Dijkstra, Pieterse and Pruyn (2008) found that individual differences in the effects of color could to a large extent be explained by "the ability to screen out irrelevant stimuli within the environment (Mehrabian, 1977a). Some people have a natural tendency to effectively reduce the complexity of an environment (high-screeners), where others are not capable of this information reduction (low-screeners)."
High-screeners appeared to be significantly less aroused/stressed by (both warm and cold) colors than low-screeners.
This finding could - tentatively of course - be generalized to people on the autistic spectrum if they have a significant tendency to be high- or low-screeners. My guess is they are on the whole very low-screeners, mainly due to low central coherence. Other common autistic weaknesses in the area of cognitive styles seem to point in the same direction.
Article Individual differences in reactions towards color in simulat...