Influencing Human Behavior with Noninvasive Brain Stimulation: Direct Human Brain Manipulation Revisited

Abstract

The use of tools to perturb brain activity can generate important insights into brain physiology and offer valuable therapeutic approaches for brain disorders. Furthermore, the potential of such tools to enhance normal behavior has become increasingly recognized, and this has led to the development of various noninvasive technologies that provides a broader access to the human brain. While providing a brief survey of brain manipulation procedures used in the past decades, this review aims at stimulating an informed discussion on the use of these new technologies to investigate the human. It highlights the importance to revisit the past use of this unique armamentarium and proceed to a detailed analysis of its present state, especially in regard to human behavioral regulation.

Introduction

Many interventions developed to treat neurological or psychiatric disorders can also transiently affect and even enhance cognitive functions or behaviors in otherwise healthy individuals. Legal access to approved pharmacologic agents for such enhancement of human function is generally limited by the need for a prescription by a physician. Even so, there are ample cases of abuse or at least questionable practice, for example, in regard to the use of stimulants (e.g., Hyman 2011). However, for a number of reasons including current gaps in regulations, such challenges seem substantially more concerning for devices. The history of such device-based approaches is rich, and it is tempting to argue that current approaches have overcome the limitations of the past and can now deliver on the promise of safe and effective enhancement of human behavior while preventing past undesirable effects. Is that truly the case?

Some of the presently available devices that enable noninvasive brain stimulation are relatively easy to manufacture and have given rise to a direct-to-consumer industry and a do-it-yourself movement that has begun to capture the imagination of individuals eager to improve their abilities in various domains and their performance in a wide array of activities. This is likely only the beginning. Powerful new technologies, such as optogenetics, are facilitating unprecedented insights into investigation of neural circuit function. Globally, under the auspices of recently announced large-scale “brain projects,” not only animal models but also human brain circuits will gain greater attention whether computationally (e.g., European Brain Project), through new basic tool development (e.g., US BRAIN project initiatives funded by the National Institutes of Health [NIH] and National Science Foundation), or as targets of new devices meant for clinical deployment (e.g., US BRAIN project initiatives funded by the NIH and Defense Advanced Research Projects Agency). Thus, the near future will almost certainly see the optimization of existing methods and the development of new invasive and noninvasive technologies for brain circuit modulation. Although initially oriented toward research or disease treatment issues, these novel procedures have the potential to influence normal function, possibly outside medical supervision. This article aims at contributing to such efforts by providing a reminder of the fascinating but often forgotten history of brain circuit modulation and describing current use of these devices in research in order to encourage a discussion on the possible lessons that can be drawn from such a perspective.

History of Direct Human Brain Manipulation

Early Research Programs

In the 17th and 18th centuries, pioneers such as Giovani Aldini (1804) led investigations on animal electricity that paved the way to human brain stimulation (Parent 2004; Fig. 1). This approach underwent a marked expansion in the first half of the 20th century, when various novel brain manipulation procedures were developed and applied to animal models as well as to patients with brain disorders. For example, in the 1920s, Walter Rudolf Hess from the University of Zürich proved that it was possible to manipulate cat’s behaviors, such as rage, hunger, and sleepiness, with cortical electrical stimulation, and was awarded a Nobel Prize in 1949 for this work. Egas Moniz from the University of Lisbon shared the 1949 Nobel Prize for his development of leucotomy, a form of psychosurgery that severed prefrontal white matter. This initiated a period in which severe mental illness was often treated with prefrontal “lobotomies.” Despite a broad rejection of the indiscriminate use of psychosurgery, smaller and more carefully made brain lesions continued to be studied and are still used today (e.g., in treatment refractory obsessive compulsive disorder; Sheth and others 2013). In the 1950s, Jose Manuel Rodriguez Delgado from Yale University invented brainchips, which he called stimoceivers or intracerebral radio stimulators. Following animal work, he implanted patients, most of them with schizophrenia or epilepsy, with the therapeutic goals of modulating emotions and manipulating behaviors (e.g., reduce aggressiveness; Delgado and others 1968). Later, collaborating with Frank Ervin from Tulane University and Harvard University/Massachusetts General Hospital and Vernon Mark from Massachusetts General Hospital, Rodriquez Delgado created chemitrodes, which integrated chemical and electrical stimulation. They reported they could induce fear, rage, lust, hilarity, garrulousness, and other reactions depending on the exact placing of the electrodes. During the same period, Robert Galbraith Heath led a program on electrical brain stimulation at Tulane University School of Medicine. His team implanted patients diagnosed with schizophrenia or epilepsy who had not benefited from the conventional therapies of that time (insulin coma, electroconvulsive therapy, psychotherapy; Heath 1954). Early in the 1990s, Alim Louis Benabid and coworkers in Grenoble combined implantable pacemaker technology with chronically implanted deep brain electrodes for the successful treatment of certain movement disorders (Benabid and others 1991). These experiments contributed important insights into the biology of neuropsychiatric disorders and initiated the development of new therapeutic approaches that have become broadly accepted. For instance, the US Food and Drug Administration (FDA) has approved devices for deep brain stimulation of different subcortical brain targets for the treatment of Parkinson’s disease (subthalamic nucleus or globus pallidus internus), dystonia (globus pallidus internus), essential tremor (ventral intermediate nucleus of the thalamus), and severe obsessive-compulsive disorder (anterior limb of the internal capsule). In addition, an active research field continues to explore deep brain stimulation for other indications, including epilepsy, major depression, Alzheimer’s disease, obesity, or substance use disorders, although current regulations remain unclear (Chandler and others 2021). However, the early research efforts of the 1950s also revealed the risks of this line of investigation.

