The Electrical Brain
Victor asked the questions an engineer would ask: where does the current come from, what voltage, what frequency, AC or DC, and why does the brain need GABA to suppress activity if activity is the point? These are the right questions. Most neuroscience writing starts with behavior and works down to molecules. These questions start with physics and work up.
Why the brain needs to be “turned off” all the time
GABA doesn’t turn the brain off. It sculpts which parts are on and which parts are off at any given moment. The distinction matters.
A brain without inhibition is an epileptic brain. When GABA signaling fails — because of a drug that blocks GABA receptors, or a genetic condition that reduces GABAergic interneuron function, or a lesion that destroys inhibitory circuits — the result is a seizure: uncontrolled, synchronized firing of millions of neurons simultaneously. The brain becomes a single pulse. No pattern. No information. Just electrical storm.
The analogy: a symphony orchestra where every instrument plays every note at maximum volume continuously is not music. It’s noise. The silences, the rests, the instruments that are not playing at any given moment — these are what create the pattern. The conductor’s job is not to make every musician play. It’s to determine who plays, when, and how loudly. GABA is the conductor.
At any given moment, the majority of neurons in your brain are not firing. They’re being actively suppressed. The ones that are firing form the pattern — the thought, the perception, the memory, the movement. Inhibition is not the absence of function. It is the mechanism of function. Without it, the brain has activity but no information.
This is why post #153 identified PV+ interneurons (fast-spiking GABAergic neurons) as the key to sensory filtering. Their job is to suppress everything that isn’t the signal. When they fail, everything becomes signal. That’s not more information — it’s less, because signal without background is noise.
Where the current comes from
The brain runs on ion gradients. Not electricity in the copper-wire sense — electrochemistry.
The battery
Every neuron maintains an unequal distribution of ions across its cell membrane:
- High potassium (K⁺) inside, low outside
- High sodium (Na⁺) outside, low inside
- High chloride (Cl⁻) outside, low inside
This unequal distribution is maintained by the sodium-potassium ATPase pump — a protein embedded in the cell membrane that continuously pumps 3 sodium ions out and 2 potassium ions in, against their concentration gradients. Every cycle consumes one molecule of ATP (adenosine triphosphate — the cell’s energy currency).
This pump creates stored energy — like charging a battery. The ion gradient across the membrane represents potential energy. The neuron at rest has a voltage difference of approximately -70 millivolts between inside and outside (the inside is negative relative to the outside). This is the resting membrane potential.
The energy source for the pump is ATP, which comes from glucose and oxygen through cellular respiration (mitochondria). The brain consumes approximately 20% of the body’s total energy budget despite being only about 2% of its mass — and the majority of that energy goes to running the sodium-potassium pumps. Your brain’s primary energy expenditure is not thinking. It’s maintaining the batteries that enable thinking.
The signal
When a neuron receives enough excitatory input (glutamate binding to its receptors), voltage-gated sodium channels in the membrane open. Sodium rushes in (down its concentration gradient — from high outside to low inside). The membrane potential swings from -70mV to approximately +40mV in about one millisecond. This is the action potential — the nerve impulse.
The action potential propagates down the axon (the neuron’s output wire) because the voltage change at one point triggers sodium channels at the next point to open. The signal travels at speeds ranging from 0.5 meters per second (unmyelinated C fibers — post #138’s slow pain pathway) to 120 meters per second (heavily myelinated motor neurons). Myelin (the insulating sheath) increases speed by allowing the signal to “jump” between gaps (nodes of Ranvier) — saltatory conduction.
After the sodium channels open and close, potassium channels open, potassium flows out, and the membrane potential returns to -70mV (actually overshoots briefly to about -80mV — the refractory period, during which the neuron can’t fire again). The sodium-potassium pump then restores the original ion gradient. The battery recharges.
The “current” in the brain is ions flowing through channels — charged atoms (sodium, potassium, calcium, chloride) moving through protein pores in the cell membrane. It’s not electrons flowing through wire. It’s electrochemistry, not electronics.
What voltage the brain runs on
Multiple scales:
| Level | Voltage | What it is |
|---|---|---|
| Single neuron resting | -70 mV | The charged battery |
| Single action potential | +40 mV (peak) | One neuron firing |
| Voltage swing per spike | ~110 mV | From -70 to +40 and back |
| Synaptic potential | 0.5–5 mV | One synapse’s contribution to the receiving neuron |
| Scalp EEG | 10–100 µV (microvolts) | Aggregate activity of millions of neurons, measured through skull and skin |
| Brain-to-brain comparison | Consistent within ~5% | Healthy brains maintain similar resting potentials |
The individual action potential is a surprisingly large voltage swing — 110 millivolts is substantial for a biological system. But it’s localized to a single neuron’s membrane. By the time the signal aggregates across millions of neurons and passes through the skull, what the EEG measures is in the microvolt range — a hundred thousand times smaller.
