Quantitative Sensory Biophysics · Retina

Tens of Picoamperes in the Dark: Why Eyes Burn Energy in Darkness

Unlike digital image sensors, retinal photoreceptors remain electrically and metabolically active in absolute darkness. When you turn off the lights, an unceasing inward stream of charge known as the dark current makes the cells consume oxygen and ATP. In tissue- and method-specific comparisons, retinal energy demand can exceed that of the cortex. Daylight acts as a metabolic brake that reduces this expenditure.

Human reference
20 pA
Peak photocurrent reported in direct human rod recordings; not a universal constant.
Light reduction
>75%
Reduction in total rod energy use in Okawa and colleagues' mammalian model.
Tissue ratio
2–3×
Retina-to-cortex ratio under the cited rabbit-retina conditions; context matters.

1. The Visual Energy Paradox

Everyday intuition suggests that our eyes rest when we close our eyelids or step into a dark room. Cellular biophysics shows that rods remain electrically and metabolically active when there is not a single photon in sight; each human retina contains approximately 120 million rods.

The dark current is on the order of tens of picoamperes, but its value depends on species and protocol. In 1979, Baylor, Lamb, and King-Wai Yau measured saturating responses up to 27 pA in toad (Bufo marinus) rods. In human rods, Kraft, Schneeweis, and Schnapf recorded flash-evoked photocurrents up to about 20 pA in 1993. This page uses 20 pA as an illustrative upper-bound reference, not as a universal human dark-current constant. The current is carried by sodium and calcium ions flowing through cyclic nucleotide-gated (CNG) cation channels in the outer segment membrane.

Due to this inward positive current, the resting membrane potential of the rod in darkness remains depolarized at approximately −40 millivolts, substantially higher than the typical −70 millivolt resting baseline of cortical neurons. At this depolarized voltage, photoreceptor ribbon synapses tonically release the neurotransmitter glutamate at maximum rates.

2. The Hefty ATP Price of Molecular Pumps

In an upper-bound model based on 20 picoamperes, each rod receives approximately 1.25 × 108 equivalent ionic charges per second. This converts current into elementary-charge units; it is not a count of sodium ions because the actual current also includes calcium.

To preserve ionic balance, the rod inner segment is packed with Na+/K+-ATPase molecular pumps. If the equivalent charge is treated as a didactic proxy and the 3-to-1 sodium-to-ATP stoichiometry is applied, the model estimates about 4.2 × 107 ATP molecules per second for one rod. This is an illustrative calculation, not a measurement of every human rod.

Across the approximately 120 million rods in one retina, the same model reaches about 5.0 × 1015 ATP molecules per second. For both eyes, the value is about 1.0 × 1016 molecules per second. Studies of rabbit and cat retina show high oxygen demand, but the 2–3× comparison with cortex depends on tissue, species, and method.

Phototransduction Workbench & Ion Current Oscilloscope

Adjust the incident photon flux to observe the single-photon enzymatic cascade: cGMP breakdown, closure of ion channels, collapse of the 20 pA reference-model current, and membrane hyperpolarization.

Incident Photon Flux 0 photons / s (Pitch Black)
Ion current
20.0 pA
Reference model; Na⁺/Ca²⁺ current
Membrane potential
−40.0 mV
Maximum depolarization
ATP burn per rod
4.2 × 10⁷ / s
Didactic pump proxy
Open cGMP channels
100%
Full permeability

3. The Photonic Cascade: Why Light Acts as a Brake

When a single quantum of light (one photon) is absorbed by a rhodopsin molecule in the outer segment discs, its 11-cis-retinal chromophore isomerizes into all-trans-retinal in less than 200 femtoseconds. This conformational switch sets off an immense enzymatic cascade:

In intact mouse rods, Yue and colleagues estimated approximately 12–14 effective transducin–PDE complexes for each activated rhodopsin; other preparations and models can yield different values. The active complexes stimulate PDE6, which hydrolyzes cGMP and closes CNG channels.

The localized plunge in cGMP concentration forces CNG channels to close. The dark current collapses and approaches zero under bright illumination. The rod hyperpolarizes to −70 millivolts, glutamate release halts, and the metabolic burden on the sodium pumps falls. In essence, light acts as an electrical breaker that reduces ATP expenditure.

Personal Retinal Bioenergetics Calculator

Estimate the equivalent ionic charge and ATP molecules used by the rods in both eyes during sleep. This is an explicitly labelled 20 pA upper-bound model, not a personal measurement.

Equivalent ionic charge pumped during sleep
8.63 × 10²⁰
20 pA model; both retinas (240 million rods)
ATP molecules consumed overnight
2.88 × 10²⁰ molecules
~0.24 g; 20 pA upper-bound model
Modelled reduction in light
>75%
Reduction observed in a mammalian rod model
Guided Body Experiment: The Scotopic Foveal Scotoma

You can directly verify the anatomical segregation between cones and rods on your own eyes in a dimly lit room or under a starry night sky:

  1. Fixate on a faint star or tiny distant light source: Look straight at the faint light using central vision. Notice that the faint point of light vanishes from perception.
  2. Avert your gaze by 10–15 degrees: Shift your gaze slightly to the left or right of the target. The faint light immediately reappears substantially brighter.
  3. The anatomical reason: Your central foveola measures 0.35 mm across and is packed exclusively with high-threshold cones and strictly 0 rods. At 10°–15° eccentricity, rod density reaches its peak of approximately 160,000 rods per square millimetre, powering dim-light scotopic vision.

Cellular Biophysics: Resting Energy Expenditure

Cell type Resting potential Dark / Resting state Relative O₂ burn Response to stimulus
Retinal rod −40 mV (depolarized) Open cGMP channels (20 pA model) High in cited studies Hyperpolarization (−70 mV, metabolic brake)
Cortical neuron −70 mV (polarized) Closed channels, basal pumping Moderate (~3.3 mL/100g/min) Depolarization (+30 mV, increased ATP burn)
Cardiac myocyte −85 mV (polarized) Basal rhythmic pumping High (~8.0 mL/100g/min) Depolarization and mechanical contraction
Retinal cone −40 mV (depolarized) Open channels, rapid recovery Very high Hyperpolarization with ultrafast kinetics

Methodological Note & Scientific Bibliography

The biophysical values presented come from experiments on toad rods, human rods, rabbit retina, and mouse rods, together with retinal oxygen and topography measurements. The quantitative model uses an explicit 20 pA upper-bound proxy rather than a universal human constant; oxygen ratios and cascade gain depend on species and method.