Navigation – Plan du site

AccueilNuméros76ArticlesExcess in the avant-garde of the ...

Articles

Excess in the avant-garde of the data subject

Language and quantification in digital psychiatry
Danielle CARR
Cet article est une traduction de :
Le Sujet Numérique d’avant-garde et sa part irréductible [fr]

Résumé

Although quantification has marked psychiatric practice since the rise of experimental psychology, the past two decades have seen an intensified use of data, implemented in tandem with a general move toward big data and “scored societies” across the biosciences. Today, a new, hyper-quantified psychiatry is aiming to map the neurological substrates of affective disorder through extracting neural data from brain implants. The classic story that is often rendered by many critical scholarly accounts about this new form of hyper-quantified psychiatry is that in their drive to render psychiatry profitable to new forms of data capitalism, such technologies evaporate language and the subjective into a cascade of numbers. This paper examines a case study of a subject enrolled in an experimentation in psychiatry for the first brain implant system coupled to an artificial intelligence in order to trouble such easy narratives about quantification. How would our account of quantification change if we saw language not as an extraneous surplus to the experimental production of numbers, but at the heart of how these new technologies of quantification are being developed?

Haut de page

Texte intégral

1This paper is based on two years of living and working with a laboratory in the United States, where researchers are developing permanent brain implants in humans to treat psychiatric mood disorders. This experimental treatment is known as deep brain stimulation (DBS), a pace-maker-like device that implants electrodes directly into “malfunctioning” portions of the brain in order to collect data about neural activity and distribute electrical stimulation (Fins & Schiff 2010). DBS is currently the most physically invasive psychiatric treatment, offering the promise of synoptic, real-time data about the brain’s electro-chemical activity. The rise of DBS sits at the juncture of two developing tendencies within contemporary capitalism: (1) new techniques to perfect predictive control over the world through data; and (2) the incorporation of the body into these regimes of data as a site of value production.

2DBS is the cutting edge of what is often described as a new “digital psychiatry,” a paradigm that rejects older psychoanalytic therapeutic models of psychiatric healing in favor of a hard-biological metaphysics of the subject. While psychoanalysis prioritizes language as a medium of exchange between subjects transacting games of recognition, digital psychiatry considers the subject to be the outcome of objective bodily processes that can be measured and analyzed through vast computational power. A common theme found in critiques of digital psychiatry is the idea that it has successfully eradicated language from the scene of psychiatric treatment, replacing the psychoanalytic subject with the “data subject.” In these accounts, qualities like subjectivity, language, and meaning that typified older forms of psychiatric treatment like psychoanalysis are disappearing into a morass of numbers bent on rendering new sites of value open to capitalist exploitation. What I will demonstrate in this paper is that these kinds of narratives better describe the aspirations of capitalism than they do the process of quantification as it actually happens. The problem with these lamentations over the triumph of quantification is that they are too credulous about the idea that every quality can be turned into a quantity, too quick to agree that, while it is something to be strenuously resisted, the quantification of the subject has, indeed – unfortunately – been successful.

3This paper tells a different story. I argue that although the drive to quantify psychiatry is driven by a push to make psychiatry profitable within emerging forms of data capitalism, the production of quantities is confounded by the very logics of language that digital psychiatry insists that it has replaced. It is only by understanding this point, rather than acceding to the idea that the numbers have, indeed, replaced language, that we can begin to formulate an adequate critique. In what follows, I will demonstrate that, for all of the drive to quantify psychiatry in a way that will facilitate profit-making, language permeates the creation of the “datafied self,” warping digital psychiatry’s regime of numbers with the very thing it was supposed to replace.

The Data Factory

Hope is a car that drives through the slums of depression
Then Hope emerges when the serotonin wins the election
Dopamine kicks in and I have good intentions
The way the brain works is quite impressive
Created by the one who is beyond perception
and I follow His messenger who is granted intercession
My serotonin got to the end of the neuron and my brain said Hold this shit
So now I start my conversations with “I’m so depressed.”

