Tools of Measurement: EEG, Hyperscanning, and the CaMBRAIN Project — How We Know What We Know About Brainwaves

In Part 1, we explored the five brainwave channels. But how do we know the brain operates at 0.5 Hz in deep sleep and 40 Hz during peak cognition? The answer lies in a century of measurement technology — from a German psychiatrist's homemade electrodes to AI-powered real-time inference models that can read your brain state from raw electrical noise.

How EEG Works: The Physics of Brainwave Detection

The electroencephalogram (EEG) is the oldest and most widely used method for measuring brainwave activity. Its principle is surprisingly simple: when large populations of neurons fire in synchrony, their combined electrical activity generates an electrical field strong enough to be measured from electrodes placed on the scalp.

A single neuron's electrical activity is vanishingly small — measured in microvolts (millionths of a volt). But when hundreds of thousands of cortical pyramidal neurons fire in phase, their fields summate. The resulting signal — typically 10–100 microvolts — can be detected by silver-silver chloride electrodes placed on the scalp.

The physics is straightforward: EEG measures the difference in voltage between pairs of electrodes over time. These voltage fluctuations are amplified (typically 1,000–10,000 times), digitized, and recorded as a continuous waveform. The waveform is a complex mixture of all the brain's oscillatory activity — delta, theta, alpha, beta, and gamma all superimposed on one another.

The challenge is separating them.

The 10-20 System: Mapping the Scalp

To make EEG measurements comparable across individuals and laboratories, the international 10-20 system standardizes electrode placement. Named for the fact that electrodes are placed at 10% and 20% intervals of the skull's major landmarks (nasion to inion, and left to right pre-auricular points), the system includes:

• Fp1/Fp2 — Prefrontal cortex (executive function, planning)
• F3/F4, F7/F8 — Frontal lobes (attention, motor planning)
• C3/C4, Cz — Central sulcus (sensorimotor cortex)
• T3/T4, T5/T6 — Temporal lobes (auditory processing, memory)
• P3/P4, Pz — Parietal lobes (sensory integration, spatial awareness)
• O1/O2, Oz — Occipital cortex (visual processing)

A standard clinical EEG uses 19–21 electrodes. Research-grade systems often use 64, 128, or even 256 electrodes for higher spatial resolution. The position matters: alpha waves are strongest over the occipital cortex; frontal theta dominates during meditation; gamma is most prominent over fronto-central regions.

From Raw Signal to Frequency Bands: The Fourier Transform

Here is where the real magic happens. The raw EEG signal is a chaotic-looking waveform — a squiggly line that seems to have no pattern. But embedded in that squiggle are multiple overlapping frequencies.

To extract them, neuroscientists use a mathematical technique called the Fourier transform (specifically the Fast Fourier Transform, or FFT, developed by Cooley and Tukey in 1965). The Fourier transform decomposes any complex waveform into its constituent sine waves at different frequencies.

Think of it like separating a musical chord into its individual notes. A C-major chord contains C, E, and G played simultaneously — they sound as one rich sound, but a Fourier analysis reveals each note. Similarly, the raw EEG contains delta, theta, alpha, beta, and gamma all at once. The FFT reveals how much power (amplitude) exists at each frequency.

The result is a power spectrum — a graph showing the amount of electrical energy at each frequency. Peaks in this spectrum tell us which frequency bands are dominant at any given moment. A healthy waking adult with eyes closed will show a prominent alpha peak around 10 Hz. Someone in deep meditation may show a theta peak. Someone with Alzheimer's may show reduced gamma.

The History: Hans Berger and the Discovery of Brainwaves

In 1924, German psychiatrist Hans Berger recorded the first human EEG by placing silver wires under the scalps of patients with skull defects. He called the resulting signal the Elektrenkephalogramm. His work was met with skepticism — other scientists could not replicate his results with less sensitive equipment.

Berger persisted. Over the next 14 years, he published 14 papers documenting the alpha rhythm (which he called the "Berger rhythm"), the effects of eye opening and closure on brainwaves, and the changes that occur during sleep and mental activity. He also noted that brainwave patterns change with age and that certain abnormalities correlate with epilepsy.

Modern neuroscience stands on Berger's shoulders. Every EEG recorded today — from clinical diagnostics to consumer neurofeedback to the CaMBRAIN project — traces its lineage to his homemade electrodes and relentless observation.

Modern Quantitative EEG (qEEG) and Source Localization

Raw EEG is useful, but modern analysis adds layers of sophistication.

