This document is intended to bring engineers and technical collaborators up to speed on the core concepts behind our neurotechnology work. It is written for people who are already comfortable with product/software development and design, but who may be new to brain-computer interfaces, neuroimaging, neurostimulation, neurofeedback, and neuromodulation.

Our work is oriented toward an ecosystem of neurotechnology products, training protocols, and clinical tools. One side of that ecosystem is performance-oriented: tools, simulations, and training experiences for athletes and other users who want to improve focus, regulation, recovery, and cognitive or psychomotor performance. The other side is clinical: neurotechnology intended for psychological and therapeutic contexts such as ADHD, anxiety, mood regulation, and related conditions. Finally, our work also considers brain-computer interfaces and how they can be deployed for assistive technologies and beyond. Throughout this primer, those three application contexts are kept in view.

---

A brain-computer interface (BCI) is a system that measures brain activity, extracts useful information from that activity, and converts it into a meaningful output. That output could be control of a cursor, a selection in a user interface, a change in a simulation, a feedback signal for training, or an adaptive response in software.

At a systems level, a BCI usually consists of the following pipeline:

1. Signal acquisition - measuring neural activity using a sensing modality such as EEG.

2. Preprocessing - cleaning the signal and removing artifacts such as eye blinks, muscle tension, and environmental noise.

3. Feature extraction - identifying useful signal components such as oscillatory power, event-related responses, or hemodynamic changes.

4. Decoding or classification - mapping those features to an intended state, command, or estimate.

5. Output and feedback - turning that estimate into an action, control signal, or closed-loop training response.

A useful early correction is that BCIs are usually not mind-reading machines in the science-fiction sense. Most real systems decode constrained and task-relevant signals: attention to a flickering stimulus, attempted movement, workload, arousal, speech-related motor patterns, or changes in self-regulatory state.

BCIs can be grouped in several useful ways.

- Non-invasive BCIs record from outside the skull, most commonly using EEG and sometimes fNIRS.

- Minimally invasive BCIs move sensors closer to the brain while avoiding some of the burden of full intracranial implants.

- Invasive BCIs place electrodes on or in the brain, usually gaining signal quality and bandwidth at the cost of surgery, risk, and complexity.

This is one of the most important design trade-offs in neurotechnology. The closer you get to neurons, the better the signal tends to be, but the greater the burden on the user.

- Active BCIs depend on intentional self-generated neural activity, such as motor imagery.

- Reactive BCIs decode brain responses to external stimuli, such as visual flicker or auditory rhythms.

- Passive BCIs do not require explicit control at all. Instead, they estimate the user’s state, such as fatigue, attention, stress, workload, or emotional arousal.

This passive category is especially relevant to our work because many real products are better framed as neural state interfaces than pure thought-control devices.

- Discrete selection systems choose letters, menu items, or targets.

- Continuous control systems drive cursors, robotic devices, or speech synthesis.

- Closed-loop training or therapy systems use the brain signal to adapt feedback, stimulation, or training conditions.

For our purposes, the third category is especially important. Many of the most practical near-term applications are not about replacing mouse and keyboard input, but about helping users regulate, train, recover, or optimize their state.

The field has advanced rapidly, but progress is uneven across applications.

- The strongest near-term clinical successes are in assistive communication and device control for people with severe motor impairment.

- Non-invasive BCIs, especially EEG-based systems, remain the most practical path for portable and consumer-adjacent products.

- Invasive BCIs currently offer the highest performance for speech decoding and fine motor control, but they come with major surgical and regulatory burdens.

On the consumer side, most currently available systems are better described as brain sensing, neurofeedback, or attention-state platforms than as high-bandwidth thought-control interfaces. That does not make them trivial. It simply means the most realistic value is often in state monitoring, training, and adaptive experience design rather than cinematic command-and-control.

Consumer and developer-oriented neurotech products today often combine portability, ease of use, and lower signal fidelity. These systems are useful for meditation, focus training, cognitive-state monitoring, neurofeedback, experimental interfaces, and prototyping.