4The experiment is conducted in a satellite campus of the Massive Corporate Hospital System (MCHS). A structure build by the United States Navy, the cavernous concrete building has been converted into a brain science Data Factory. In and out of the building scurry droves of young research assistants (RAs), freshly graduated from bachelor’s programs in cognitive neuroscience or computer science, and eager to accept salaries that come out to less than minimum wage for the chance of making it into graduate school. They are the shop-floor workers of the Data Factory. At first glance, you could easily mistake the experimental Subject for being one of them. Like the RAs, the Subject is 25 years old, and reports weekly to one of the labs in the Navy building; like the RAs, he will see none of the massive profits made from the experiments that absorb his life. But unlike the RAs, the Subject is the first person in history to have been implanted with a neuromodulation device that stimulates and records brain activity with a “closed-loop” system. The fact that the experiment is being run on only one subject speaks to the extreme difficulty of recruiting patients for this type of research: patients must be willing to have an experimental device implanted, while passing a byzantine array of inclusion criteria to ascertain that this experimental treatment is indeed being used as a last resort. While his rationale for participating is that the device may treat his mental illness, the Subject is the centerpiece of an intricate network of public, private, and military funding, an n of 1 in a series of proliferating experiments bound together less by a hypothesis than by the unparalleled opportunity for harvesting data that he embodies. The Subject’s body has become an organ of the Data Factory. Signals from his brain are captured by the neural implant; signals from his movements, social interactions, and mood are gathered by his smartphone; all are collected in terabytes of information, signals the RAs will fashion into data. The “closed-loop” system allows the implant to intervene in neural circuitry by distributing electrical stimulation in real-time stimulation through the electrode’s contacts lodged deep in the brain, a change from previous DBS systems, which only allowed stimulation at a pre-set parameter. At the same time, the electrodes are recording data about neuro-electrical activity. As both research and therapy, the data harvested by DBS systems offer a path toward one of digital psychiatry’s goals: to replace the linguistically constructed nosology of the previous Diagnostic and Statistical Manual with categories drawn directly from the body.

5The basic conditions and structure of the experiment – or, to put it differently, of the Subject’s life – are as follows. The Subject has suffered intensely from mental illness since early childhood. The exact nature of his illness remains unclear: over the course of his 14 years of psychiatric treatment, he has been diagnosed with obsessive compulsive disorder (OCD), major depressive disorder (MDD), and bipolar I disorder. He lives with his family, and is dependent on his father, who has for years relentlessly sought the cure for what ails his son. His course of treatment, which has largely been to no avail, has included at least twelve drugs, repetitive transcranial magnetic stimulation (rTMS), electroconvulsive therapy (ECT), and ongoing psychotherapy.

6These treatments have done little to nothing for the Subject’s suffering, which manifests as repetitive thoughts, anxiety, and chronic feelings of emptiness and sadness. Desperate for a cure after all standard treatments have failed, the Subject was recommended to the lab for experimental treatment in a clinical trial using a neural implant called deep brain stimulation. DBS has benefited from enormous state and private funding operating in conjunction. President Obama’s 2013 announcement of the Brain Initiative modeled on the Human Genome Project unveiled a bolus of federal funds to map the brain, a project that relies on DBS as a central technology (Markoff & Gorman 2013). Federal support for the project had reached $434 million by 2017, and half of the funds have been supplied through the Defense Advanced Research Projects Agency (DARPA), the experimental science branch of the military. The 14-fold increase of intellectual property claims for brain stimulation technology since 1995 (Roskams-Edris et al. 2017) attests to the potential value of the data being extracted from the brain. At the time of fieldwork in 2019, the global brain stimulation market was projected to be worth $6.2 billion by 2021, with pharma giants like GlaxoSmithKline vying with venture capitalists like Mark Zuckerberg and Elon Musk to establish neurotechnology labs (Masunaga 2017). DBS lacks Food and Drug Administration (FDA) approval for mood disorder, and the results on whether it “works” have been equivocal: several studies with small cohorts have shown promise, but the only two large studies with more participants were shut down when they failed to achieve efficacy benchmarks. The Subject was implanted with a DBS system as part of one of the two larger trials that were shut down, and he was one of the test subjects who didn’t experience any significant benefit from the device. But the lab recently got a crop of new grants from a conjunction of military, public, and medical device funders, and they needed an experimental subject. At the urging of his father, the Subject enrolled in the new study, and was implanted with the new, first-ever-in-human closed-loop device.

7In some ways – in fact, in most ways – the Subject is not the ideal candidate for this study. It is not clear precisely what mental illness he has. Are his repetitive thoughts about his depression a symptom of the depression itself, or an atypical form of OCD in which his rituals are not physical rituals like handwashing, but mental repetitions? The fact that his diagnosis is unclear means that even if the DBS causes him to “get better,” it won’t be clear what pathophysiology the treatment has “cured.” But the name of his diagnosis is not of paramount interest to the principal investigators (PIs) running the study, who have dismissed the nosology provided by the Diagnostic and Statistics Manual as being unscientific and imprecise. Instead, they are interested in mapping the correlation between neural activity and specific behaviors. In any case, even if the study doesn’t cure the Subject, the experiment will still have been valuable: as one RA explained to me over lunch, “To the PIs, in a way this experiment is a test of the experimental protocol itself.” What they meant by this was that while most of the lab’s research assistants had reservations about the scientific validity of the trial – not to mention the likelihood that the Subject would get better from the experimental treatment – they understood that the senior scientists running the trial had multiple objectives in play, only one of which was the therapeutic outcome of the experiment. The PIs needed to provide an experimental subject to keep from losing the grants that were supporting the project, and they needed to test that this sort of experimental design would actually work. These were merely the facts of the matter, neither good nor bad. The other fact of the matter was that I never doubted that everyone involved sincerely wanted the Subject to get better, and that the PIs genuinely believed that the treatment had a good chance of relieving his suffering. These things can both be true at the same time.

