Brain, mind, cognition — a clinical tour of the most complex object known to science.
~86 billion neurons in the human brain[1], each with thousands of inputs. Specialized for one job: receive a signal, decide, transmit.
When inputs push the membrane past about −55 mV, voltage-gated Na⁺ channels snap open. The cell depolarizes, then K⁺ channels reset it. Repeat down the axon at up to ~100 m/s in myelinated fibres[1].
An action potential reaches the terminal, calcium floods in, vesicles dump neurotransmitter into a 20 nm gap. The next cell’s receptors decide what to do. The whole thing takes about half a millisecond.[1]
Roughly 100 trillion connections in an adult brain — an order of magnitude more than there are stars in our galaxy.
Glutamate (excite), GABA (inhibit), dopamine, serotonin, acetylcholine, norepinephrine, and many peptides.
Long-term potentiation: repeatedly co-active synapses get stronger[2]. Hebb’s rule — "neurons that fire together, wire together."
Time from arriving spike to postsynaptic response — the rate-limiting step in neural circuits.
Cortex is organized as topographic maps: neighbouring neurons code for neighbouring features. Discovered by carefully poking the cortex with electrodes — literally.
Hubel & Wiesel (1959) found neurons that fire only for edges at a specific orientation[1]. The retinal image is mapped point-for-point onto V1 in the occipital lobe.
A strip of cortex along the central sulcus where each patch controls a specific muscle group. Stimulate it, the body twitches.
Penfield’s wartime maps showed lips, hands, and tongue claim disproportionate cortical real estate — sensitivity, not size, sets the scale[2].
In congenitally blind people, "visual" cortex is recruited for Braille reading. Maps re-write themselves when the input changes.
In 1953, surgeons removed both hippocampi from a young man named Henry Molaison — "Patient H.M." — to stop his seizures. They worked. But H.M. could no longer form new long-term memories.[1]
H.M. founded modern memory research. We learned that declarative memory (facts, events) needs the hippocampus, but procedural memory (skills) does not — he could still learn new motor tasks without remembering having practiced them.
Two 19th-century clinicians, two patients, two distinct deficits — and the first localized account of any cognitive function in the brain.
Damage here produces Broca’s aphasia: speech is halting, ungrammatical, effortful — but comprehension is largely intact. The patient knows what they want to say.
Damage here produces Wernicke’s aphasia: fluent, grammatical, melodic speech — but comprehension is impaired and the words come out as nonsense. Patients are often unaware.
Modern fMRI shows language is more distributed than the classical model suggested[1] — but Broca’s and Wernicke’s patients are still where every neurology resident starts.
Camillo Golgi invented the silver-nitrate stain that, mysteriously, blackens only a few neurons at random — rendering each one in exquisite detail.[1]
Santiago Ramón y Cajal, a Spanish histologist, used Golgi’s stain to draw thousands of neurons by hand. He concluded that the brain is made of discrete cells, not a continuous web — the neuron doctrine.
They shared the 1906 Nobel Prize. They disagreed bitterly in their acceptance speeches. Cajal turned out to be right.[2]
Twenty years ago neuroscience meant electrodes and post-mortem slices. Today we can image, manipulate, and reconstruct circuits at multiple scales simultaneously.
Functional MRI tracks blood-oxygen changes (BOLD signal) at ~1 mm resolution — the workhorse of cognitive neuroscience since the 1990s.[1]
Channelrhodopsins from algae let researchers switch specific neurons on or off with millisecond pulses of blue light.[2] Causation, not just correlation.
GCaMP fluoresces when calcium rises during a spike. Two-photon microscopy can record thousands of neurons simultaneously in a behaving mouse.
Serial electron microscopy + deep learning reconstructs every synapse in a tissue volume. Full Drosophila brain mapped in 2024[3]; mouse cortex underway.
Despite a century of progress, the deepest questions remain genuinely open. These are the problems a young researcher could bet a career on.
Why does any physical process feel like something? The "hard problem"[1]. Leading frameworks: Global Workspace Theory, Integrated Information Theory — neither yet decisive.
Color, shape, motion, and location are processed in separate cortical regions. How are they fused into a single, unified percept of "red car moving left"?
REM sleep is universal among mammals and birds. Dreams may be memory replay, threat rehearsal, noise — or essential to learning. We still don’t know.
If memory is stored physically in the brain, where? Recent work (Tonegawa lab) tags and reactivates specific memory traces in mice[2] — but the full code is still cryptic.
Roughly one in three people will be affected by a neurological or psychiatric disorder in their lifetime[1]. The clinic is where neuroscience meets the world.
Amyloid-β plaques and tau tangles; first hippocampal, then cortical neurodegeneration. Lecanemab (2023) is the first disease-modifying therapy — modest but real.[2]
Loss of dopamine neurons in the substantia nigra. L-DOPA (1968) was a miracle drug; deep-brain stimulation now helps refractory patients.
SSRIs treat symptoms imperfectly. Ketamine and psilocybin show rapid antidepressant effects via NMDA / 5-HT2A receptors — reshaping the field.[3]
Lifetime prevalence around 1% globally. Heritable, dopaminergic and glutamatergic dysregulation, but cause and definitive biomarker still elusive.