Abstracts

Rationale: Despite the widespread clinical use of intermittent photic stimulation (IPS), EEG laboratories conduct the procedure with a high degree of variability. No study has identified the range of flash stimulus duration required to achieve a photoparoxysmal response (PPR). The aim of this study was to determine the range of IPS required to produce a PPR to aid in developing a standardized protocol. Methods: A retrospective analysis of the EEGs from the Division of Pediatric Neurology at Boston Medical Center between 2005 and 2009 yielding a PPR was performed. In 2007 the IPS protocol was changed from 5 second to 10 second flash trains, and data from before and after this protocol change were collected for comparison. All participants 30 years of age or younger were included. All EEGs were independently reviewed by two blinded epileptologists and data on PPR were collected, including time of onset, duration, distribution, morphology, presence of a driving or a photomyogenic response, and evidence of a clinical seizure. Discrepancies between the two reviewers were discarded. The morphology and distribution of the PPR were categorized using the Waltz 4 stage grading system (Waltz et al., Electroencephalogr Clin Neurophysiol 1992;83(2)138-45). Summary statistics were produced for time of PPR onset, and these were compared among the 4 grades of PPR. Results: Of 175 reports identified, 117 were excluded due to either absence of PPR, frequent background epileptiform activity, or presence of only a driving response. Among the remaining 58 studies, 245 frequency trains produced a PPR, 33% from the 5 second flash protocol and 67% from the 10 second protocol (table). Overall mean PPR onset was 2.33 seconds (range 0.08 to 9.54). Mean PPR onset was not significantly different among individual flash frequencies (p=.18). Mean PPR onset for the 5 second trains (1.44 seconds 1.50) was significantly different from the 10 second trains (2.67 seconds 2.51, p<.001). Within the 10 second protocol 30 of 165 PPRs (18%) began at least 5 seconds after IPS onset (figure). Of the 40 studies with 10 second flash trains, 7 (18%) would have been misclassified as negative if stimulation only lasted 5 seconds, including two Grade III (5.1 to 7.7 seconds) and five Grade IV (5.6 to 7.6 seconds). Conclusions: These findings support an IPS protocol of flash trains lasting at least 8 seconds, a duration which would have yielded positive findings in all studies. Although a small percentage of PPR began after 8 seconds, all of the studies in which this occurred had earlier PPR in other frequencies. The majority of the PPRs detected between 5 to 10 seconds were Grade IV, which are most closely associated with a diagnosis of epilepsy. Our study had several limitations, most importantly our protocol only included frequencies up to 25 Hz and could have missed PPR brought out by higher frequencies. In summary, the modification of existing protocols to include longer flash trains would likely increase the yield of IPS by 15 to 20%, and could improve the management of patients with epilepsy.