8Over the two years I spend in the lab, the experimental protocol expands to collect more and new types of data. Still, for all its fluctuations, the basic structure of the experiment is as follows. The Subject has an implanted electrode system that stimulates two parts of his brain: one deep target and one on the surface of the cortex. Simultaneously, these electrodes are also recording brain activity, conveyed to a simple compact computer called an IPG (implantable pulse generator) that has been implanted in the shoulder, which can store about two weeks’ worth of data. In tandem with this harvesting of brain activity data, the Subject is equipped with an app on his phone that tracks his movements. This app is so precise in its collection of granular movement data that researchers can infer whether he has gotten out of bed, is upstairs or downstairs, is exercising or running errands. Every week, the Subject comes to the lab for either a “data dump” or a clinical visit. During these visits, he is filmed so that the footage can be fed into an AI algorithm to extract correlations between his facial expressions, the sound of his voice, and the other data being collected.

9These visits are often his most significant, or only, social interaction of the week. Many times, when researchers at the lab ask him how he’s been since the last time he came in, he will respond that for the few days after his last visit he “felt like a human being.”

10The goal in each arm of the experiment is to produce entities known as “biomarkers,” or objective and measurable physiological states that correlate consistently with psychological behaviors or states; the idea is that with sufficient data, psychological states can be inferred from neurological or behavioral data. The promise of a biomarker is to circumvent the patient’s self-report and get direct access to what a psychiatric illness “really is.” Biomarkers provide a truth the body can speak directly without having to be refracted through the subject, its language, or its tendency to exceed the quantifications that will facilitate digital psychiatry’s profit-making aspirations. This truth beyond language will tell us what an illness really is. Evicted from the prisonhouse of language, the subject will become a body, and the body will be forced to speak the true name of its illness, encoding it into numbers. At last, the biomarker says, we will arrive at the Thing Itself: the illness in its pure form, reality itself decoded.

The avant-garde of the data subject and “surveillance” biocapitalism

11In many ways, the Subject at the center of the experiments is at the cutting edge of the creation of a “data subject,” a dynamic that many scholars and social theorists have articulated to describe the developing relationship between data, the body, and emerging forms of capitalism. While there are many differences in these individual accounts, the common threads add up to a story that goes something like this. As the development of capitalism requires ever-new territories and sites where it can extract value, a subject that appears to be made out of data is being created (see, e.g., Raley 2013; Goriunova 2019). The rise of the “information economy” is the product of a reconfiguration of production and labor that meshed with the ascent of digital computation technologies in the 1970s. With the rise of computerized capabilities of amassing large quantities of data, by the early 2000s many critics were arguing that the global digital data economy was operating through control over modes of communication (Thrift 2005; Lash 2007; Castells 2010; Kitchin 2014). Within “informational capitalism,” the quantification of everything into data enables capital to turn its eye to the body in new ways. While prior modes of capitalist production relied on commodifying a worker’s bodily labor, the “new economy” now equally relies on profiting from information gathered on individuals’ behavior, consumption, and habits (see, e.g., Ritzer 2014; Thrift 2005).

12The rise of the information or digital economy facilitates the creation of what many theorists have called “biocapital.” Per this formulation, contemporary capitalism is being reorganized around the reproductive capacities of life itself at the “suborganismic level of the body” (Cooper & Waldby 2014: 11). As anthropologist Kaushik Sunder Rajan has noted, capitalism’s turn to life produces new forms of and capacities for life itself (Sunder Rajan 2006: 7-15, 138-144), which hinge on technical and bureaucratic processes that “alienate information from material biological source” (ibid.: 17). Most presciently for the present case study, biocapital’s drive to render life abstract and quantified is exemplified in the welter of bureaucratic and technological apparatuses – brain stimulation, weekly visits, ratings scales, location tracking through apps, voice and facial analysis, and so on – that turn the Subject’s life and body into numbers to be mined.

