To study the effects of filtering bandwidth on the two-global-flash multifocal electroretinogram (mfERG) responses in primary open-angle glaucoma (POAG) compared with control subjects. A two-global-flash mfERG (VERIS 6.06™, FMS III) was recorded in 20 healthy subjects and 22 POAG patients with a band-pass filter (BPF) of 1-300 Hz (103 Hexagons, M-sequence stimulus: Lmax 100 cd/m(2), Lmin < 1 cd/m(2), global flash: 200 cd/m(2)). The root-mean-square average of the central 10° was calculated. Three response epochs were analysed: the response to the focal flash, at 15-45 ms (DC), and the following two components induced by the effects of the preceding focal flash on the response to the global flashes at 45-75 ms (IC1) and at 75-105 ms (IC2). The following BPF settings were analysed: 1-300 Hz, 3-300 Hz, 10-300 Hz, 100-300 Hz, 200-300 Hz, 1-10 Hz, 1-100 Hz and 1-200 Hz. Filtering at 1-300 Hz showed significantly lower responses in POAG than in control subjects (p < 0.001) for all epochs analysed. At 1-100 Hz, this also held true even though the difference between the groups became smaller. At 1-10 Hz, responses were extremely small and did not differ between POAG and control (p > 0.5). This would suggest a filter setting of 10-300 Hz for mfERG recordings in POAG. However, when a filter setting of 10-300 Hz was compared to 1-300 Hz, with a filter setting of 10-300 Hz, the DC in POAG differed more (p < 0.0001) from normal than with 1-300 Hz (p = 0.0002). For IC1 and IC2, the stronger difference between POAG and control was found with 1-300 Hz (p < 0.0001) rather than with 10-300 Hz (p < 0.0001 and p = 0.0005, respectively). For the 'oscillatory potentials' at 100-300 Hz, POAG and control differed significantly in IC1 and IC2 (p < 0.05), but not in DC (p = 0.8). However, filtering at 200-300 Hz did not show a difference between POAG and control (p > 0.5). Thus, we applied a filter setting of 1-200 Hz, which seemed to be most sensitive in detecting glaucomatous retinal dysfunction (p < 0.0001). A filter setting of 1-200 Hz appears most sensitive to detect glaucomatous damage if using a two-global-flash mfERG: using a band-pass filter a with lower low-frequency cut-off, containing the 10 Hz component, may be especially important in the small induced components that show glaucomatous damage most sensitively. High frequencies of 100-300 Hz also contain information that differentiates glaucoma from normal and thus should be included in the analysis.