefore need to understand their customers' needs. To get a better understanding of the directions that neurohackers' and designers' neurotechnology projects are taking, this paper examines current applications of neurotechnology, categorizes them and focuses specifically on applications and concepts that surpass or move beyond pure medical or therapeutical purposes.
While neuroscience generally attracts considerable public interest, this is particularly true because of the rapidly expanding discourse on the merging of human corporeality with technology. In this discourse, neurotechnology often features as the harbinger of a future in which the body is transformed in a process of ever-increasing "technologization." It is expected that such technologization will happen in two ways: 1. Invasive: technological modifications of the body in which surgical interventions allow technology to replace or augment bodily functions (implants, prostheses).
2. Noninvasive: the use of technologies that modify the body without such interventions (noninvasive neurostimulation technologies), and brain-machine interface technology capable of coupling humans and artifacts.

| TOOLS
With the wide availability of neurotechnology tools for DIY use such as electroencephalography (EEG) based brain-computer interfaces (BCI) and transcranial direct current stimulation (tDCS), the possible applications are more and more only dependent on the available or programmable software and, of course, imagination. Sites such as GitHub.com provide a wide range of free downloadable software codes to use and many websites provide tutorials on how to use neurotechnology tools (NeuroTechX, 2019). Therefore, even people without any sophisticated background in neurology or computer science can start to use these applications at home.
An EEG headset consists of one or multiple electrodes placed on the scalp, a ground electrode and a reference electrode. The electrodes measure the changes in voltage potential created by the ion current within the nerves of the brain. More electrodes allow more detailed measurements. Simple headsets can measure the level of general focus or relaxation of an individual and show this as biofeedback. If the focus surpasses a defined threshold, one could couple an automated action to it such as extinguishing a candle (Chierico, 2014). Other biofeedback systems are based on the visual cortex: if the wearer of the EEG looks at a screen with blinking lights, the frequency of the blinking can be traced in the brainwaves in the visual cortex, thereby determining what the wearer is focusing on. This can be used to select something on a screen with multiple different options (displayed with different frequencies) to make decisions (Emmerson, 2018).
Where EEG headsets measure brain activity, transcranial direct current stimulation (tDCS) is an example of a technique that allows one to influence/stimulate brain activity. tDCS uses at least two electrodes placed on the scalp that apply a small current (0.5-2 mA) through the brain, which can be used for a wide range of effects, some of which are still poorly understood (Dubljevi c et al., 2014). This technique is closely related to transcranial alternating current stimulation (tACS).
There are other neurotechnological tools including transcranial magnetic stimulation (TMS), functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG) and spinal cord stimulation. However, to our knowledge, they are, mostly due to the costs and sometimes regulatory hurdles, currently not available to neurohackers.
If the thresholds of access and accessibility are lowered, tools and applications can be appropriated by individuals from a variety of fields.
This can be observed, for example, in the use of neurotechnology in art and art-science projects at the annual Ars Electronica Festival in Linz, Austria, an event well known for gathering artists, designers, and scientists to experiment, explore and innovate (ARSElectronica, 2019).

| METHODS
To ensure broad coverage of scientific research, artists' work and DIY use of neurotechnology, a large number of different sources and databases were used for this work. First, all the exhibitions from the Ars Electronica festival (2013)(2014)(2015)(2016)(2017)(2018) were examined in search of neurotechnology related projects, as Ars Electronica is the largest and most prestigious media arts festival, frequently linking art and science in unusual ways. Second, vimeo.com, newness.com and youtube.com were used to find art and design projects around neurotechnology. Vimeo.com showed the best and most results, which were narrowed down by selecting videos that were posted no longer than 4 years ago, and which seemed to be an art-science project or neurotechnology product. Furthermore, the first two pages of the category tabs "Art & Design" and "Animation" were checked. Some of the videos were the outcome of actual scientific research, which often led to peer-reviewed papers. DIY use was found by snowball sampling through different forums such as www.reddit.com and learn.neurotechedu.com. The latter houses a long list of examples of neurotechnology. These examples led to other projects that were related or functioned as inspiration. Often, DIY projects referred to scientific work to back-up or substantiate their claims.
In addition, Google Scholar was used to research neurotechnology techniques and related technosocial models (such as Collingridge's dilemma and Technology Adaption Lifecycle) covered in this article.
Research into neurotechnology techniques also opened up the existence of additional applications. The majority of the peerreviewed articles we found was in the medical field; however, as they were often purely medical we decided to discard them based on a lack of enhancement opportunities. The reference sections of the peerreviewed articles were searched for additional articles. Finally, the websites of the applications and tools showed us the scientific research they based their claims upon.