13This brief sketch of the entwined relationship between the “data economy” and biocapitalism is necessary to understand how digital psychiatry is situated at the nexus of the two dynamics. Digital psychiatry relies on the creation of a “data double” of the subject (Pickersgill 2019; Stark 2020). It is this data aggregate, rather than the speaking subject, that is of interest to the clinician or researcher. These digital signals will allow “digital phenotyping,” or the creation of markers presumed to correspond to a real biological state of things (Birk & Samuel 2020). The reorganization of psychiatry away from language is not only an epistemic shift, but is also linked to psychiatry’s increasing subordination to profit-making incentives. To the same extent that clinicians are supposed to “listen as machines” (Semel 2019), digital psychiatry seeks to analyze “objective measurements” in such a way that the labor of a clinician or researcher can be replaced by a machine (Rosenberg 2002). Andrew Lakoff has theorized the forms of diagnosis that result as displaying “diagnostic liquidity,” or an ability to toggle fungibly between the clinic, the lab, and the market (Lakoff 2005). The forms of data collection undertaken in the experiment on the Subject aim to produce the correlation between behavioral data and neural data that undergirds the explosion in profit-making enabled by digital psychiatry. With the expensive labor of clinicians replaced by AI, venture capital is currently pouring money into digital applications in the mental health space, investing $637 million in digital mental health startups in 2019 alone (Runwal 2020); this number has only increased with the enormous opportunity for digital mental wellness apps presented by the COVID-19 pandemic (Cosgrove et al. 2020).

14The DBS experiment sits at the intersection of these threads. DBS is a valuable research tool for its two-pronged capacity to both stimulate and record electrical activity in the brain. Eletrodes placed at neural locations hypothesized to be responsible for psychiatric treatment distribute electrical stimulation, intervening in real time into neural circuitry and allowing researchers to test hypotheses about the relation between neural circuits and behavior. At the same time, these electrodes are recording neural activity and storing the data, which can then be downloaded and correlated to a variety of measures such as mood ratings scales produced during clinical visits, and the data collected by the Subject’s smartwatch and smartphone. Several months after I join the lab, I trek to the office of the PI in charge of the arm of the experiment concerned with gathering phone data to correlate with the readings of the electronic neural data. He explains why the experiment has recently become an object of investment and collaboration with Gargle, a global data conglomerate whose parent company has launched a life sciences venture. This PI has his fingers in this pie, as well as many others, and runs a tech-based initiative eager to collaborate with the plethora of startups looking to use digital psychiatry to turn a profit. He explains why it is that the Subject is of such interest to the new collaborations between digital psychiatry and data capitalism:

The ability to stimulate and record from neurons in the live human brain is something that is relatively new, at least in an ambulatory setting. People are in their real world. And so for us, the opportunity to capture naturalistic neural data, where we can also be looking at naturalistic behavioral data in tandem, both at high temporal density, was really a key opportunity for us. So for our role in the study, we’re just opportunistically latching on here, because we’re trying to understand ways of capturing objective behaviors that map onto these subjective constructs. … For us, success looks like the ability to accurately estimate somebody’s subjective state from his other objective data types.

Our approach says, “You know, let's sideline the question of language from the perspective of meaningful categories – diagnoses or subjective states – and say, okay, how did these datasets vary and like how can we make them talk to each other in a way that doesn’t necessarily have to do with meaning in that sentence?”

15A few months after our conversation, I will accompany the Subject to the Gargle headquarters, where he will be the star of the company’s Grand Rounds, the living showpiece of the new digital psychiatry.

Language all the way down

Use your intellect to understand my metaphors
Put much into writing this for people to just ignore
I’m articulating the pain in my brain that won’t go away
It comes in waves
happens every day
This is my status
people won’t know what to say
I want more social interaction because making people laugh is my passion
I’m still the authentic me just looking for a distraction to help give me spiritual satisfaction
I’m always me 

16If the aspiration of digital psychiatry is to excise the difficult problem of language – its tendency to slip into metaphors that exceed quantification – the attempt to banish it isn’t going so well. It’s a weekday morning just before lunch, and we are all huddled into the tiny office of Tony, the lab’s chief postdoc. The Subject has taken a shine to this group: the three RAs and Tony, who run data collection on him. “The Lab is like my family,” he tells me multiple times over the years. He has been growing a mustache that echoes the distinctive one Tony sports. Every week, the Subject arrives earlier for his data dump, and finds increasingly elaborate reasons to stay past the end of the visit.

17Today is the first day that the lab is running a protocol with the face and vocal recordings added, courtesy of the tech initiative PI’s lab. The room is extra crowded because of the additional person in the room, an RA sent from the consulting tech PI’s lab to set up the apparatus. There are wires looping around the Subject everywhere, connecting his IPG to the tablet that will collect the data, the video camera trained on his face to the laptop, the microphone sets collecting everyone’s voice data. While everyone sets up the apparatus, Tony chats with the Subject, as he always does.

