Theories of induction propose that the brightness of a test patch within a complex surround is explained by local contrast or by integrating contrasts from various regions within the surround, weighted inversely with the distance from the test. Results here corroborate that brightness induction from a patterned background depends on both contiguous and non-contiguous surrounding light, but the measurements were inconsistent with any linear integration of contrast at edges within the scene. In some conditions, assimilation rather than contrast to contiguous surrounding light was observed, depending on the luminance of the light in non-contiguous regions. This finding implies that brightness induction from patterned backgrounds depends on neural processes that can cause contrast and/or assimilation, depending on the luminance relation between contiguous and non-contiguous regions. A model in the literature postulating that the influence of a non-contiguous edge is regulated by the amount of contrast at the contiguous edge can accommodate brightness induced by these patterned backgrounds.