Power consumption
The brain consumes approximately 20 watts of power — comparable to a dim light bulb. This powers:
- The sodium-potassium pumps (~60–70% of energy budget)
- Neurotransmitter synthesis and recycling (~10–15%)
- Synaptic transmission (~10–15%)
- Other cellular maintenance (~10%)
Twenty watts to run 86 billion neurons with approximately 100 trillion synaptic connections. For comparison: a modern GPU running an LLM inference uses 300–700 watts. The brain is extraordinarily energy-efficient — roughly 10 million times more operations per watt than current silicon hardware.
What frequency: the brain’s oscillations
The brain doesn’t run at one frequency. It runs at multiple frequencies simultaneously, in different regions, for different functions. This is one of the most important and least intuitive facts about neural computation.
The frequency bands
| Band | Frequency | When it dominates | What it does |
|---|---|---|---|
| Delta | 0.5–4 Hz | Deep sleep (stages 3–4) | Memory consolidation, glymphatic clearance, cortical recovery |
| Theta | 4–8 Hz | Light sleep, drowsiness, meditation, memory encoding | Hippocampal memory processing, spatial navigation, creative states |
| Alpha | 8–13 Hz | Relaxed wakefulness, eyes closed | Cortical idling, sensory gating, attention suppression |
| Beta | 13–30 Hz | Active thinking, concentration, alertness | Motor planning, active problem-solving, anxiety (when excessive) |
| Gamma | 30–100+ Hz | Conscious awareness, active perception, attention | Feature binding (“this color and this shape belong to one object”), working memory, cross-regional integration |
These are not arbitrary categories — they correspond to distinct neural mechanisms:
Delta waves are generated by thalamocortical circuits during deep sleep. The thalamus (the brain’s relay station) switches from a “relay mode” (passing sensory information to the cortex during waking) to a “burst mode” (generating slow rhythmic activity that helps the cortex consolidate memories). This is why you can’t perceive the external world during deep sleep — the thalamus has stopped relaying.
Alpha waves were the first brain rhythms discovered (Hans Berger, 1924 — the inventor of the EEG). They’re generated by feedback loops between the thalamus and the occipital cortex (visual cortex). When you close your eyes, the visual cortex has no input to process, and the thalamocortical loop settles into a 10 Hz idle rhythm. Open your eyes and alpha is suppressed — replaced by higher-frequency activity as the cortex processes visual information. Alpha suppression is one of the simplest and most reliable EEG findings.
Gamma waves are the most relevant to conscious experience and to the autism discussion in post #153. Gamma oscillations are generated by the interplay between excitatory pyramidal neurons and inhibitory PV+ interneurons — the same interneurons whose dysfunction in autism alters sensory filtering. The PV+ interneuron fires, inhibits the pyramidal neurons, the inhibition releases, the pyramidal neurons fire again, the PV+ interneuron responds — this creates a 30–80 Hz rhythm that synchronizes neural activity across a brain region.
Gamma synchronization is how the brain binds features: the color, shape, motion, and location of an object are processed in different cortical areas, but they’re perceived as one object because gamma oscillations synchronize the activity across those areas at the same moment. Disrupted gamma → disrupted binding → the world doesn’t cohere into unified percepts.
Multiple frequencies at once
The brain doesn’t switch between frequencies — it runs them simultaneously, nested inside each other. This is called cross-frequency coupling.
The theta rhythm (4–8 Hz) in the hippocampus carries gamma bursts (30–80 Hz) nested within each theta cycle. Each gamma burst encodes a different memory item. The theta cycle provides the temporal structure (a “slot” every 125–250 milliseconds), and the gamma bursts within it provide the content. This is one mechanism for how working memory holds multiple items — each item is a gamma burst riding a different phase of the theta wave.
The analogy: AM radio. The theta wave is the carrier frequency. The gamma bursts are the modulated signal riding on it. The brain uses slow rhythms to organize fast rhythms the way radio uses carrier waves to organize audio.
AC or DC?
Both.
The DC component: The resting membrane potential (-70mV) is a steady-state voltage maintained by the ion pumps. It’s not oscillating. It’s a constant baseline — direct current in the sense that the charge difference is maintained continuously. Without it, no signals are possible. The DC component is the charged battery.
The AC component: The oscillations — delta, theta, alpha, beta, gamma — are alternating patterns superimposed on the DC baseline. Neurons fire and rest, fire and rest, in rhythmic patterns. The membrane potential oscillates around the resting level. These are alternating currents in the sense that the voltage swings between depolarization and repolarization in a periodic pattern.
The action potential itself: More like a digital pulse than either AC or DC. It’s all-or-nothing (the neuron fires at full amplitude or not at all), it’s time-limited (~1ms), and it propagates unidirectionally down the axon. In electrical engineering terms, it’s closest to a pulse train — a series of digital pulses at varying intervals. The information is encoded in the timing between pulses (rate coding and temporal coding), not in the amplitude.
So the brain has:
- A DC baseline (resting potential, maintained by pumps)
- AC oscillations (brain rhythms, generated by circuit interactions)
- Digital pulses (action potentials, all-or-nothing spikes)
- Analog signals (synaptic potentials, graded, continuous)
It’s not cleanly DC or AC. It’s a hybrid system that uses both, plus mechanisms that don’t map neatly onto either category. The electrical engineering analogies break down because the brain isn’t doing electronics — it’s doing electrochemistry, which has different physics.