Figure 1.

Timeline of the main past brain stimulation devices.

Decline of Early Research Programs

The field of brain manipulation with electricity eventually lost its momentum, despite robust scientific findings and intriguing clinical results. The beginning of chemical theories and drug development of mental illness in the 1950s certainly contributed to this decline. In addition, significant transgressions of broadly accepted social and ethical boundaries, and pursuit of fringe scientific questions, casted a negative light on this field as a whole. For example, Jose Manuel Rodriguez Delgado suggested using brain stimulation, especially the chemitrodes, to suppress violence that occurred in the context of demonstrations by African Americans in several American cities during the late 1960s. He ended up calling for a “psycho-civilized society” in which brain stimulation would be used to regulate and align individual character and behavior for the good of society (Delgado 1969). In 1972, Heath used brain stimulation to try to eliminate a male subject’s homosexuality, which was considered a psychiatric disorder by the DSM-II (Diagnostic and Statistical Manual of Mental Disorders, 2nd edition). Deep brain stimulation targeting his septal region was delivered while offering him the services of a female prostitute (Heath 1972; Moan and Heath 1972). Fred Mettler, a previous mentor of Heath, was among those who publicly castigated this work. Rodrigues Delgado’s ideas were also strongly condemned by the medical and scientific community (e.g., Mark and Ervin, 1970, Violence and the Brain). Nonetheless, approaches to behavioral control entered public imagination: consider, for example, the 1972 novel The Terminal Man from Michael Crichton and the 1996 novel The Kindling Effect from Peter Hernon.

Current Human Brain Manipulation Interventions

A modern field of brain manipulation using invasive and noninvasive stimulation has been advancing and broadening, spurred by advances in computation, microelectronics, molecular biology, and brain imaging technologies, and by a growing recognition of the limits of drugs for treating mental and neurological illnesses. Much of the work revolves around electromagnetic approaches, and yet other solutions to manipulate brain activity directly and in a controlled manner without requiring surgical exposure of the brain or implantation of hardware are rapidly being introduced, including some relying on light and other on ultrasound.

Invasive and Minimally Invasive Brain Stimulation Techniques

There are several invasive or minimally invasive approaches to modulate neural activity. Techniques such as deep brain stimulation and vagus nerve stimulation consist of implanting a stimulator that sends electric impulses in deep structures of the brain or to the vagus nerve, respectively (Perlmutter and Mink 2006). There is epidural cortical stimulation that also uses electricity to modulate neural activity using a device installed on the surface of the brain (Lefaucheur 2008). There are more precise methods available to control neural circuits, such as optogenetics, albeit still limited to animal models. Optogenetics exploits light activated cation and anion channels that can be expressed in specific neural cell types permitting selective activation or inhibition of neural circuits (Deisseroth 2010).