Quantitative EEG (qEEG) applies statistical analysis to the raw signal, producing maps of brain activity that can be compared against normative databases. A qEEG report typically shows:

• Absolute power — How much energy at each frequency band
• Relative power — The percentage of total power in each band
• Coherence — How synchronized different brain regions are
• Phase lag — The timing relationship between regions
• Asymmetry — Differences between left and right hemispheres

Source localization (e.g., LORETA — Low Resolution Brain Electromagnetic Tomography) attempts to solve the "inverse problem" — estimating where in the brain the measured scalp activity originated. While mathematically challenging (many possible source configurations could produce the same scalp pattern), modern algorithms can localize activity to specific cortical regions with reasonable accuracy.

Hyperscanning: Measuring Two Brains at Once

Perhaps the most intriguing development in EEG methodology is hyperscanning — measuring the brain activity of two or more people simultaneously while they interact.

Pioneered by researchers like Uri Hasson at Princeton and Guillaume Dumas at the Institut Pasteur, hyperscanning studies have revealed extraordinary phenomena:

• During conversation, speakers' and listeners' brain waves synchronize word by word, with the listener's brain activity mirroring the speaker's with a short delay [1].
• During cooperative tasks, participants show increased inter-brain coherence in the gamma and alpha bands.
• During eye contact, the brains of two people synchronize in the theta band.
• During musical performance, the brains of ensemble musicians synchronize during coordinated passages.

These findings have profound implications for the "broadcast receiver" model of consciousness explored in the earlier series. If brains synchronize across skulls during interaction, the metaphor of "broadcasting and receiving" may be more literal than metaphorical.

The CaMBRAIN Project: Real-Time EEG Inference (2026)

Published on arXiv in May 2026, the CaMBRAIN (Causal state space Models for BRAIN) project represents a quantum leap in EEG technology [2]. Developed by researchers combining causal state space models with neural networks, CaMBRAIN achieves real-time, continuous EEG inference with unprecedented accuracy.

What makes CaMBRAIN revolutionary:

• Real-time processing — Previous EEG analysis required significant post-processing. CaMBRAIN processes and classifies brain states in milliseconds.
• Continuous inference — Rather than classifying short windows of data, CaMBRAIN tracks brain state as a continuous trajectory through state space.
• Causal modeling — The model captures causal relationships between brain regions, revealing not just which frequencies are present but how they interact.
• Generalizability — CaMBRAIN works across individuals without extensive calibration, addressing one of the major barriers to practical EEG.

The implications for frequency tuning are enormous. A system that can detect your current brain state in real time and trigger entrainment stimulation on the fly represents the holy grail of closed-loop brainwave modulation. CaMBRAIN makes this practical for the first time.

Consumer Neurofeedback Devices: What They Measure (and Miss)

The availability of consumer EEG devices has exploded in recent years. The major players:

| Device | Channels | Measures | Limitations | |--------|----------|----------|-------------| | Muse (InteraXon) | 4 (AF7, AF8, TP9, TP10) | Alpha, beta, theta, delta ratios | Limited spatial coverage; frontal-only | | Emotiv EPOC+ | 14 | Full band power, facial expressions, mental commands | Wet electrodes; setup time | | Neurosity Crown | 7 (dry electrodes) | Focus, flow state detection | Proprietary algorithms; developer-focused | | Versus | 2 (dry, ear-clip) | Mental load detection | Extremely limited; designed for cognitive load monitoring | | Muse S | 4 + PPG sensor | EEG + heart rate for sleep tracking | Accurate enough for sleep staging but not clinical |

Consumer devices are excellent for training and self-experimentation but have significant limitations. With only 2–14 electrodes (versus 64–256 in research), they miss most of the brain. They are also prone to artifact from eye blinks, jaw clenching, and head movement. A consumer device can tell you whether you are generally in alpha or beta, but it cannot localize activity to specific brain regions or reliably measure gamma.

Nevertheless, for the purpose of frequency tuning practice, consumer devices are sufficient. You do not need clinical-grade EEG to learn to shift between delta, theta, alpha, beta, and gamma states. The devices provide enough feedback to make the learning process dramatically faster than practicing blind.

MEG, fMRI, fNIRS: Alternative Brain Measurement Technologies

EEG is not the only way to measure brain activity. Each technology has strengths and weaknesses:

Magnetoencephalography (MEG) measures the magnetic fields produced by neural activity rather than the electrical fields. MEG offers better spatial resolution than EEG (can localize activity within millimeters), is not distorted by the skull, and does not require electrodes. However, MEG machines cost millions, require superconducting sensors cooled by liquid helium, and the patient must remain completely still.

Functional magnetic resonance imaging (fMRI) measures blood oxygenation changes (the BOLD signal) as a proxy for neural activity. It provides excellent spatial resolution (millimeters) but poor temporal resolution (seconds rather than milliseconds). fMRI cannot directly measure brainwave frequencies — it sees the metabolic consequences of neural firing, not the firing itself.