| Applications
To get a comprehensive overview of neurotechnology in society we The category "medical" encompasses neurotech applications and methods that have clear cut diagnostic and therapeutic purposes. On the other end of the spectrum, "science fiction" refers to extremely speculative or even unrealistic applications, barely grounded in technological possibilities or even surpassing the laws of physics.
The remaining three categories are of most interest for the scope of this manuscript. First, a quasi-medical category is defined, which is called Med+, because applications in this group have originally had a medical approach or purpose, but also have potential to go beyond healing or restoring functions; they can allow enhancement of specific functions (e.g., hearing ultrasound). The second category, called Enhancement, considers applications that are nonmedical, but specifically serve other purposes, such as enhancement of senses or interactions with the environment. The third category is based on research and art projects that are not yet realizable and is therefore called Speculative. It points to applications that might occur in the future, but includes work based on current research or prototypes; this category encompasses scientific research (SR) as well as speculative design (SD).
The three categories can further be subdivided into invasive and noninvasive technologies. See Table 1 for an overview of the number of cases identified. In general the number of cases found for noninvasive applications clearly surpass the invasive types.

| RESULTS
Based on our online and literature search we identified 56 applications in the three categories Med + (13), Enhancement (22) and Speculative (21) (see Table 1). The following section will give an overview of the identified applications and highlight a few representatives and/or outstanding cases in all of the three categories.

