Photoparoxysmal response (PPR) is considered to be a risk factor for idiopathic generalised epilepsy (IGE) and it has a strong genetic basis. Two genome-wide linkage studies have been published before and they identified loci for PPR at 6p21, 7q32, 13q13, 13q31 and 16p13. Here we combine these studies, augmented with additional families, in a mega-analysis of 100 families. Non-parametric linkage analysis identified three suggestive peaks for photosensitivity, two of which are novel (5q35.3 and 8q21.13) and one has been found before (16p13.3). We found no evidence for linkage at four previously detected loci (6p21, 7q32, 13q13 and 13q31). Our results suggest that the different family data sets are not linked to a shared locus. Detailed analysis showed that the peak at 16p13 was mainly supported by a single subset of families, while the peaks at 5q35 and 8q21 had weak support from multiple subsets. Family studies clearly support the role of PPR as a risk factor for IGE. This mega-analysis shows that distinct loci seem to be linked to subsets of PPR-positive families that may differ in subtle clinical phenotypes or geographic origin. Further linkage studies of PPR should therefore include in-depth phenotyping to make appropriate subsets and increase genetic homogeneity.