on: How predictive are photosensitive epilepsy models as proof of principle trials for epilepsy?In this interesting paper, 1 the authors have successfully quantified the predictive capabilities of the single-blind, Phase IIa photosensitivity model. Most importantly, their conclusion stresses that potential AEDs showing suppression of the pathological generalized epileptogenic reaction to Intermittent Photic Stimulation (IPS) also show efficacy in phase III double-blind placebo-controlled AED trials of both partial and generalized epilepsies. This is not surprising since photosensitivity (the photoparoxysmal EEG response, PPR) can occur in all different types of epilepsies.We would, however, like to put two main issues that are addressed in the Discussion section into a broader perspective and give advice for future proof of principle (PoP) intrapatient photosensitivity studies based on what we have learned so far from all trials, including those referred to by Yuen and Sims. 1 1. Apparent discrepancy in results of CBZ in the photosensitivity model-dose and formulation impact the effect of CBZ Binnie et al. 2 found that CBZ produced a suppressive PPR (positive effect) in the photosensitivity model; yet recently, French et al. 3 did not. The difference in conclusion reached likely can be explained by: (a) the formulation of CBZ used (CBZ liquid solution 2 and tablets 3 ) and by (b) the total time of IPS EEG recording post-dose.CBZ tablets have a slow, variable absorption rate with peak serum concentrations at hour 15 post-dose 4 ; since CSF concentrations are linearly correlated with serum, peak CSF/brain CBZ concentrations would occur at 15 h post-single CBZ dose. 5 Binnie's 2 use of a liquid CBZ formulation ensured an earlier rise to, and greater peak CBZ concentration, 6 and enhanced pharmacodynamic effect, 7 explaining its positive effect within a one-day period.The latest adaptive, double-blind CBZ trial 3 was long, with four distinct treatment days, comparing replicate placebo days with days for single-dose CBZ 400 mg and with LEV 1000 mg. This trial burden, with recurrent IPS sessions over weeks-due to wash-out periods-was such that a choice was made to determine the effect of drug or placebo for only 6 h post-dose, instead of the usual testing over 32 h in Phase IIa AED development. Time is important pharmacodynamically; while an AED PPR effect can coincide with its' peak plasma concentration, this is not always the case. While valproate (VPA) is well-known to be clinically effective in patients with photosensitive epilepsy, Rowan found a continued, delayed effect of VPA in the model. 8 Additionally, a muted effect of intravenous VPA was seen in the model at 12 h post-continuous infusion, postulating that VPA is an AED that may require both appropriate concentration and time to exert its' optimal effect. 9