Noninvasive Brain Stimulation Techniques

There are also noninvasive brain stimulation techniques (Fig. 2). Transcranial magnetic stimulation (TMS) delivers electromagnetic pulses to induce electric current in the brain via electromagnetic induction by passing current pulses through a coil held over the subject’s head (Barker and others 1985; Hallett 2000). Transcranial stimulation with direct or alternating current (tDCS, tACS), or random noise (tRNS) deliver a low electric current through surface electrodes applied over the subject’s head (Nitsche and Paulus 2000). Cranial electrotherapy stimulation also uses electricity by sending electric pulses across the subject’s head (Bystritsky and others 2008). Another technique, the low-intensity focused ultrasound, uses sound waves to induce neuronal excitation and inhibition (Bystritsky and others 2011). Transcutaneous vagus nerve stimulation delivers electric impulses to the vagus nerve by applying the device to the surface of the ear (Carreno and Frazer 2016).

Figure 2.

Main current noninvasive brain stimulation devices: transcranial magnetic stimulation (TMS); transcranial electrical stimulation (with direct or alternating current, or random noise; tDCS, tACS, tRNS); low-intensity focused ultrasound stimulation (LIFUS); cranial electrotherapy stimulation; transcutaneous vagus nerve stimulation (tVNS).

Contribution of Noninvasive Stimulation Tools to Human Brain Investigations

The noninvasive brain stimulation tools, considered easy to apply and relatively inexpensive and safe, are particularly increasingly used by neuroscientists, cognitive scientists, medical doctors, as well as psychologists.

Contributions of Noninvasive Brain Stimulation to Our Understanding of the Core Principles of Brain-Behavior Interactions

Although we lack a clear understanding of their mechanisms of action, noninvasive brain stimulation techniques have proven to alter brain function, cognition, and behavior. This can lead to valuable insights into the core principles of brain function and brain-behavior interactions. Noninvasive brain manipulation methods, albeit with limited resolution, are being increasingly used to induce transient changes in brain activity, experimentally test psychological and cognitive theoretical models, test for cause and effect relationships between a behavior (e.g., motor or cognitive) and activity in specific brain regions or across distributed brain networks, and to manipulate brain activity in a controlled manner. Here are examples of the capacity of noninvasive brain stimulation methods to reveal causality relationships between behaviors and brain activity across specific brain networks enabling testing of several important theoretical principles of brain function:

1. Localization of function. TMS and repeated TMS (rTMS) were first used to map motor representations in patients and healthy individuals (e.g., Levy 1987; Wassermann and others 1991; Fig. 3) during different states such as wakefulness and sleep (e.g., Hess and others 1987). This method was also shown to be useful not only in the investigation of motor activity per se but also in the study of high-level motor planning (e.g., hand preference; Ammon and Gandevia, 1990). Since then, the FDA has cleared the Nexstim eXimia system for noninvasive presurgical hand motor function mapping evaluations. This line of work supports pioneered studies on motor homunculus by Wilder Penfield, Edwin Boldrey, and Theodore Rasmussen. It is still among the most fruitful scientific work in the field of noninvasive brain stimulation, leading to advances in basic research to clinical potential, even in children (e.g., Gilbert and others 2019).

2. Localization of dysfunction. rTMS can be used to disrupt cognitive functions, which contributes to localize brain regions associated with these functions (Fig. 4). When targeting Broca’s area with rTMS, adult volunteers had difficulties producing simple words (Pascual-Leone and others 1991). rTMS induced speech arrest and paraphasic errors in healthy subjects, similar to those encountered in patients with aphasia, by interfering with activity in the bihemispheric language network. This work supported the causal role of Broca’s area in speech production. The use of rTMS was also shown to contribute to localize critical speech areas in patients with medically intractable epilepsy (Michelucci and others 1994). Since then, the FDA has cleared the Nexstim eXimia system also for such noninvasive presurgical language mapping evaluations. Furthermore, rTMS is useful for topical localization of functions. Beckers and Hömberg (1991) found that rTMS over the occipital cortex disrupted visual perception and memory scanning, but not visual stimulus encoding or storage. Such approaches have been termed “virtual lesions” (Pascual-Leone and Walsh 2005) to emphasize that they shed light into the basis of neurological and psychiatric brain dysfunctions. This line of work illustrates how acquired brain insults impact functions as well as functional connections, first postulated in the 1900s by pioneers such as John Call Dalton and Constantin von Monakow.