18“Have you been writing more?” Tony is referring to the poems the Subject writes and sends to the RAs on his experiment, plus Tony and now me. We get them via texts and emails, several times per week, and often the Subject reads us his latest during the lab appointment itself. The Subject says, yes, he is always writing, and asks if Tony got the poem he sent last night. Tony say he didn’t, but as if to make up for it, asks, “Do you feel some relief writing them?” He means relief from the Subject’s symptoms – the racing thoughts, obsessions, and black mood. The Subject doesn’t exactly say yes, but explains – as much to the new RA as to Tony – “I have to use metaphors to get people to understand what I’m going through.” He pulls out his phone unprompted and reads his poem:

Outside I’m a lucid human being
Praying on my knees
Inside I can’t die
I’ve been on a treadmill tied to a leash
I ran so fast
That I collapsed
friction took the skin off my back
Agonizing pain
Forgot my name
Trying to scream but I can’t breathe
I’m dying inside but fuck ECT
Waiting for the cortical leads
To set me free
It’s been like this for the last 10 years
people talk to me I know they care
But my brain doesn’t register their words
I just nod my head and steer
Lay in the hospital with guilt and fear
Nurse checking on me
Makes sure I’m still there
Got home thankful for my own bed
Met up with my two best friends
anguish and dread
Anticipating dinner
But the Luvox is too bitter
I had to trust my gut all along
I knew I wasn’t going to get better

19The RAs are half-listening, half-completing the elaborate experimental set-up and initiating the data dump. Tony murmurs to one of the Ras, “There’s limited memory space. We’ll have to download what we can then clean the data.” The Subject finishes reading his poem. As the senior-ranking lab member in the room, there is a feeling that it is on Tony to respond. His attitude is encouraging, but I write in my notes that he seems out of his depth. “Well,” he says awkwardly, “keep writing.” “I’ll send you the one I wrote today,” says the Subject. He wants to read it to us, but it’s time to begin the battery of ratings scales that are used to keep track of his symptoms over the course of the study.

20The ratings scales are only one of multiple levels of the experiment’s attempt to transform language into quantities. The first rating scale lists a series of emotions (exhausted, relaxed, angry, and so on) and asks the Subject to numerically rate how much he is at that moment experiencing it, from one to five. Over the course of the experiment, I never see him endorse an emotion that would indicate that he is having an anti-social feeling like boredom, rage, or annoyance. He is too eager to gain the approval of the lab members; endorsing an emotion like anger or annoyance would mean that he is not among his friends. (He acknowledges this one day in a joke: “Imagine if I said I was furious! It would be awkward.”) The second rating scale is administered by the new RA. She begins by asking the Subject a series of questions related to his daily functioning and cognitive habits, questions intended to be answered with a yes or a no. Quickly, the situation begins to get away from her. She asks him to define his diagnosis. He says quickly that he has OCD and depression, “but not like what you might have already seen.” He begins describing his daily experience of living, his words tumbling out in his eagerness to explain before she cuts him off. “It’s like I feel like I’m being buried in my own grave. It’s all day, every day anguish, like I lost the World Cup.”

21She writes down the string of metaphors, bullet points in neat handwriting written in the blank margin of the pre-printed questionnaire. Only the questionnaire results will be entered into a database that will be used to assess the efficacy and results of the neural stimulation; these numeric quantities will be correlated to the data being collected by both the neural implant system and the Subject’s smartwatch and mobile phone. Everything else the Subject has said is treated as so much surplus, discarded because it cannot be easily quantified and correlated. The last test asks him to interpret the meaning of everyday metaphors or idioms: “There’s more than one way to skin a cat,” or “A bird in the hand is worth two in the bush.” He gets most of them right, but eventually he gets stuck. “That could mean a lot of different things,” he says. “It’s okay,” she reassures him. On the floor, a series of beeps is emitted from the tablet to indicate that the memory storage is full. The tablet’s case says “Medtronic: Your smarter DBS journey begins here.”

The gift of the symptom

22The aspiration to render the subject as matrices of quantified correlations hinges on establishing dyadic relationships between quantities and qualities. Digital psychiatry’s search for biomarkers is predicated on the idea that X subjective state will collapse into Y measurable quantity, with one folding neatly into the other without excess. Michel Callon refers to this process as “enframing,” or the practices of rendering objects “disentangled” from their environment and thus amenable to calculation (Callon 1998). The modes of enframing at the heart of capitalist accumulation rely on getting rid of the excess that does not collapse into the regime of numeric equivalence. Crucially, this maneuver is an ideological one: the regime of numbers that enables such accumulation only works by insisting that it has successfully gotten rid of the excess that doesn’t fit into its matrices of calculability.