At the frontier industry end, invasive BCI companies are pushing speech restoration, computer control, and high-bandwidth motor decoding. These efforts matter strategically, but they solve a different problem than the one most product teams face. They are optimizing for maximum neural bandwidth and clinical restoration, not for lightweight deployment, repeated use, or home-based performance training.

Our opportunity is probably not to compete directly in the high-risk implant race. A more strategic contribution lies in usable, adaptive, non-invasive, and closed-loop neurotechnology.

The major bottlenecks in real-world BCI are still deeply engineering-shaped:

- setup friction

- signal noise

- artifact sensitivity

- user variability

- long calibration times

- poor UX

- weak transfer from lab tasks to real life

This is good news for a multidisciplinary team. It means progress will not come only from better electrodes or better decoders, but also from better interaction design, multimodal sensing, adaptive training systems, confidence-aware software, immersive feedback, and product-quality experiences.

For us, the strongest framing is likely this:

That framing naturally supports both sides of our ecosystem: performance training and clinical intervention.

---

Neuroimaging methods used in neurotechnology generally fall into two big families:

- methods that measure electrical activity directly

- methods that measure hemodynamic consequences of neural activity, such as blood oxygenation and blood flow

Electroencephalography (EEG) records voltage fluctuations from electrodes placed on the scalp. The recorded signal mainly reflects the summed and synchronous postsynaptic activity of large populations of cortical neurons, especially pyramidal cells.

Because the signal has to pass through brain tissue, cerebrospinal fluid, skull, and scalp before reaching the sensor, it becomes attenuated and spatially blurred. Even so, EEG remains one of the most important tools in neurotechnology because it captures neural dynamics at very high temporal resolution.

- Excellent temporal resolution. EEG is very good for tracking fast changes in brain state, oscillations, event-related responses, and timing-sensitive dynamics.

- Portable and relatively affordable. It is far more deployable than MRI-based methods.

- Suitable for real-time feedback and BCI. EEG can be streamed with low latency, making it useful for neurofeedback, adaptive systems, and brain-state monitoring.

- Strong ecosystem. Hardware, software, and analysis workflows are relatively mature.

- Poor spatial precision. EEG is much better at telling you when something changed than exactly where it came from.

- Mostly cortex-biased. It is not a strong direct window into deep structures.

- Artifact-prone. Eye movements, facial muscles, body motion, and electrical noise can all distort the signal.

- User-to-user variability. Anatomy, hair, skin, impedance, and state can all affect signal quality.

A good shorthand is that EEG gives you a fast but blurry view of brain dynamics.

Functional near-infrared spectroscopy (fNIRS) uses near-infrared light to estimate changes in oxygenated and deoxygenated blood in superficial cortex. Rather than measuring neural firing directly, it measures the vascular response that follows local neural activation.

This makes fNIRS a hemodynamic method rather than an electrical one. It mainly accesses cortex close to the scalp and is especially useful when the target is a relatively slow change in frontal or motor cortical state.

Hemoencephalography (HEG) is best thought of as a simpler, often more feedback-oriented cousin of fNIRS. It is generally more lightweight and narrower in scope, often used in neurofeedback contexts rather than rich multichannel imaging.

- Better superficial localization than EEG for some cortical regions

- Portable and quiet

- Tolerant of more naturalistic environments than many imaging methods

- Useful for slow state estimation, executive function monitoring, stress regulation, and frontal activation tracking

- HEG can be simple and pragmatic for frontal neurofeedback applications

- Slow. fNIRS and HEG track hemodynamic changes that unfold over seconds, not milliseconds.

- Indirect. They do not measure neural activity directly.

- Limited depth. They mainly capture superficial cortex.

- Sensitive to physiology and placement. Hair, scalp blood flow, anatomy, and systemic changes can affect the signal.

