Improving the Reliability of the Flash Visual Evoked Potential P2 (FVEP-P2)

Unlike the pattern-reversal visual evoked potential, the flash visual evoked potential (FVEP) has had limited clinical use because of lower reproducibility and higher inter- and intra-individual variability. One potential explanation involves the time-locked nature of FVEP recordings, where activity such as unintentional suppression of eyes-closed alpha can lead to overlapping potentials that might obscure the FVEP-P2. The present investigation evaluated the reliability of the FVEP-P2 in young healthy controls across five sessions after applying narrow-band filtering (1-7 Hz) and additional notch filtering to remove alpha and beta activity. It was hypothesized that spectral entrainment would occur within the alpha and that correction for this entrainment would significantly improve the reliability of FVEP-P2. 27 younger, healthy participants aged 18-39 years (M = 24.48, SD = 5.26; 17 females) took part in the study, all reporting no history of photosensitivity or seizures, neurological disorders, or color deficiency. Each participant experienced five sessions of 100 strobe flashes with their eyes closed. The FVEP-P2 associated with each trial was identified using an automated algorithm. Narrow-band filtering (1-7 Hz) significantly improved FVEP-P2 latency reliability compared to the uncorrected condition. Notch filters targeting alpha and beta separately did not produce significant improvements, suggesting additive or overlapping contributions. FVEP-P2 amplitude reliability also improved with filtering, primarily because of attenuating alpha-band interference. Narrow-band and notch filtering mitigated alpha- and beta-band interference, enhancing the test-retest reliability of the FVEP-P2. These findings establish a clearer physiological basis for FVEP unreliability and demonstrate that spectral correction can elevate the FVEP-P2 to clinically acceptable reliability levels, supporting the continued clinical utility of the FVEP, particularly in patients for whom pattern-reversal stimuli are unsuitable (e.g., infants, comatose, or uncooperative individuals) and as a candidate biomarker in mild cognitive impairment and Alzheimer's disease (Arruda et al., 2020; Fix et al., 2014).