Functional near-infrared spectroscopy (fNIRS) uses light to measure blood oxygenation through the skull. It is portable, inexpensive, and relatively robust to movement. Like fMRI, it measures metabolic activity rather than electrical activity, so it cannot directly detect frequency bands.

For frequency tuning, EEG remains the gold standard. It is the only widely available technology that directly measures the oscillatory activity we care about — and at the millisecond timescale at which brainwaves operate.

The Limitations of Brainwave Measurement

It is important to be honest about what EEG can and cannot tell us:

What EEG can do:

• Detect the dominant frequency band at any moment
• Measure relative changes in band power across conditions
• Track coherence between brain regions
• Identify abnormal patterns (epileptiform activity, slowing in dementia)
• Provide biofeedback for training specific brain states

What EEG cannot do:

• Read thoughts or decode mental content with precision
• Measure activity in deep brain structures (thalamus, hippocampus, amygdala) — EEG is primarily cortical
• Determine the cause of a brain state (correlation ≠ causation)
• Work well in high-movement environments
• Provide definitive diagnosis without clinical context

The most important limitation for frequency tuning: EEG measures what your brain is doing, not what your brain is capable of. A person who shows low gamma in a resting EEG may be capable of high gamma in the right conditions. The measurement captures a snapshot, not the full range.

The Future: Wearable EEG and Closed-Loop Systems

The trajectory of EEG technology is clear: smaller, cheaper, more channels, better artifact rejection, integrated with AI.

• Dry electrodes have already replaced wet electrodes in most consumer devices, eliminating the messy gel application
• Wearable form factors are evolving toward invisible integration — behind-the-ear, in-ear, headband, and even glasses
• Closed-loop systems — the holy grail — detect your brain state and trigger stimulation (sound, light, vibration) in real time to entrain the target frequency
• AI integration — models like CaMBRAIN can separate signal from noise better than traditional filtering, enabling accurate measurement in real-world conditions

Within the next five years, a wearable device that tracks your frequency band throughout the day and provides targeted entrainment when you need it — gamma in the morning, alpha at noon, theta for creativity, delta at night — will be technologically feasible. The question is whether the market and the evidence base will support it.

Key Takeaways

• EEG measures the summed electrical activity of cortical neurons, extracting frequency bands through Fourier analysis and source localization through inverse modeling
• The international 10-20 system standardizes electrode placement, enabling reproducible measurements across individuals and laboratories
• Hans Berger recorded the first human EEG in 1924, establishing the foundation for a century of brainwave research
• Quantitative EEG (qEEG) adds statistical analysis and normative comparison, while source localization (LORETA) estimates where signals originate
• Hyperscanning reveals that brains synchronize during interaction — a finding with profound implications for the broadcast receiver model
• The CaMBRAIN project (2026) achieves real-time continuous EEG inference through causal state space models, making closed-loop frequency tuning practical for the first time
• Consumer neurofeedback devices (Muse, Emotiv, Neurosity) are sufficient for frequency training but limited in spatial coverage and gamma measurement
• EEG remains the gold standard for frequency measurement because it directly captures oscillatory activity at millisecond resolution
• The future is wearable, closed-loop, AI-integrated EEG — detectable brain states triggering real-time entrainment

References & Further Reading

1. Hasson et al. (2012) — "Brain-to-brain coupling: a mechanism for creating and sharing a social world" Trends in Cognitive Sciences — Foundational hyperscanning review.
2. CaMBRAIN (2026) — "Real-time, Continuous EEG Inference with Causal State Space Models" arXiv — The AI-powered real-time EEG inference project.
3. Berger (1929) — "Über das Elektrenkephalogramm des Menschen" Archiv für Psychiatrie und Nervenkrankheiten — The first human EEG paper.
4. Nunez & Srinivasan (2006) — Electric Fields of the Brain: The Neurophysics of EEG — The definitive textbook on EEG physics and methodology.
5. Buzsáki (2006) — Rhythms of the Brain — Neural oscillations and their measurement.
6. Dumas et al. (2010) — "Inter-brain synchronization during social interaction" PLOS ONE — Early hyperscanning demonstration.
7. Marzbani et al. (2016) — "Neurofeedback: A Comprehensive Review on System Design, Methodology and Clinical Applications" Basic and Clinical Neuroscience — Consumer neurofeedback technology review.
8. MIT News (Jan 2024) — "Study reveals a universal pattern of brain wave frequencies" — Universal brainwave scaling study.

Next in series: Entrainment — The Frequency Following Response

This article is Part 2 of the Brainwave Frequency Tuning series. View series overview →

Also explore: Brain as Broadcast Receiver Series — the theoretical companion to this practical series.

Back to Consciousness