3. Functional consequences. rTMS can test interhemispheric interactions and their role in shaping behavior. Hilgetag and others (2001) showed that disruption of the left or right intraparietal sulcus region with rTMS impaired visuospatial attention in the contralateral hemifield, while it improved it ipsilaterally due to release of interhemispheric inhibition. Knecht and others (2002) induced a virtual lesion of the left frontal operculum with rTMS in healthy volunteers to disrupt language processing. Even though the local impact was the same in all participants, those with a more lateralized language network, as characterized by functional magnetic resonance imaging, displayed greater functional consequences than those with a less lateralized network, that is recruiting more bilateral regions. Using functional magnetic resonance imaging–guided TMS targeting different brain regions, Sack and others (2005) showed that rTMS applied to the left parietal cortex disrupted generation of mental images but did not affect performance in a mental imagery task because the right parietal cortex was able to compensate. These approaches contribute to test hemispheric dominance and interhemispheric compensation models (e.g., Morrell and Salamy 1971; Smith and Burklund 1966).

4. Experience-dependent brain compensation. Noninvasive brain stimulation is used to test for brain reorganization that can occur in the context of diseases or disorders, as illustrated by a series of TMS studies on Braille reading (Kupers and others 2007). Most notably, the use of rTMS reveals that early “visual” cortex supports some fundamental aspects of tactile, auditory, and even language processing, in early blind individuals, and that such cross-modal plasticity changes are associated with functional gains. These rTMS experiments support previous work on experience-dependent brain compensation (Sadato and others 1996), such as the more occipital activation, the better auditory localization (Gougoux and others 2004) and verbal memory (Amedi and others 2007) in the blind. Hence, these and other applications of TMS have allowed us to understand how the brain reorganizes itself and compensates for various functional losses according to experience.

5. Modulation of cortical representation and output. TMS can be used to map changes in motor cortical representation associated with the acquisition of new skills and relate them to different cognitive processes (Muellbacher and others 2002). Depending on its stimulation parameters, rTMS can also modulate cortical activity and cortical output (Pascual-Leone and others 1994), inducing changes in motor cortical maps, and disrupting or promoting skill acquisition (Kobayashi and others 2004). Thus, it is possible to gain insights into neural processes of learning, and their relation to the system level plasticity.

6. Feedback cortico-cortical interaction. Focusing on visual awareness, and leveraging the fact that TMS can induce phosphenes (brief percepts of light) when targeting parts of the visual cortex, Pascual-Leone and Walsh (2001) used pairs of TMS stimuli applied with variable intervals to different parts of the human visual cortical system to demonstrate that feedback connections from human area MT+/V5 to V1 are critical for awareness of visual motion. More generally speaking this study illustrates the potential use of TMS to examine conscious awareness.

7. Modulation of distributed brain networks. The combination of TMS with brain imaging enables the analysis of the effects of a controlled perturbation of specific brain networks (e.g., Bestmann and others 2003). Importantly, this combination offers to the possibility of examining the relation between behaviors and specific neural network activities, as elegantly shown in the context of associative memory (Wang and others 2014) and working memory (Rose and others 2016).

