Neural adaptation

Changes in responsiveness during sustained or repeated input. Adaptation makes activity depend on recent stimulation and can invalidate a model whose rate depends only on the present stimulus.

Changes in visual response associated with prevailing light input. Receptor gain, later neural gain and appearance reference points can adjust on different timescales; threshold loss is not automatically tissue damage.

Adjustment of visual responses to the spectral statistics of prevailing stimulation. Changes in gain or reference levels can partially stabilize color judgments despite different illumination or receptor signals. Adaptation is an explanatory mechanism to test; similarity on a color test does not uniquely identify its neural implementation.

A diagonal model applying gains $g_j$ to cone signals. If a change transforms every relevant catch as $Q_j\mapsto k_j{Q}_j$, gains $g_j\mapsto g_j/k_j$ preserve the normalized outputs exactly. A general illuminant change need not produce the same channelwise factors for every surface, so exact cancellation is conditional.

Changes following exposure to variation around a mean color or luminance. Holding the mean input fixed helps distinguish adaptation to variation from a simple account based on changed average cone catch, while receptor nonlinearities and task factors still require controls.