- HEG is less standardized and less information-rich than full multichannel systems.

A useful comparison is:

- EEG reads the electrical weather

- fNIRS and HEG read the vascular afterglow

Our current projects should remain focused on non-invasive approaches, but it is useful to understand the sensing ladder beyond that.

Sub-scalp EEG places electrodes under the scalp but above the skull. It is a middle-ground approach that can improve signal quality relative to scalp EEG without reaching full intracranial invasiveness.

Endovascular systems place electrodes within blood vessels adjacent to the brain. The Stentrode concept is a prominent example. These approaches aim to get closer to cortical signals without a craniotomy.

Electrocorticography (ECoG) places electrode arrays directly on the cortical surface. This provides better signal quality and access to higher-frequency activity than scalp EEG, but still requires surgery.

These electrodes are inserted into deep and superficial brain regions. They are useful when the important dynamics are distributed across three-dimensional brain structures.

These are penetrating electrodes inserted into cortex. They offer the richest signals and support the highest-performance BCIs, but they are also the most invasive and technically demanding.

The ladder looks roughly like this:

- EEG: safest, easiest, fastest, blurriest

- fNIRS / HEG: slower, hemodynamic, practical for state tracking

- Sub-scalp / endovascular: middle-ground attempts to improve signal quality

- ECoG / SEEG: high-quality clinical and research interfaces

- Intracortical arrays: highest bandwidth, highest burden

---

If neuroimaging is about reading brain activity, neurostimulation is about perturbing it. The goal is usually not to inject a precise thought or force a deterministic output. Rather, it is to shift the brain’s operating conditions by changing excitability, synchrony, plasticity, or network state.

Different methods use different kinds of energy:

- light

- sound

- electrical current

- magnetic fields

- ultrasound

- infrared light

These methods differ in depth, focality, strength, safety profile, maturity, and ease of deployment.

Photic stimulation uses rhythmic flashes of light to evoke time-locked responses in visual cortex. When brain activity begins to align with the stimulation frequency or its harmonics, this is often called photic driving.

This is one of the most accessible ways to interact with brain rhythms. It is inexpensive, easy to prototype, and naturally compatible with displays, games, immersive experiences, and visual feedback systems.

However, photic stimulation is highly state-dependent and must be handled carefully because rhythmic visual stimulation can provoke abnormal responses in susceptible individuals.

Auditory stimulation uses tones, clicks, rhythms, or modulated sound to influence brain activity. It can entrain auditory pathways and is especially relevant for sleep protocols, arousal shaping, and adaptive sound-based environments.

Closed-loop auditory stimulation during sleep is a particularly interesting area, where sound is timed relative to ongoing slow-wave activity in an attempt to enhance beneficial rhythms.

Binaural beats involve presenting slightly different frequencies to each ear so that the listener perceives a beat frequency equal to the difference. They are often marketed as brainwave entrainment, but the evidence is mixed.

The sensible stance is to view binaural beats as a lightweight state-modulation tool rather than as a precision neuromodulation technology.

Transcranial electrical stimulation (tES) is an umbrella category for methods that pass weak electrical current through scalp electrodes.

Transcranial direct current stimulation (tDCS) applies a weak constant current. It is generally understood as biasing cortical excitability and plasticity rather than directly driving firing.

Transcranial alternating current stimulation (tACS) applies oscillating current at a chosen frequency. It is especially attractive for interacting with ongoing brain rhythms and attempting rhythm-specific modulation.

One important additional variant is tRNS, or transcranial random noise stimulation, which applies randomly varying current. It has been explored for perceptual learning, excitability modulation, and cognitive effects.

- relatively affordable

- portable

- compatible with repeated and even supervised home use

- conceptually useful for rhythm and excitability modulation

- not highly focal

- variable across individuals

- sensitive to montage, task, timing, and anatomy

- effect sizes can be modest and inconsistent

Transcranial magnetic stimulation (TMS) uses rapidly changing magnetic fields to induce currents in cortex. Compared with tES, it can stimulate more strongly and more focally.