8. Modulation of brain connectivity and oscillations. Brain function involves the dynamic coordination and configuration of different brain regions upon a relatively stable anatomic substrate. Therefore, understanding brain function also requires a comprehensive assessment of the temporal dynamics of neural activity. Oscillatory mechanisms play a key role in the dynamic coordination of neuronal activity in local and distributed brain regions (Varela and others 2001) and different brain regions have distinct capacities to support different oscillation patterns (Groppe and others 2013). Brain oscillatory patterns are altered in patients with neuropsychiatric diseases, such as epilepsy (Chowdhury and others 2014). Similar oscillatory alterations were documented in the unaffected first-degree relatives of these patients (Chowdhury and others 2014), suggesting that these disturbances in neural synchrony may represent an endophenotype. However, our knowledge of the relevance of brain connectivity and oscillatory patterns is still limited. Numerous studies have correlated connectivity and oscillatory patterns with cognitive functions or disease states, but their direct manipulation with noninvasive brain stimulation techniques allows assessment of their behavioral relevance. TMS (in which an applied electromagnetic flux leads to a spatially localized burst of neuronal activity) and tACS (characterized by the application of a low-amplitude electric field oscillating at a specific frequency) can be used to causally probe neural activity and neural interactions. Previous work shows that when combined with tools such as EEG (Fig. 5), noninvasive brain stimulation can be used to assess (1) the causal connectivity of brain regions in different behavioral and disease states (Massimini and others 2005), (2) the frequency response of different brain regions to brief stimulation (Rosanova and others 2009), (3) the compensatory frequency and connectivity changes that occur with more durable perturbations (Vernet and others 2013), and (4) the role of specific spatially localized oscillations in complex cognitive functions (Polanía and others 2014). Furthermore, these techniques can be employed to directly modulate the oscillatory rhythms that are involved in cognition, and that may be dysregulated in disease states.

Figure 3.

Contribution of noninvasive brain stimulation to our understanding of the core principles of brain-behavior interactions: localization of function. Illustration of the use of transcranial magnetic stimulation (TMS) for hand function mapping. Motor-evoked potentials (MEPs) are induced when the TMS coil delivers stimulation over the brain areas relevant to hand functions. Noninvasive brain stimulation consists of a new tool to support pioneered studies on motor homunculus by Wilder Penfield, Edwin Boldrey, and Theodore Rasmussen.

Figure 4.

Contribution of noninvasive brain stimulation to our understanding of the core principles of brain-behavior interactions: localization of dysfunction. Illustration of the use of repetitive transcranial magnetic stimulation (rTMS) for mimicking dysfunction such as speech arrest or paraphasia. Speech arrest is induced when the TMS coil is applied over the relevant brain areas. Noninvasive brain stimulation provides a new way to study the impact of acquired brain insults on functions and functional connections, sustaining pioneer work by John Call Dalton and Constantin von Monakow.

Figure 5.

Contribution of noninvasive brain stimulation to our understanding of the core principles of brain-behavior interactions: modulation of brain connectivity and oscillations. Illustration of noninvasive brain stimulation closed-loop system.

Contributions of Noninvasive Brain Stimulation to Our Understanding of Higher Order Human Cognitive Functions and Behaviors

In the context of studies designed to understand principles of brain function and brain-behavior relations, noninvasive brain stimulation has been found to be capable of affecting high order behaviors and cognitive processes. This can be valuable and lead to tests of model of cognitive and psychological processes. However, it can also illustrate the capacity of noninvasive brain stimulation to alter human mood and emotional state, decision making, social interaction, moral judgment, and deceptive skills.

1. Modulation of mood and emotion. Early work shows that rTMS applied over the prefrontal cortex capable of inducing a speech arrest also led to dysphoric reaction to the point of overwhelming sadness and crying (Pascual-Leone and others 1991). This is consistent with findings from direct cortical stimulation (Penfield 1959) and also from sodium amytal testing (WADA). Subsequent experiments confirm the possibility of decreasing happiness and increasing sadness in healthy individuals depending on the lateralization of stimulation and its specific parameters (George and others 1996; Pascual-Leone and others 1996). This set of studies has contributed to the elaboration of theoretical models on hemispheric contribution to mood (Sackeim and others 1982) and has encouraged translational application to patients with mood disorders, ultimately resulting in the FDA approval of rTMS devices (e.g., Neuronetics, Brainsway, Magstim, MagVenture) to treat major depression of patients who have failed to achieve satisfactory improvement from one prior antidepressant medication (O’Reardon and others 2007). At the same time, it has led to a fruitful line of research establishing the capacity of noninvasive brain stimulation to modify emotional perception in otherwise healthy individuals (Harmer and others 2001).

