Although many observations in patients with this intriguing type of epilepsy have been described and detailed studies have been performed, only a few meet the current criteria of class 1 or 2 evidence-based studies. In general, the selection bias is due to studying a referral population instead of the general population, and to different age and sex distributions of the subjects under study. Comparing the various studies is often difficult, because of differences in the populations studied (single seizures, epilepsy centre population, etc.), but also because of different methods (photic stimulator, flash frequencies, eye conditions, etc.) and the terminology used. Finally, and most crucial, in many studies there is often no information on how the data were actually obtained (EEG or clinical data or both?). The popular term "photosensitive" is used widely and applied to patients with a history of visually induced seizures, with and without a photoparoxysmal response (PPR), and to those with only a PPR. An overview of the "hard" data is given with future needs for a better understanding of this type of epilepsy and for improving the endophenotype for genetic research. It is important to standardise the studies as much as possible and describe in detail the methodology of the study, taking at least the above variables into account.