23One example of this “enframing” of dyadic equivalence at the heart of capitalist regimes of fungibility is provided by Marcel Mauss in his The Gift. In a market exchange like the wage, two parties trade a good for the quantified representation of that good’s value (in this case, the monetary representation of labor’s value). These entities – a unit of labor time and a quantity of monetary value – are held to be equivalents that collapse into each other without remainder. In contrast to this structure of market exchange, Mauss shows that the gift is animated by the logic of a remainder, a third space that exceeds the dyad of X exchanged for Y, and it is this nonequivalence that animates the social relation (Mauss 2000). Yet here is the point that cannot be missed: Mauss explicitly understands the dyadic logic of market exchange to be an ideological veil cast over a gift relation that remains, at its core, fundamentally social. In the closing pages of The Gift, Mauss rebukes the capitalist employer who schemes to obscure the fundamentally social relationship through the veil of the wage relationship. Of the hidden social relation occluded by dyadic quantification of the wage, he writes: “The producer who carries on exchange feels once more – he has always felt it, but this time he does so acutely – that he is exchanging more than a product of hours of working time, but that he is giving something of himself – his time, his life. Thus he wishes to be rewarded, even if only moderately, for this gift” (Mauss 2000: 77, emphasis added). What Mauss means by this is that capitalist accumulation rests on installing a regime of equivalence in which one term collapses neatly into the other, on the correlation of quality into quantity without remainder. But this gambit of rendering-equivalent is not a description of an already-existing “reality” as much as it is an attempt to make it so, a maneuver intended to bury the social relations that would occlude the capitalist accumulation of wealth.

24To put this differently, digital psychiatry’s attempt to get rid of language is an attempt to “enframe” the subject by getting rid of the excess that would interfere with biocapital’s aspiration to turn the body into an organ of profit-making. It rests on installing a regime of numbers that cast an ideological veil over the proliferation of language in the very scene of the production of quantities.

Trading symptoms for time

25It is another weekday afternoon in the lab, another standard data dump. The ratings scales have been completed, the data from the Subject’s IPG safely downloaded. The Subject is still here, but there are no tasks left to do. It is clear that he does not want to leave. He begins to stall for time, saying that he just took his afternoon doses of Adderall (a powerful psychostimulant), and must wait 20 minutes for the drugs to “kick in.” It’s getting more awkward, increasingly difficult to make eye contact. The subject says he is feeling “too stimulated,” and it’s unclear whether he means by the neuromodulation device or by the drug. “I’m getting a wave of anxiety,” he says.

26I write in my notebook “He is trading symptoms for time.” Time with the RAs and Tony, who he often tells me are his only friends. Time in which he gets to “feel like a human being.”

27The Subject says, “I won’t know for another 20 minutes if I’m okay or if the Adderall has kicked in. I don’t want to drive home and have an adverse reaction.” I write, “He has codeswitched into the language of the clinical trial to mark that these symptoms ‘belong’ to the lab.” He finishes awkwardly, “But I don’t want to waste your time, though.” Tony says that it’s okay, that the Subject can sit in the lobby with us while we wait. It is clear that there is a negotiation happening here. The Subject realizes that Tony is suggesting that we leave the lab office in order to initiate the process of getting him to leave. He mounts a new gambit.

  • 1 The operation and efficacy of the implant depends not only on where the electrodes are placed in th (...)

28“I’m not feeling so good with the new setting of the cortical leads,” he says. He is referring to the fact that the team has adjusted the intensity of the stimulation in his brain implant. The Subject says that he wants to go back to the other setting he had before.1 Changing the stimulation settings will change the electrical activity in the brain, which will then change the numeric quantities that are being used to construct the biomarkers that are the object of the experiment.

29We all troop out of Tony’s office, past the back labs full of mice and through the desert of cubicles until we get to the managing PI’s office. We have to wait for the PI to get off of a call about primate research. When he does, Tony explains that the Subject feels “too stimulated.” The PI hands the Subject the device controller, which the latter holds over his left chest. The PI is speaking calmly in his comforting patter: “I just toggled it back to the old settings, so now it’s on the same settings as it was when you came in, okay?” Immediately – too quickly for it to have an effect – the Subject says, “It feels more balanced now, it was too much before. With the current settings I felt a decrease in mental energy.” 

30With this task dispatched, the Subject is out of time. Tony is leading him out of the office, but the Subject is not ready to go. He stops by the cubicle where the lab’s RAs are clustered, and begins asking questions about the enormous computer screens full of data that they are working on. One of the senior RAs begins to explain: “We have a lot of numbers that we get from the electrodes you have in your head, and we have to classify them.” The RA senses that he should fill the awkward silence with banter, and begins to tell the Subject about an art project he is working on that uses neural signals and turns them into musical tones. The RA says, “That’s what art is all about, distilling experience.” Tony says to the Subject “In a way, that’s what you do with your poetry.”

31The Subject is still looking at the enormous screens filled with numbers, numbers that have been taken from his body. He says, “It’s weird to think that all of that came from inside me.” Everyone falls quiet while we look at dozens of columns of data filling the banks of screens, the algorithms analyzing the columns turning over the figures quietly, like cells in mitosis. Tony says to the Subject: “All of the cells in your body are calculating all the time, much better than any machines. So, in a way, this computer is you, or you are this computer. I mean, when your cells are calculating, they’re not using numbers, but what they’re doing can be represented by numbers.” The Subject is silent, fascinated, as if in a trance. After a few long seconds, Tony offers to walk him to his car. The Subject has no choice but to agree.