2. Modulation of decision making. Noninvasive brain stimulation has been used to study human’s ability of decision making in various contexts. It has been shown that healthy volunteers displayed an increased number of false alarms at the Go-Nogo task after they received tDCS over the dorsolateral prefrontal cortex (Beeli and others 2008), suggesting that tDCS may disrupt inhibitory control. Another example is that rTMS influenced healthy volunteers toward choosing more often an immediate but smaller reward than a delayed but greater one (Figner and others 2010) or the opposite (Cho and others 2010), indicating that rTMS may induce cognitive impulsivity or delayed gratification, depending on stimulation parameters. Moreover, noninvasive brain stimulation was reported to either increase risk taking (Knoch and others 2006a) or decrease risk taking and reward seeking (Fecteau and others 2007; Herz and others 2014), depending on the stimulation parameters. rTMS had also impaired computation of goal values of food stimuli but not estimation of food calories (Camus and others 2009). This line of work suggests that noninvasive brain stimulation may be particularly valuable to study the neurobiology of substance use disorders. Indeed, impairments in these processes (self-control, cognitive impulsivity, risk taking, reward seeking) have been reported linked to increased vulnerability craving and substance use in substance use disorders (e.g., Rogers and others 1999). Some proof-of-concept studies suggest that noninvasive brain stimulation applied over the dorsolateral prefrontal cortex may modulate these processes, as well as suppress craving and substance intake (Fecteau and others 2010).

3. Modulation of social interaction. Noninvasive brain stimulation can also affect behaviors when interacting with a human peer. It has been shown that rTMS can influence the acceptance rate of unfair offers of money when they came from a human peer but not from a computer (Knoch and others 2006b), suggesting that rTMS can influence a social aspect involved in decision making. Also, when volunteers are given money and asked to give some money back to their peers, they usually give smaller amount when they can transfer the money anonymously as compared to when their identity is known. It has been reported that rTMS can interfere with these processes: healthy volunteers gave back similar amount of money regardless of whether they were anonymous or not after they received rTMS (Knoch and others 2009). Knoch and others suggest that rTMS decreases self-control, impairing subjects’ ability to override immediate short-run benefits. There are also some data indicating that tDCS can augment or reduce sanction-induced norm compliance and voluntary norm compliance, especially during social interactions (versus interactions with a computer; Ruff and others 2013; see also Zhang and others 2016). It has been discussed whether these stimulation protocols on social norm compliance might be a valuable therapeutic avenue for personality disorders (Ruff 2018).

4. Modulation of moral judgment. rTMS can also influence judgment of attempted harms (i.e., negative beliefs) as being less morally forbidden and more morally permissible (Young and others 2010) and can increase probability of utilitarian responses, especially in high-conflict dilemmas (Tassy and others 2012). As an example, healthy volunteers were more likely to choose to kill their own baby to spare the lives of many people after they received rTMS than sparing their baby and put at risk the lives of others.

5. Modulation of deceptive skills. It has been well described that healthy subjects usually display longer response time when lying than telling the truth (Spence and others 2004). This response time associated with lying has been shortened with tDCS in various contexts (e.g., about personal or fictional information; Karim and others 2009). Other aspects of deception have also been influenced by rTMS, such as increasing propensity to lie (Karton and others 2014) and recognition of deceptive intents (Tidoni and others 2013).

While most scientific investigations aim at understanding brain functions and developing novel therapeutic avenues, they also carry ethical, social, and policy implications; and work from the past era has declined partially due to ethical and social reasons.

A Revisited but New Era of Human Brain Manipulation

The discussion above demonstrates that noninvasive brain stimulation technologies have the capacity to influence people’s brain function without the need for surgery, drugs, or even long-term wearing of a device. These technologies can influence various products of neural circuit activity, such as perception, mood, decision making, social cognition and emotion, and significant aspects of behavior. History shows that the idea of modulating behaviors with brain stimulation is not new. The current era, however, raises the possibility that not only neuroscientists but also a broader scientific community and the general public have now access to manipulate these brain stimulation devices. Indeed, a major new aspect to consider nowadays is the availability of these devices.

The Use of Noninvasive Brain Manipulation by the General Public

Thus, while solidifying the appeal of noninvasive brain stimulation as a tool to understand human cognition and its neural substrates, such studies also raise the concern of potential abuse by scientists but also by the general public. Indeed, the possibilities of manipulating the brain noninvasively that may enhance human capacities have captured the interest of the general public. Furthermore, the direct-to-consumer business and online shopping websites have empowered the general public to facilely buy brain stimulators. There is also the do-it-yourself movement on how to build brain stimulators. For instance, tDCS devices are inexpensive or relatively easy to build.

