Outline: Light/Dark Adaptation

Why have light adaptation?

Light level range of a factor of 10¹² - 10¹³
Pupil response range of a factor of 16
2 receptor systems (rods and cones)
Firing rate range of, say, 0-200 spikes/sec
Can't slow down the input response too much
Hence, move the dynamic range to match the requirements set by the input

Dark adaptation, basic data:

2 limbs of the curve = 2 mechanisms (photopic and scotopic), with different spectral sensitivities:
Log sensitivity is plotted so as to cover huge range
Shift of the peak (Purkinje shift) corresponds to peak rod vs cone sensitivity

Scotopic = rods, nonfoveal, color blind, slow temporal sensitivity, no Stiles-Crawford effect
Photopic = cones, foveal is best, discriminates colors, fast temporal sensitivity, shows Stiles-Crawford effect
All of these changes go with the two limbs of the curve
Note: You can discriminate colors if dark-adapted, just not at threshold:

Adaptation and photopigment bleaching

Doesn't explain the entire effect (see adaptation when little pigment has been bleached)
However, you can isolate one receptor by bleaching another
Adaptation is pooled over multiple receptors

Afterimages

Positive on dark (subthreshold) backgrounds (``dark light'')
Negative on bright backgrounds (hence adaptation is local)
Fades because is retinally stabilized
Returns with eyeblink
Positive & negative interchange as background is changed