Single-pulse TMS is often used to probe excitability. Repetitive TMS and patterned bursts such as theta-burst stimulation are used to induce longer-lasting changes.

TMS is one of the most clinically mature non-invasive neuromodulation technologies and is already established in the treatment of disorders such as depression.

Focused ultrasound uses acoustic energy concentrated at a target. This is strategically important because it may eventually allow deeper and more spatially precise modulation than many other non-invasive tools.

High-intensity ultrasound is already used clinically for targeted ablation in selected indications. Lower-intensity ultrasound for reversible neuromodulation is exciting but remains more frontier-lab than routine deployment.

Infrared approaches include infrared neural stimulation and transcranial photobiomodulation.

The latter is especially relevant for non-invasive neurotech. It aims to influence metabolism, blood flow, mitochondrial processes, and downstream brain function using near-infrared light. It is promising, but still not a fully mature mainstream modality.

For our work, the most immediately actionable neurostimulation tools are likely:

- photic stimulation

- auditory stimulation

- selected forms of tES

TMS is an important clinical benchmark. Ultrasound and infrared methods are strategically important to watch, but sit further out on the frontier.

---

Neurofeedback is a form of closed-loop training in which a person receives real-time information about their own brain activity and learns, through repetition, to shift that activity in a desired direction.

The core principle is usually a type of operant conditioning or reinforcement learning. When the target brain signal moves in the desired direction, the system rewards that movement with feedback. Over repeated sessions, the user may learn better self-regulation of the relevant brain state.

This makes neurofeedback especially relevant for our ecosystem because it is not just a sensing tool. It is a training framework.

1. Measure the signal.

2. Clean and process it.

3. Extract a target feature.

4. Map that feature into meaningful feedback.

5. Let the user learn through repeated interaction.

The quality of the experience matters. Neurofeedback success does not depend only on signal processing. It also depends on target choice, protocol design, feedback design, user engagement, repetition, and individual differences.

With EEG neurofeedback, the target might be:

- sensorimotor rhythm

- alpha power

- theta/beta ratio

- alpha/theta balance

- slow cortical potentials

The system turns those features into a visual, auditory, or interactive signal that the user tries to influence.

In performance contexts, EEG neurofeedback is often used for:

- attentional control

- arousal regulation

- stress management

- pre-performance state optimization

- psychomotor steadiness

In clinical or psychological contexts, EEG neurofeedback has been explored for:

- ADHD

- anxiety

- mood regulation

- trauma-related dysregulation

- epilepsy-related applications

- executive function training

The evidence base is uneven. Some areas are promising, some are mixed, and many studies suffer from inconsistent protocols and weak standardization.

With fNIRS neurofeedback, the target is hemodynamic activity in superficial cortical regions, often prefrontal cortex.

This makes it attractive for:

- executive control training

- frontal regulation tasks

- emotional regulation protocols

- cognitive effort and workload training

- stress and self-regulation routines

Compared with EEG neurofeedback, fNIRS neurofeedback is slower but often more spatially grounded for accessible cortical regions.

For a performance-focused training center, neurofeedback is most plausibly useful for:

- state optimization

- attentional stability

- stress resilience

- emotional steadiness

- recovery and self-regulation training

- targeted performance routines for specific tasks

For a clinical setting, neurofeedback is best viewed as a structured self-regulation intervention or adjunctive therapy tool rather than a miracle replacement for other treatments.

- non-invasive

- repeatable

- compatible with personalization

- naturally suited to software-rich and immersive experiences

- gives the user an active role in self-regulation

- fits both clinic and performance-center models

- strong individual variability

- protocol quality matters enormously

- not everyone learns equally well

- neural changes during training do not always translate into durable real-world benefit

- evidence quality varies a lot across conditions

For our team, neurofeedback should probably be treated as a platform capability, not just a single product. It is a way to build closed-loop training experiences that combine sensing, visualization, simulation, coaching, and adaptive difficulty.