Some noninvasive brain stimulation protocols are FDA cleared (e.g., rTMS to treat medication-refractory major depression, obsessive-compulsive disorders, migraine, or tobacco smoking cessation) and others used off-label harbor the promise of treating human disorders (e.g., attention deficit, autism spectrum disorders). There is also great effort to define and recommend standardized guidelines on training practitioners who administer stimulation (Fried and others 2021). In most countries (e.g., Canada, the United States), individuals need a medical prescription to receive these treatments. However, some companies practice direct-to-consumer sales by marketing these same brain stimulators used in hospital and research settings as “general wellness products.” These companies advertise for cognitive or motor enhancement, but also for alleviation of cravings, stress, negative emotions, and social connection (in other words, symptoms of disorders). This leads to the question (Fig. 6): Should noninvasive brain stimulation be considered as a prescribed medicine and an over-the-counter medicine? It is currently both. This is unlikely with medication: Can someone buy antidepressants for his/her well-being directly from a pharma? Should some specific noninvasive brain stimulation protocols be considered as prescribed medicine and others as over-the-counter medicine?

Figure 6.

The current use of noninvasive brain manipulation by scientists, medical professionals, and general public. Some noninvasive brain stimulation devices (regardless of the stimulation protocols) are used for investigations, medical treatments, and wellness. The wellness use* advertised by some companies includes cognitive and motor enhancement, as well as alleviation of craving, stress, emotions, social connection, and so on. The devices can be licensed medical ones or direct-to-consumer and some, especially tDCS, as do-it-yourself devices. Some devices are cleared by health agencies (US FDA, Health Canada, etc.), such as rTMS for treatment-resistant depression, and they can also be used off label, such as for autism spectrum disorders. These uses bring into question if these devices (or some devices and/or some specific protocols for given treatments) should be prescribed by medical professional or considered as an over-the-counter treatment. This seems particularly important for some populations (e.g., children).

New Challenges of Modern Noninvasive Brain Manipulation

This revisited but new era of human brain manipulation involves a community of users that is much larger than in the past, potentially putting its misuse at greater risk. In an era of self-diagnosis and self-treatment fueled by internet dissemination, some individuals attempt self-treatment of their medical conditions with noninvasive brain stimulation. Others seek to enhance their musical performance, video gaming, physical capacities, or overall well-being. The direct-to-consumer and do-it-yourself movements are likely here to stay and continue to grow. Currently, there are no clear international guidelines from the scientific and medical communities and no firm recommendations from public policies, but effort and progress are made (e.g., since May 2021, transcranial magnetic and electrical stimulation devices, regardless of their intended use, are considered as Medical Devices in the EU). True, this is not a simple task. As a scientific community expert in noninvasive brain stimulation, what can we do now? (1) Continue spearheading effort on developing safety guidelines; (2) contribute at revisiting and monitoring public policies; (3) engage all communities (physicians, ethicists, general public, patient advocacy groups, governmental regulatory bodies, etc.) in a constructive dialogue; and (4) educate the public.

The safety issue regarding noninvasive brain stimulation has been under constant surveillance by the scientific community (e.g., Bridgers and Delaney 1989; Lerner and others 2019). Yet, the potential long-term effects of this approach on health and safety are still poorly known. Under most current experimental protocols, effects from a single session of brain stimulation appear to be transient (e.g., N-acetylaspartame levels were transiently elevated within 15 minutes of 1 mA tDCS delivery, an undesirable effect in healthy adults; Hone-Blanchet and others 2016). However, we have yet to understand whether short-lived experience, such as perception of improved problem-solving skills or disrupted speech or moral judgment, may provoke positive or detrimental lasting behavioral and cognitive consequences.

Importantly, most people building or buying these devices for self-use likely deliver more than a single session of noninvasive brain stimulation. However, there are currently no available longitudinal safety studies of repeated stimulation sessions in healthy volunteers (adults, children, adolescents). Daily sessions delivered in therapeutic studies show that longer-term changes in circuit function can be attained (e.g., to treat medication refractory major depression; O’Reardon and others 2007). However, long-term effects (beneficial and/or detrimental) are not even well known in these patients with medication refractory major depression who have been receiving the FDA-cleared rTMS protocol for decades. Studies are needed to assess the potential consequences of cumulative doses over years of exposure. Furthermore, specific paradigms of brain stimulation, from single or repeated sessions, might improve or disrupt cognitive functions, but might do so differently across individuals depending on age, health status, and many other factors. Effects might even differ over time within a single individual depending on onset of illness, injury, medications taken, or even state of arousal.