32The exchange that has occurred takes the following form. The lab is not interested in the Subject’s language, so the Subject must find something else to offer in order to get what he wants. In its place, he offers psychiatric symptoms, knowing that the lab values the symptom as a bodily state producing the numeric quantities that will be used to extract biomarkers. In return, the members of the lab offer time: recognition and care coded as experimental labor. Yet while everyone involved knows exactly what is going on – the codings, displacements, and swaps that are in play – the stimulation parameters of the neural implant do really get adjusted. The numeric quantities produced by the neural implant’s recording of brain activity that will be used for constructing biomarkers do, in fact, end up being changed. The numbers that were supposed to banish the problem of language have instead ended up infused with it.

33At the core of the production of the datafied self, in the scene of its avant-garde, there is an excess haunting the attempt to produce the dyadic equivalents of biometric correlation. Try as it might to excise it, digital psychiatry cannot free itself from this excess – the logic of language at the heart of a social relation that cannot be reduced to quantification. Seen in this way, the numeric quantification of the subject is not so much a replacement for language and its metaphors, but the ideological insistence that language has been excised in favor of fungible numbers. “It is not as if X quantity is equivalent to Y subjective state,” insists a regime of capital eager to make profits on the promise of digital psychiatry’s regime of quantification, “It really is so.” But simply because capital insists that it has made it so does not mean that it actually has. Just like the wage ideologically veils the gift at the heart of the social relation, digital psychiatry’s insistence that the numbers are “real” veils the logic of language in an attempt to turn life itself into a profit-making data factory.

34There have been many critiques of digital psychiatry and the export of its promise of behavioral control into adjacent fields like advertising. These critiques often rely on the idea that there is a dangerously powerful threat of behavioral engineering through data gleaned on the datafied subject: the regime of numbers is getting too strong, and is stripping the human of its autonomy and free choice (see Stark 2018; for an example of this critique, see Zuboff 2019). Similar ethical concerns haunt neural engineering like DBS, which is often discussed as holding the concerning potential for “brainwashing” or “mind-control” (e.g. Koivumieni & Otto 2014; McCarthy-Jones 2019). In both instances, the critique of these technologies rests on the idea that the numbers are “real,” and that they “work” – possibly too well. What I have attempted to suggest in this paper is a critique with less optimism built in. What if it is the case that these dangerous, invasive experimental procedures to install the Data Factory within the Subject’s body are all in the service of numbers that aren’t as “real” as they claim to be; that they don’t, in fact, “work” to straightforwardly transform subjectivity into a matrix of fungible numbers? What if the process of speculative biocapital requires only the illusion of “real” numbers in order to prop up its inflated financialized value? What if the numbers are not, in fact, straightforwardly extracted from the body as an effective map of subjectivity, and it still doesn’t matter for the inexorable advance of capital’s colonization of the body?

35Seven months after I leave the lab, American venture capitalist Elon Musk livestreams a product launch for his neural modulation venture called Neuralink. Neuralink has raised an extraordinary $66.3 million in order to build what Musk describes as the “Fitbit of the brain,” a neuromodulation device based on the very same DBS research the lab performs on the Subject. To his investors, Musk delusionally claims that the Neuralink will be able to treat everything from blindness to anxiety (Carr 2020).

36At the time of writing this paper, Elon Musk is the richest man in the world. The Subject writes to me often. He is hoping to feel better soon.

Haut de page

Bibliographie

BIRK RASMUS H. & GABRIELLE SAMUEL, 2020. “Can Digital Data Diagnose Mental Health Problems? A Sociological Exploration of ‘Digital Phenotyping’”, Sociology of Health & Illness no. 42/8, pp. 1873-1887.

CALLON MICHEL, 1998. “An Essay on Framing and Overflowing: Economic Externalities Revisited by Sociology”,in Michel Callon (ed.), The Laws of the Market,. Oxford, Blackwell, pp. 244-269.

CARR DANIELLE, 2020. “Shit for Brains: Elon Musk Is Nowhere Near Achieving Mind Control”, The Baffler, September 29, 2020. Online: https://thebaffler.com/latest/shit-for-brains-carr.

CASTELLS MANUEL, 2010. The Rise of the Network Society, Oxford: John Wiley & Sons.

COOPER MELINDA & CATHERINE WALDBY, 2014. Clinical Labor: Tissue Donors and Research Subjects in the Global Bioeconomy, Durham, NC, Duke University Press.

COSGROVE LISA, JUSTIN M. KARTER, ZENOBIA MORRILL & MALLAIGH McGINLEY. “Psychology and Surveillance Capitalism: The Risk of Pushing Mental Health Apps during the COVID-19 Pandemic”, Journal of Humanistic Psychology no. 60/5, pp. 611-625.

