Applications
What Optogenetics Is Used For
The physics in Parts 1 to 3 is not decoration: every one of those steps sets a limit on what the technique can do. The colour the protein absorbs decides how deep you can reach, the channel's closing time decides how fast you can drive a cell, and the threshold current decides how much light you need. This page follows each constraint to the application it shapes.
1. Turning correlation into causation
This is the application that earned the prize. Before optogenetics, a neuroscientist could record which neurons fired during a behaviour and argue that they mattered. Electrodes could stimulate, but they excite everything nearby — axons passing through, several cell types at once — so the result was hard to attribute.
Genetics solves the targeting problem that physics cannot. A promoter expresses channelrhodopsin only in one cell type, so the light finds only those cells even though it floods the whole volume. That combination — genetic specificity, optical timing — is what Francis Crick had asked for and what the technique delivered.
Laboratories now use it routinely to test circuits for memory, fear, reward, feeding, sleep and movement: silence a projection and see whether the behaviour disappears, drive it and see whether the behaviour appears without the usual trigger.
2. The toolkit is a catalogue of physical parameters
Protein engineers vary exactly the quantities this course derived. Each row below is a knob from Parts 1–3 turned deliberately.
| What is changed | Which physics | What it buys |
|---|---|---|
| Absorption colour, 470 → 590–630 nm | The opsin shift of Part 1 | Deeper tissue reach, less scattering |
| Channel closing time, 10 ms → ~4 ms | The rate kclose of Part 3 | Spike trains above 100 Hz instead of ~40 Hz |
| Ion selectivity: cations → chloride | The reversal potential of the primer | Silencing instead of driving |
| Bistable “step-function” variants | A metastable state in the photocycle | One flash on, one flash off; far less light |
The third row is worth dwelling on. Channelrhodopsin reverses near 0 mV, so as Part 3 showed it can only ever depolarise. To silence a neuron you need a different reversal potential — a chloride-conducting channel or an ion pump — which is why inhibitory tools are separate proteins rather than a setting.
3. In the clinic: restoring a kind of sight
In retinitis pigmentosa the photoreceptors die while the retinal ganglion cells survive. If those surviving cells can be made light-sensitive, the eye regains a channel to the brain, even though the original detectors are gone.
In 2021 Sahel and colleagues reported partial recovery of visual function in a 58-year-old man blind from the disease for decades. An injected viral vector made his ganglion cells express ChrimsonR, and engineered goggles captured the scene and projected it onto the retina as pulses of amber light. He could locate, count and touch objects with the treated eye.
The choice of ChrimsonR is this course's physics making a clinical decision. Its absorption peak sits near 590 nm rather than 470 nm — the opsin shift of Part 1, engineered deliberately — because amber light penetrates better, constricts the pupil less, and is gentler on a retina that has to be illuminated for hours.
Sahel J.-A. et al., “Partial recovery of visual function in a blind patient after optogenetic therapy”, Nature Medicine 27, 1223–1229 (2021), doi:10.1038/s41591-021-01351-4. A single-patient case report, not a completed trial.
4. Why depth is the binding constraint
Part 3 found that a neuron needs of order 1 mW/mm² at its membrane. Tissue is not transparent, so the question every experiment faces is how far from the fibre tip that irradiance survives. Two effects work against you at once.
Geometric spreading. Light leaves the fibre into a cone whose half-angle in tissue follows from the numerical aperture, sin θ = NA/n. The beam area grows, so irradiance falls as 1/(r + z tan θ)² even in a perfectly clear medium.
Attenuation. Scattering and absorption remove light exponentially, with an effective penetration depth δ that is strongly wavelength-dependent: roughly half a millimetre in the blue, about twice that in the red-to-amber. Together:
\[ I(z) = \frac{P}{\pi r^2}\left(\frac{r}{r + z\tan\theta}\right)^{2} e^{-z/\delta} \]
Setting I(z) = 1 mW/mm² gives the reach. For a 200 µm fibre at 10 mW, blue light reaches about 1.3 mm and red about 2.0 mm — half again as deep, and more than three times the volume.
The exponential also explains why you cannot simply turn the laser up. Going from 1 mW to 10 mW — a factor of ten — buys only a factor of two in depth, and the discarded power becomes heat in the tissue. Reaching deeper is a materials problem, solved by shifting the protein's colour, not a power problem.
What this model leaves out. Real tissue scatters strongly rather than simply absorbing, so light does not travel in straight lines and δ lumps a great deal of physics into one number. Treat the output as an order of magnitude: enough to see why red-shifted opsins were worth engineering, not enough to design an implant.
5. Beyond neurons
Nothing in Part 3 is specific to a neuron. Any excitable cell with a membrane, a capacitance and a threshold responds the same way, which is why cardiac muscle became the second home of the technique: hearts expressing channelrhodopsin can be paced with light, and arrhythmias started and stopped optically in animal work.
The same logic reaches further still. Because the trigger is a photon and the output is a conductance, the protein can be dropped into any system where you want an event at a known time — which is also why it is such a clean testbed for the quantum-biology questions raised in Exercise 6.
What is still hard
Delivering the gene safely to the right cells in a human; reaching structures deeper than a few millimetres without implanting hardware; heating; immune responses to a protein borrowed from algae; and the gap between making cells fire and restoring a function the brain can actually interpret. The 2021 vision work is a case report of one patient, not a treatment.