| Med+
Where there are many medical applications of neurotechnology available, the Med + category only include medical applications that have a prospect to provide enhancement over normal human capabilities (see Table 2). Often this is found in the restoration and improvement of senses such as hearing and sight. Since the first invention of cochlear implants in the 1960s, and seminal developments in the 1970s, cochlear implants have now become increasingly used (Mudry & Mills, 2013), restoring hearing in deaf patients. In the early 2010s, this is followed by a retinal prosthesis called Argus II (SecondSight, 2013), which (partly) restores vision for retinitis pigmentosa patients (Weiland & Humayun, 2014). These are both examples of technically advanced medical implants that restore lost functions. However, such implants might in the future also be used for enhancement, allowing the individual future patients of a visual prosthesis, for instance, to see in more detail or in frequencies outside the spectrum normally available to humans (e.g., infrared or ultraviolet).
The cochlear implant, for example, does partially restore hearing loss, but can under specific circumstances, such as in a very noisy environment, endow the patient with better than human hearing capabili- hears through bone conduction. This device partly restores his color vision, but also allows him to "see" (or rather: hear) color such as infrared and ultraviolet light (Jeffries, 2014) that are not part of the spectrum visible to humans. Implanting an experimental device such as the cyborg antenna is not something that is easily done: Harbisson was rejected by bioethics committees before he found, "a doctor willing to do the surgery anonymously, and [he] found one in Barcelona." (Gartry, 2015 Speller software, for instance, to be used together with an EEG headset is one of those applications available for consumers and often used at the BR41N.IO hackathon. Originally designed for use in the medical field, for example, for locked-in patients, it can be used to spell words by watching a screen and counting the moments a specific letter or symbol blinks. This counting can be registered via the p300 paradigm by the EEG, and the correlation with the blinking of the icon on the screen allows the software to determine which symbol the users want to choose (Ortner, Aloise et al., 2011;Ortner, Prueckl et al., 2011). This is a rather slow approach to communication, but useful for people unable to speak or move. Another example is Alessio Chierico's interactive installation "Trataka," wherein the user wears a BCI device and is asked to focus on a flame. An airflow is directed to the flame, and once the user puts full attention on it, it is extinguished.
A technique using blinking frequency and synchronization in the visual cortex captured by EEG (as explained in the Tools -EEG section) allowed motor-impaired patients to participate in making music with a string quartet (Emmerson, 2018 -Increase learning for learning disabled children (Fernández et al., 2003) -Controlling a prosthetic arm by thought (Levy & Beaty, 2011) -Rehabilitation for stroke patients (Kober et al., 2015) -Controlling a robotic arm by thought for paralysis (Chadwick et al., 2011) -Determine consciousness of people in coma/locked-in syndrome (Guger et al., 2017) -Decrease stress and increase sleep (Tyler et al., 2015) -Paralyzed influence music by brain waves (2018) -Spell by thinking/watching the Unicorn Speller ( Ribas is known for her implants, which are connected to online seismographs and which resulted in giving her additional senses that allowed her to "see" movement behind her back and to feel earthquakes all over the world (CNN, 2018;Garcia, 2015;Quito, 2016).
Among the most accessible implants we find the insertion of magnets, for example in the tip of a finger (Robertson, 2017).
This allows the individual to feel magnetic forces and also control magnets or magnetizable objects. DIY YouTube videos are available to teach interested parties how to implant a magnet by themselves (TheThoughtEmporium, 2017).
Closely related to implanting magnets is Cyborg Nest's "North Sense," a device pierced into the chest that vibrates when the wearer is facing north. The mission statement of this self-titled "Mindware Company" is to, "contribute to human evolution. By experiencing the hidden parts of nature through new senses, we will evolve towards a richer life experience." (CyborgNest, 2018).
Apart from implants, be they integrated into One practical possible future application arising from the hackathon is the control of home products also referenced under the phrase "Smart Home." This application allows users to measure brain waves with an EEG headset and to put on any device including lights, a heater or an AI helper like Apple's Siri. The applications of the hackathon make use of the "Unicorn Speller," which, as described previously in the section "Med+," requires users to focus on specific icons flashing on a screen.
While companies such as the streaming service Netflix (2019) and some music video clips (Coldplay, 2014) are experimenting with userdefined narratives-meaning that the user influences the way the story develops-, neurotechnology is doing the same, however, working with the unconscious. Rachel and Richard Ramchurn's "The MOMENT" project is a brain-controlled film: Based on brain activity and blinking measured by EEG headsets, the film can be uncon- and applied electrodes to their arms. One of them could move the other's arm. This is done by measuring the electrical current in the first person's arm, while they are moving it, and applying a correlated electrical current to the second person's arm (Gage, 2015). While this connection uses electrodes on muscles, as early as 2013, researchers at the University of Washington achieved a human-to-human brain interface using EEG on the sender's end and transcranial magnetic stimulation on the receiver's end (Armstrong & Ma, 2013). 2 Finally, we find another method in enhancements: transcranial direct current stimulation (tDCS). Closely related to the tDCS applications for medical purposes, tDCS is used in a similar way for enhancement. When used in the right way (and what exactly the right way is, is still up for debate), tDCS can lead to enhanced planning ability (Dockery et al., 2009), enhanced working memory (Fregni et al., 2005) and enhanced learning capacity (Meinzer et al., 2014). Users report they experience increased ease, fewer distractions and a decreased number of "background" thoughts (Adee, 2012). Most importantly for the purposes of this paper, tDCS devices are readily available and, ranging between USD 150 and USD 300-or building one yourself for just the material costs 3 -, and thus fairly accessible in terms of pricing, as well as reviewed in recommendation lists online (neurogalMD, 2018; TotaltDCS, 2019). Table 3. the core of most of these projects encompasses measuring the brain waves of two persons through EEG data and asking them to actively try to synchronize their brain states. In one project this is done by dancing the tango, in two other projects people are allowed more liberty to find their own way of connecting (Dikker & Oostrik, 2019). This is also related to research done on the effect of brainwave synchronization between teacher and students on learning outcomes.

