.png){kind=logo}
Chapter 2: Harnessing Emotion through Neurotechnology
- JM Wesierski
- Sep 21, 2025
- 8 min read
Are emotions good for us?
Many are considered beautiful but undeniably destructive when left unchecked. We each seek love and happiness, but such quests can lead to unhealthy obsession or dependence. Without natural and regular fulfilment, a person may turn to dangerous alternatives like sexual promiscuity or drug abuse.
Is the occasional smile and laughter worth the potential for anxiety and depression? Some practices, such as Greek Stoicism or the Buddhist Upeksha, suggest developing cognitive strength and equanimity — calmness and composure, especially in a difficult situation — to prevent our emotions from interfering with rational and mindful thought.
While the idea of tempering the full emotional spectrum, including both the highs and the inevitable lows, may seem appealing, for many of us emotions remain an intrinsic part of being human. Rather than suppressing or eliminating them, our real desire is learning to harness and channel these affective states in ways that serve us — before they take control.
Rewiring our Affective States
Affective states encompass the range of psycho-physiological ways in which we are affected by an experience. These states include both specific emotions, such as fear or joy, and lasting sentiments, like depression or irritability.
Hess and Thibau, 2009, building on earlier work by Charles Darwin, describe these states as a link to our primal past. In an age of hunting and being hunted, emotions developed to foster co-operation and reproduction, forming part of the bio-evolutionary concept of ‘fitness’. Affective states evolved in complexity along with the rest of our psyche, yet while our knowledge compounded to build shelters into sky scrapers and turned game into cuisine, emotions became confused, transferring anxiety to public speaking and envy to Instagram feeds. Psychologist Jonathan Haidt likened this confusion to a rational rider steering an overpowering, emotional elephant.
These affective states can be rewired, yet modern methods of assessing them often fail to reliably diagnose the experience — or its cause — due to limitations of subjective descriptions, and overlapping or unrecognizable psychosomatic symptoms, that is, physical symptoms triggered by emotional states.
Neuropsychology
Whether someone considers emotions to be good or not, neuroscientists assign a positive or negative valence to differentiate their affects. The strength of reaction, whether long lasting or short, is referred to as arousal. Russell (1980) proposed a circumplex model, shown below, on which most emotions, feelings, moods, and attachments of affective states can be plotted as a degree of valence and arousal. For example, emotions with low arousal are categorized positively as ‘relaxed’, or negatively as ‘bored’.
Emotional valence is considered to be a bottom-up experience, wherein an individual’s physiological arousal in response to stimulation, such as having to give a presentation leading to increased heartbeat or sweat, is interpreted by the mind in the context of the given the situation and accordingly assigned a valence. “They are guesses that your brain constructs in the moment,” claims Dr. Lisa Feldman, Professor of Psychology at Northeastern University. Therefore, “when faced with an uncomfortable task, like public speaking, don’t allow yourself to succumb to negative anxiety by dwelling on what could go wrong. Instead, remind yourself that you prepared, and are simply pumped up with positive excitement,” advises J.M. Wesierski, Lead of Emotion Research at MindsApplied, “This way, you can regulate arousal and reinterpret your emotional states in service of wellbeing.” Through non-invasive and commercially available neurotechnology, such as an EEG, arousal can be reliably discerned via power-spectral density by calculating the ratio of alpha to beta frequency band activity, with greater beta indicating arousal. Other technologies used to ascertain emotion may rely on eye-tracking, facial expressions, skin conductivity and heart rate.
However, identifying objective valence with consistent accuracy requires a machine learning (ML) system to extract meaningful patterns in the variability and volume of neural data that correspond to specific emotional states.
MindsApplied
This article explores the creation of such a system by MindsApplied, an organization purveying commercial neurotechnology.
Specifically, we trained a generalizable (i.e. transfer learning) ML model to predict degrees of valence in real-time for commercial purposes, called your Minds AI. Such emotion-recognizing technology is being used across the fields of:
Art — visualizing affective states
Gaming — creating environments which react to an individual’s affective states
Mental Health — visual graphics to support interoception, informing diagnostics and treatment
Neuromarketing — monitoring the neurological response of consumers towards products
Samples of such applications have been included in MindsApplied’s Neural User Interface (your Minds UI).
Powered by Artificial Intelligence (AI)
The ML model is first trained on a large base of neural data taken from many people during their experience of target affective states, and subsequently applied to the neural data of an individual experiencing an unknown emotion to be predicted.
Calibration consists of all participants being exposed to various short movie scenes or song clips discerned to induce emotion, after which they provide feedback on their emotional experience using a self-assessment manikin (SAM) scale for valence and arousal, shown below. After calibration, the user is then provided with a personalized predictive model designed to classify aspects of their emotional state such as strength and multitude.
Some techniques for classifying emotional state from EEG data include spatially informed signal filters to enhance EEG’s low spatial resolution, physics-informed filters to consider activity across brain regions, and data augmentation to help reconstruct and generate further useful data from what is available. Model prediction accuracy can be reduced due to limitations in data quality from EEG interference, inaccurate SAM reporting, and delays between calibration and real-time prediction.