The safety issue also comes from a lack of clear mechanistic knowledge of human brain manipulation (e.g., Rounis and Huang 2020; Bestmann and Walsh 2017). One notion of unintended consequences is due to the fact that even focal noninvasive brain stimulation can induce distal effects transsynaptically, as well as potentially through homeostatic control and/or endogenous neuroplastic response mechanisms that remain poorly characterized. One of the earliest examples comes from Bridgers and Delaney (1989), who reported that rTMS induced a decline in serum prolactin, a finding replicated by Wassermann and others (1996), suggesting that some effects are not captured by common tests (e.g., behaviors, cognition, brain activity) and likely engage de- and/or activation in distal brain regions (e.g., potentially the hypothalamus in regard to prolactin). In addition, animal work reveals that rTMS can engage epigenetic mechanisms similarly as psychoactive drugs (e.g., Etievant and others 2015). There is also increasing recognition of state- and trait-dependent factors that can shape inter- and intraindividual variability in the response to brain stimulation (e.g., Goldsworthy and others 2021). Furthermore, reproducibility is increasingly being recognized as a critical concern for noninvasive brain stimulation studies overall and this may be especially problematic for manipulations of cognitive functions and behaviors without an independent physiologic measure of the effect. This has also been an issue in the past (Fig. 7). These issues should incite some caution and considerations from advancing the applications of such techniques to widely before more is known.

Figure 7.

The importance of deepen mechanisms of noninvasive brain stimulation to induce behavioral changes that are replicable and further improve safety guidelines.

The scientific community expert in noninvasive brain stimulation, along with institutional review boards, carry shared responsibility to conduct safety studies as well as systematically assess and report side effects in every type of study, but again, the recent and increasing use outside the scientific and clinical settings complicates safety monitoring. In addition to safety recommendations, there is an urgent need for regulation, open debate, and awareness of potential abuse, which is particularly critical for children and other vulnerable populations. Specific questions should be carefully addressed, including: (1) Should there be regulated applications, such as delivering repeated stimulation sessions in healthy adults or in children? (2) Is the use of such technologies and their further development as consumer tools desirable or should they be restricted as medical interventions requiring a prescription? (3) What are the standards and built-in safety features for medically prescribed versus over-the-counter or do-it-yourself devices?

Conclusions and Perspective: How to Care for Modern Noninvasive Brain Manipulation?

The use of brain manipulation outside the scientific community was not a real concern in the past, but we are now facing important issues similar to the concerns about neurotechnology that were raised following the work of Delgado, Ervin, Mark, Heath, and many others in the 1960s and earlier. We should do better, because we have (1) the possibility of learning from history, (2) the benefit of better defined ethical guidelines and research regulations, and (3) the capability of faster and wider communication within the scientific community enabling us to rapidly report abuses as well as actual data about the effects of brain manipulation. However, we also face more health and ethical risks than our predecessors because of easy access to noninvasive technology and increasing political, military, media, and general public interest. The experiments on modulating high order behaviors are surely sensational topics that arouse media interest. However, this also elicits concerns of risk of abuse especially with the general public, enabled by internet dissemination and do-it-yourself and direct-to-consumer devices. So are we better prepared for manipulating the human brain or are we actually at higher risk of abuse and deleterious impact on the healthy human brain given ready access to technology? It is timely to open the debate and lay down the potential future impact of these tools, in order to avoid having to confront and explain past actions. The open debate should bring together ethicists, philosophers, clinicians, engineers, scientists, and so on. This is critical, as it is a distributed responsibility. There are already some welcome initiatives addressing these questions, including meetings at the National Academies of Sciences, Engineering, and Medicine (2017) and papers (e.g., Luber and others 2009; Wurzman and others 2016), but policies are sill lacking. There is thus a call for recommendations and regulations—research under an international review board that defines how such work should be monitored.