| Speculative
It is shown that social closeness between teacher and student are of much higher influence on the learning outcomes than brain wave synchronization (Bevilacqua et al., 2019). Even more speculative is the concept of an installation designed by Susanna Hertrich over the span of several years titled "Prostheses for Instincts" that explores a device attached to the human body, which creates a new sense of awareness and emotional extension: The inputs vary from stock market data, currency exchange rates, natural disasters or crime rates, and a change of the data input (like a spike or drop) incites the device to react. The user then experiences sensations that are similar to physical reactions to immediate danger: goosebumps, shivers etc. (Hertrich, 2019).
Hertrich's work is closely related to research done by David Eagleman that uses a haptic vest to input information into the back of the wearer (Novich & Eagleman, 2015). The research explores the possible information streams that our brains can learn to understand and feel as an additional sense. In a TED talk, Eagleman hypothesizes about how this could be used to get a feeling for multi measurement data, like all the airplane measurements for pilots, all quality measurements of the International Space Station (ISS) for an astronaut, or a common feel for changes in the stock markets. Due to the many variables influencing these measures it is difficult for humans to consciously analyze them, but the vest could give us a feeling for it that we might understand (Eagleman, 2015).  (Emmerson, 2018), for pain relief (Kapural et al., 2010), or, in fact, unraveling the mystery of consciousness (or unconsciousness) (Guger et al., 2017). But Med + devices can also work beyond the conventional medical uses-a cochlear implant can function as an aid to hear sounds, not within the usual human accesible frequency range, and the research put into a retinal prosthesis may well serve as a base for bionic eyes. The noninvasive devices in this category have also been first conceptualized from a medical or therapeutical standpoint, but hold the potential or actually can be used for other purposes.

| MOTIVATIONS
On the border between Med + and Enhancement, we find health tracking, self-optimization or self-monitoring as the driving forces.
Similarly to how users around the world are currently using apps and smart watches to track and self-optimize their selves, these activities could (in the near future) also be done via invasive or noninvasive neurotechnology. These activities are mainly motivated by self-control and self-improvement (Wexler, 2017) or self-actualization, and are mostly embedded in a competitive market economy environment.
The Enhancement category is defined by practical reasons, identity or self-esteem factors, but also the motivation to change how life is experienced. Practical reasons are defined by goals that range from the general enhancement of human abilities and intelligence
Another reason for Enhancement neurotechnology is for the sake of a gimmick or to perform a party trick (Robertson, 2017). This can be motivated by the desires to have fun, gain respect and attention or build self-esteem, for instance. Here, neurotechnology gains attraction from people because of its futuristic elements. People might feel the desire to have the latest gadgets before anyone else (Thaddeus-Johns, 2017). Others may wonder how the addition of extra senses will affect their perception and enjoyment of reality.
They may want to discover the limits of human perception, and whether implants or additional senses can lead to new forms of expression, creativity or understanding in general. In the case of invasive technologies and implants, another possible consideration for individuals may be the desire to attain the next level of body modification-beyond tattoos and piercings (Thaddeus-Johns, 2017).
Another group of distinguishable motivations is strictly found with DIY or home users of neurotechnologies. Similarly to what is seen in other fields of DIY science, these groups explore neurotechnology to democratize the tools of science, increase learning outcomes (Wexler, 2017), adapt the tools to their specific needs, and create applications without a commercial interest. While for the most part the intentions of the hackers are benign and constructive, they often lack in professional knowledge and skills, which can lead to some unintended self-harm, for example when attempting to remove an implanted magnetic bead (aixre, 2017a, 2017b).
Finally, the category Speculative firmly points to the future, with speculative design concepts that explore possibilities. Some of the motivations overlap with motivations within the Enhanced group, such as the exploration of expression and potentials: additional senses (Hertrich, 2019), how it might "feel" to deal with multiple data inputs from unusual sources (Eagleman, 2015;Eagleman & Novich, 2019), what brainwave synchronization or interaction between two people may feel or look like (Dikker & Oostrik, 2019;Lancel, 2019). But the Speculative category also involves existing scientific research that serves as a basis for potential future applications. In the case of devices like AlterEgo, a wearable headset that can translate subvocalizations, or "silent speech," and can also reply "silently" via bone conduction, the practical motivations are found in enhanced privacy and the fact that devices are strictly personal and not influenced by distance or background noises (Kapur et al., 2018). This consideration also applies to all implanted devices. And yet, the research, conducted at the Massachusetts Institute of Technology (MIT), also speaks to future applications that could come close to how we imagine "telepathy" (Kapur et al., 2018).
This is also the only category, which includes selected examples of animal research, because they contain a speculative dimension toward T A B L E 4 Identified neurotechnology applications in the category Speculative
These experiments may or may not see applications in humans, though given the fact that China is conducting a first clinical trial using experimental deep brain stimulation (DBS) on drug addicts with mixed reports, the topic seems of some relevance (Navarro, 2019