As emotion-prediction technologies advance, safeguarding neural data becomes essential. Emotional states derived from EEG and related biosignals may be considered sensitive health information, especially when used in clinical or commercial settings. Adopting data privacy frameworks such as GDPR, CCPA, and HIPAA, as well as prioritising privacy by design through informed consent, data minimization, and secure storage, is key to ensure ethical integration of neurotechnology into everyday applications.
Cutting-Edge Applications
Art
Neuroart has been used by communities like EDGE to visualize affective states digitally and physically to provide introspective meaning to their artwork. Windows of neural data corresponding to target states, such as laughing or being told “I love you”, provide the basis for ornate and uniquely meaningly works.
For live visuals, Neurovision makes use of a user’s brain’s polarity and changes in arousal to create colorful, dynamic, and abstract experiences: slow transitions in the artwork reflect a calm brain, while explosions and higher speed movement signify spikes of excitement.
Such expressive technologies are being used by performers to give audiences a deeper insight into their creative process, as shown in the video above. Introducing emotion prediction into such technology provides novel ways to understand the workings of an artist’s mind, such as sadness through a thunderstorm or stress via a volcano on the brink of eruption. Beyond aesthetic purposes, emotion prediction expands use cases of neurotechnology such as in gaming experiences, neurofeedback modulation of target states, and marketing evaluations.
Extended Reality (XR)
Similarly to art, the ability of emotion-predicting neurotechnology to affect a reactive environment can increase the immersiveness of gaming and virtual realms. Neurotechnology has made a significant jump into the commercial space as part of biosensing XR technology that already makes use of the head, such as Meta’s glasses and Galea’s VR headset.
Specific applications are widespread: For horror games, prolonged anxiety can be used to assess the perfect time to scare a user. In a contrasting scenario, detecting unwanted anxiety can allow an experience to adjust to be happier using features like sunlight or calming music. This example indicates how predictive neurotechnology expands the intersection of gaming — a widely adopted public pastime — and mental healthcare, which is slow to be sought due to poor accessibility, stigma, limited personal health literacy, and lack of interoceptive awareness.
Mental Health
Mental Health stands at the forefront of the neurotechnology discourse. Almost a fifth of American adults regularly suffer from a degree of anxiety or depression. Both invasive and non-invasive neurosensing devices have been paired with solutions spanning robotics, gaming, mobile applications and more to address this epidemic faced by societies worldwide.
Digitally, an effective neurotechnological solution for therapists and patients to engage with mental health is neurofeedback — the practice of using real-time signals of brain activity to modulate affective states. Neurofeedback applications also often focus on visualizations and gaming.
Emotion-prediction technology can provide more granular and informative insights into affective states discussed in the context of mental health therapy or long-term monitoring than first hand descriptions, which may be minimally reliable. In combination with expressive visualization and gamified technologies, it can make cognitive and emotional regulation more achievable for both clinicians and individuals. Using AI to improve brain signal processing, we can better understand individuals experiences — this is a foundational step that can enable personalized diagnostics and treatments.
Neuromarketing
Emotion-prediction AI enables access to deeper, previously inaccessible insights regarding how consumers respond to products. Neuromarketing typically relies on measurement of physical or physiological signals corresponding to arousal and attention, such as facial expression and skin conductance, to objectively approximate implicit cognitive processes that traditional self-report methods often miss, or misrepresent.
However, these signals alone fail to reflect the full nuance of a person’s emotional spectrum. By leveraging AI-based decoding of neural data in real time, marketers can accurately and dynamically interpret a user’s cognitive and emotional state, enabling adaptive experiences that tailor content, interfaces, and product offerings to the individual’s live engagement. Emotion-prediction neurotechnology provides richer data to make creation and advertising of products more beneficial for consumers and companies alike.
Final Feelings
Defining emotion is not simple, and harnessing true feelings, whether for art, health, or business is equally challenging. While some traditions emphasize detachment or emotional discipline, a more balanced goal is not to eliminate emotion — but to hold a steady hand on the elephant’s reins.
The reward for accessing our untapped, objective emotional states will include the revolutionization of mental health care and the deepening of content immersion. AI, when enabled by accurate and secure data, allows us to provide more personalized and targeted experiences that make the most of our emotional spectrum.
Though they may be categorized as positive or negative, no one can truly say whether emotion is inherently good or bad. Yet, it is perhaps our most human trait which we must control within ourselves and use to its full potential.
Written by JM Wesierski, edited by Lars Olsen and Benjamin Schornstein, with original artwork by Lina Cortéz.
JM Wesierski has a background in business, technology, art, and neuroscience. He is the co-founder of MindsApplied, a company focused on fostering commercial adoption of neurotechnology. They use cutting edge artificial intelligence to create and improve neuro-applications targeting communication, interaction, and mental health.
Lars Olsen is a regulatory medical writer. He works in the pharmaceutical industry writing submission-level documents, and has additional experience with medical devices and pharmacovigilance.
Benjamin Schornstein became interested in BCIs during his undergraduate experience at Brown University, where he was a double major in computer engineering and neuroscience working in the BrainGate lab to restore autonomy of movement for individuals with ALS, spinal cord injury, and other forms of tetraplegia. He is a Fulbright scholar in Switzerland, where he is working on developing BCIs at the NeuroRestore lab.
Lina Cortéz is an electronics engineer and neurotechnology enthusiast who is highly interested in the application of brain-computer interface in robotics control.
Comments