Seasonal affective disorder double peak

Could seasonal affective disorder be caused not just by length of day (and hence the amount of light received), but also rate of change in length of day. This would generate a double peak in the systems on an annual graph, one at the winter solstice, one at the autumn equinox. If this was the case, the extent to which these two causal variables stimulate the effect would no doubt vary from person to person. And understanding this would be helpful in treatment.

The rate of change in length of day stimulus would be a logical one to have evolved, as it would stimulate earlier preparations for winter than a short day stimulus.