| Normalization levels
Yet, in this text, the neurotechnological applications discussed move beyond the purely medical, and are explored outside that field. The cases collected and categorized here are diverse and encompass both invasive and noninvasive technologies, goal-driven as well as speculative applications. Whether or not these become normalized, is not easy to answer, as multiple disciplines field the issue of what a "social norm" is, or more specifically means, differently: in law social norms can be seen as something that is necessary to modify, in order to keep up with socio-economic, environmental or technological changes (Spector, 2018). In psychology "normality" is frequently defined by juxtaposing it with "abnormality," and diagnosis of the severity of that abnormality based on several criteria including statistical infrequency or violation of social norms (McLeod, 2018). Within Sociology what is "normal" is defined as collective or individual perceptions of acceptable conduct, largely dependent on social norms, which in turn, can often be endorsed separately or additionally in smaller groups, and are often situational dependent (Hechter & Opp, 2005;Schultz, Nolan, Cialdini, Goldstein, & Griskevicius, 2007). Instead of asking whether a certain technology is "normal" or not, or whether a neurohacker and cyborg is perceived as "normal" or not, it serves the scope of this Apart from the crucial work of sense-making and building a community of practice through such experiments, the first fun applications of basic research are in fact essential to overcome the trough of disillusionment within the Gartner hype cycle (Fenn, 2007)-which follows the peak of inflated expectations-by gathering enough early adopters as indicated by the technology adoption lifecycle (Rogers, 2003) for the further development of the field. These gimmicks are perfect examples of what kind of everyday applications can be expected in the near future. "Near future" is indeed appropriate for these applications, as they are already available and most often easy to use or program. The claim of "easy to use" is emphasized by the text inviting people to join one of the BR41N.IO hackathon events: "Anyone can participate who has interests in BMI (brain machine interface) […] Participants do not have to be a BMI expert to participate on a team!" (BR41N.IO, 2019).

| Final thoughts
A study in the area of 3D-printed weapons has shown how difficult it is to regulate DIY activities once published online (Bryans, 2015). To a large extent the same applies to the area of DIY neuroscience (and other DIY fields), as both practices are done at home and the materials are of everyday use.
In the near future, the influence of neurotechnology might increase due to the development and accessibility of the technology.
Where some argue that by increased performance of wearables, invasive technology will become less popular (Robertson, 2017), other research shows that 8 out of 10 smartphone users envision a near future with implanted augmentation or body monitoring devices (internables) (Ericsson, 2015). Either way this would increase the use of neurotechnology and the need of more focus and care from users, developers, companies, and policy makers. Even though neurotechnology might have already been used for centuries (Sarmiento, San-Juan, & Prasath, 2016), in most aspects it is still in its infancy. As described by Collingridge's dilemma of control, this is the time to shape the future of the technology (Worthington, 1982). It requires responsibility of its users and for developers to focus on their customers, for example, as outlined in the Responsible Research and Innovation (RRI) frameworks (Wickson & Carew, 2014).

ACKNOWLEDGMENTS
The authors acknowledge the financial support from the ERA Net 2 This work is continued in animal experiments listed in Table 4, coupling animal brain via brain implants.
3 However, we certainly do not encourage readers to try this at home. 4 Experience from the past, however, shows that it does not take very long for someone to hack a new biometric system, for example, fingerprints, retina scan, vene scans, and so forth. See also: https://media.ccc. de/v/35c3-9545-venenerkennung_hacken or https://media.ccc.de/v/ biometrie-s8-iris-en