This is where our technical-art and simulation capabilities can become a major differentiator. Neurofeedback experiences do not need to look like a bar graph from 2007. They can become rich, embodied, motivating environments that train self-regulation in a way users actually want to repeat.

---

Neuromodulation is the strategic use of stimulation to shift how the brain functions. If neurofeedback is about teaching a person to regulate their own state through feedback, neuromodulation is about using external stimulation to bias the underlying system itself.

In practice, neuromodulation can be used to:

- increase or decrease cortical excitability

- reinforce or disrupt rhythms

- influence connectivity between regions

- prime learning and plasticity

- bias emotional, attentional, or motor states

The most important idea here is that neuromodulation is usually state-dependent. The same protocol can help, do nothing, or impair performance depending on the target, timing, task context, and individual.

In performance settings, neuromodulation is often discussed in relation to:

- focus and concentration

- reaction time

- motor output and skill acquisition

- fatigue perception

- stress regulation

- recovery

- cognitive readiness

The evidence is intriguing but inconsistent. There is real reason to think stimulation can sometimes improve performance-related states or learning conditions, but effect sizes are often modest, protocol-sensitive, and not uniformly reliable.

For this reason, neuromodulation is probably best used as a targeted amplifier within a training protocol, not as a standalone enhancement gadget. It may work best when paired with:

- cognitive drills

- psychomotor training

- immersive simulation

- neurofeedback

- breathwork and regulation routines

- task-specific performance preparation

In clinical contexts, neuromodulation is easier to justify because the goal is often to shift dysfunctional network states toward healthier ones.

Relevant targets include:

- mood regulation

- anxiety and threat reactivity

- emotional control

- executive dysfunction

- sleep and arousal regulation

- trauma-related dysregulation

This is one of the strongest current application areas, especially for TMS and increasingly for tDCS.

Neuromodulation is being actively explored for anxiety disorders, OCD, PTSD, and related presentations. The evidence is encouraging in some areas but still heterogeneous.

Sleep is a rhythm-rich process and therefore an obvious target for neuromodulation. However, the field is still developing, and sleep-focused neuromodulation should currently be treated as promising R&D terrain rather than fully settled standard practice.

For our ecosystem, neuromodulation should be viewed as a state-shaping layer that can work alongside sensing, feedback, and training.

- In the performance center, it may help improve readiness, regulation, learning efficiency, and focus.

- In the clinic, it may support interventions for mood, anxiety, attention, sleep, and self-regulation.

The long-term opportunity is likely closed-loop neuromodulation: systems that sense brain state, apply stimulation intelligently, and integrate that stimulation into guided training, immersive simulations, and adaptive protocols.

The real frontier is not just stimulating the brain. It is building systems that stimulate the right network, in the right state, for the right goal, inside the right experience.

---

Across all of these domains, a few themes repeat:

1. State matters. Brain function is dynamic, context-sensitive, and variable.

2. Closed loops matter. The most interesting systems are not passive measurement tools but interactive loops that adapt to the user.

3. Usability matters. Product value depends as much on setup, comfort, engagement, and repeatability as on raw signal quality.

4. Multimodal systems matter. Combining brain signals with other physiological and behavioral signals will likely be more powerful than relying on one modality alone.

5. Experience design matters. Visualization, interaction, simulation, and training structure are not cosmetic. They are part of the intervention.

6. Evidence matters. Some applications are mature, some are promising, and some are still speculative. We should remain ambitious without becoming gullible.

If we do this well, our work does not need to look like generic gadget neurotech. It can become a serious ecosystem of tools, training environments, and clinical experiences that help people regulate, learn, recover, and perform better.

That is likely where our strongest contribution lies: not just building sensors or stimulation devices in isolation, but creating integrated neurotechnology experiences that are scientifically grounded, technically robust, and genuinely compelling to use.

---