FINS JOSEPH J. & NICHOLAS D. SCHIFF, 2010. “Conflicts of Interest in Deep Brain Stimulation Research and the Ethics of Transparency”, Journal of Clinical Ethics no. 21/2, pp. 125-132.

GORIUNOVA OLGA, 2019. “The Digital Subject: People as Data as Persons”, Theory, Culture & Society no. 36/6, pp. 125-145.

KITCHIN ROB, 2014. The Data Revolution: Big Data, Open Data, Data Infrastructures and Their Consequences, London, Sage.

KOIVUNIEMI ANDREW & KEVIN OTTO, 2014. “When ‘Altering Brain Function’ Becomes ‘Mind Control’”, Frontiers in Systems Neuroscience no. 8, 202.

LAKOFF ANDREW, 2005. “Diagnostic Liquidity: Mental Illness and the Global Trade in DNA”, Theory and Society no. 34/1, pp. 63-92.

LASH SCOTT. “Power after Hegemony: Cultural Studies in Mutation?” Theory, Culture & Society no. 24/3, pp. 55-78.

MARKOFF JOHN & JAMES GORMAN, 2013. “Obama to Unveil Initiative to Map the Human Brain”, The New York Times, April 2. Online: https://www.nytimes.com/2013/04/02/science/obama-to-unveil-initiative-to-map-the-human-brain.html

MASUNAGA SAMANTHA, 2017. “A Quick Guide to Elon Musk’s New Brain-Implant Company, Neuralink”, Los Angeles Times, April 21. Onlie: https://www.latimes.com/business/technology/la-fi-tn-elon-musk-neuralink-20170421-htmlstory.html

MAUSS MARCEL. The Gift: The Form and Reason for Exchange in Archaic Societies, New York, W.W. Norton.

McCARTHY-JONES SIMON, 2019. “The Autonomous Mind: The Right to Freedom of Thought in the 21st Century”, Frontiers in Artificial Intelligence no. 2, 19.

PICKERSGILL MARTYN, 2019. “Digitising Psychiatry? Sociotechnical Expectations, Performative Nominalism and Biomedical Virtue in (Digital) Psychiatric Praxis”, Sociology of Health & Illness no. 41, pp. 16-30.

RALEY RITA, 2013. “Dataveillance and Countervailance”, in Lisa Gitelman (ed.), “Raw Data” Is an Oxymoron, Cambridge, MA, MIT Press, pp. 121-146.

RITZER GEORGE, 2014. “Prosumption: Evolution, Revolution, or Eternal Return of the Same?” Journal of Consumer Culture no. 14/1, pp. 3-24.

ROSENBERG CHARLES E., 2002. “The Tyranny of Diagnosis: Specific Entities and Individual Experience”, The Milbank Quarterly no. 80/2, pp. 237-260.

ROSKAMS-EDRIS DYLAN, STACEY ANDERSON-REDICK, ZELMA H. KISS & JUDY ILLES, 2017. “Situating Brain Regions among Patent Rights and Moral Risks”, Nature Biotechnology no. 35/2, pp. 119-121.

RUNWAL PRIYANKA, 2020. “5 Takeaways from Venture Capital’s Growing Focus on Mental Health”, STAT Health News, September 18. Online: https://www.statnews.com/2020/09/18/venture-capital-mental-health-tech/.

SEMEL BETH MICHELLE, 2019. “Speech, Signal, Symptom: Machine Listening and the Remaking of Psychiatric Assessment”, Ph.D. thesis, Massachusetts Institute of Technology.

STARK, LUKE, 2018. “Algorithmic Psychometrics and the Scalable Subject”, Social Studies of Science no. 48/2, pp. 204-231.

STARK, 2020. “The Emotive Politics of Digital Mood Tracking”, New Media & Society no. 22/11, pp. 2039-2057.

SUNDER RAJAN Kaushik, 2006. Biocapital: The Constitution of Postgenomic Life, Durham, NC, Duke University Press.

THRIFT NIGEL, 2005. Knowing Capitalism, London, Sage. Online: https://sk.sagepub.com/books/knowing-capitalism

ZUBOFF SHOSHANA, 2019. The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power, London, Profile Books.

Haut de page

Notes

1 The operation and efficacy of the implant depends not only on where the electrodes are placed in the brain, but also on the amount of electrical stimulation (amplitude) the electrode distributes. Typically, a higher setting will produce an intensified effect.

Haut de page

Pour citer cet article

Référence électronique

Danielle CARR, « Excess in the avant-garde of the data subject »Terrain [En ligne], 76 | 2022, mis en ligne le 03 mai 2022, consulté le 28 mars 2024. URL : http://journals.openedition.org/terrain/23363 ; DOI : https://doi.org/10.4000/terrain.23363

Haut de page

Auteur

Danielle CARR

Columbia University, Department of Anthropology

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search