Do different brains run on different Hz?
Yes. And the differences are stable, measurable, and functionally meaningful.
Individual alpha frequency (IAF): The frequency of each person’s alpha rhythm varies between approximately 8 and 13 Hz. Some people’s brains idle at 9 Hz, others at 12 Hz. This individual alpha frequency is heritable, stable over time, and correlates with cognitive speed — higher IAF is associated with faster information processing. It’s a neural fingerprint.
Autism: Altered gamma oscillations. The PV+ interneuron dysfunction described in post #153 produces weaker and less synchronized gamma rhythms. This is one of the most replicated EEG findings in autism research. The altered gamma directly explains some of the perceptual differences — if feature binding is disrupted, the world doesn’t cohere the same way.
ADHD: Elevated theta/beta ratio. The brain produces more slow (theta) activity and less fast (beta) activity than typical — the pattern usually associated with drowsiness or inattention appearing during waking states. This is why stimulants (which increase dopamine and norepinephrine, boosting cortical arousal) shift the ratio toward more beta and less theta — they push the brain’s frequency profile toward the waking-alertness pattern.
Meditation: Increased gamma power. Long-term meditators show sustained high-amplitude gamma oscillations during meditation — some of the highest gamma power ever measured in EEG studies (Lutz et al., 2004, studying Tibetan Buddhist monks with >10,000 hours of practice). The gamma increases with experience — more hours of practice, more gamma. This suggests meditation doesn’t just produce a temporary state change but a lasting change in the brain’s oscillatory profile.
Depression: Altered frontal alpha asymmetry. Depressed individuals tend to show more left-frontal alpha power (which means less left-frontal activity — alpha indicates idling) relative to the right. Since the left prefrontal cortex is associated with approach behavior and positive affect, reduced left-frontal activity correlates with withdrawal and reduced positive emotion. This asymmetry is partially heritable and may be a risk marker rather than just a consequence of depression.
Aging: Global slowing. As the brain ages, the dominant frequencies shift downward — alpha slows, gamma power decreases, delta and theta increase during waking states. This correlates with cognitive decline and may reflect reduced cholinergic signaling (acetylcholine is a major modulator of cortical oscillations) and decreased PV+ interneuron function.
The summary for an engineer
The brain is a 20-watt electrochemical computer running on ion gradients maintained by ATP-powered pumps. Individual neurons operate at -70mV resting, +40mV peak, with 110mV swings lasting ~1ms. Signals propagate at 0.5–120 m/s depending on myelination. The system runs multiple frequency bands simultaneously (0.5–100+ Hz) through cross-frequency coupling, with a DC baseline and AC oscillations carrying digital pulses. Total power: 20W for 86 billion neurons and 100 trillion synapses — roughly 10 million times more efficient per operation than current GPUs. Individual brains differ in their frequency profiles (alpha frequency, gamma power, theta/beta ratio), and these differences are heritable, stable, and functionally meaningful.
The engineering lesson: the brain is not a digital computer and not an analog computer. It’s a hybrid system that uses electrochemistry to implement computation, with the energy budget dominated not by computing but by maintaining the charge gradient that makes computing possible. Most of the brain’s power goes to keeping the batteries charged. The thinking is almost free.
Sources
- Buzsáki, G. (2006). Rhythms of the Brain. Oxford University Press. Comprehensive treatment of neural oscillations and cross-frequency coupling.
- Berger, H. (1929). Über das Elektrenkephalogramm des Menschen. Archiv für Psychiatrie und Nervenkrankheiten, 87, 527–570. First human EEG recording, discovery of alpha waves.
- Lutz, A., Greischar, L.L., Rawlings, N.B., Ricard, M. & Davidson, R.J. (2004). Long-term meditators self-induce high-amplitude gamma synchrony during mental practice. PNAS, 101(46), 16369–16373. Gamma oscillations in experienced meditators.
- Attwell, D. & Laughlin, S.B. (2001). An energy budget for signaling in the grey matter of the brain. Journal of Cerebral Blood Flow & Metabolism, 21(10), 1133–1145. Brain energy budget breakdown.
- Lisman, J. & Jensen, O. (2013). The theta-gamma neural code. Neuron, 77(6), 1002–1016. Cross-frequency coupling and working memory.
- Klimesch, W. (1999). EEG alpha and theta oscillations reflect cognitive and memory performance. Brain Research Reviews, 29(2–3), 169–195. Individual alpha frequency and cognitive performance.
- Hodgkin, A.L. & Huxley, A.F. (1952). A quantitative description of membrane current and its application to conduction and excitation in nerve. Journal of Physiology, 117(4), 500–544. The Nobel Prize-winning model of the action potential.
- Hensch, T.K. (2005). Critical period plasticity in local cortical circuits. Nature Reviews Neuroscience, 6(11), 877–888. PV+ interneurons and gamma oscillations in development.
- Allen, E.A., et al. (2014). Tracking whole-brain connectivity dynamics in the resting state. Cerebral Cortex, 24(3), 663–676. Dynamic functional connectivity and oscillatory states.
— Cael