There is some considerable uncertainty as the exact mechanisms. However, the recent seismicity and pyroclastic deposits detected around the ridge provide a unique opportunity to observe and model the baheviour of ultraslow spreading ridges.
According to the article Explosive volcanism on the ultraslow-spreading
Gakkel ridge, Arctic Ocean (Sohn et al. 2008), to overcome the pressure of the seawater, it becomes a question of the amount of volatiles (particularly $\ce{CO2}$ available in the magma chamber to cause fragmentation resulting in the glassy pyroclastic deposits observed around the volcanic centres on the Gakkel Ridge.
Sohn et al. hypothesise that the ultraslow spreading nature of the Gakkel Ridge would cause long intervals between volcanic events, which in the article, Effusive and explosive volcanism on the ultraslow-spreading Gakkel Ridge, 85°E (Pontbriand et al. 2012) suggest to be around 10,000 years.
The long interval times allow for the accumulation of large volumes of volatiles in the upper part of the magma chambers. These would erupt periodically when the pressure in the magma chamber exceeds the deep sea-water pressure. Models from Sohn et al. suggest that a weight fraction of 14% $\ce{CO2}$ (~75% volume fraction) would be required - a value considerably higher than those measured at most mid ocean ridges. Pontbriand et al. estimate that $\ce{CO2}$ supersaturation in the magma chambers of the Gakkel Ridge to occur around 1500-2000 year timeframes (i.e. well within the estimated volcanism interval).
Pontbriand et al. go further and suggest that one of the most important implications of the observations (and modelling) of explosive volcanism on deep-sea ultraslow spreading ridges challenge the hypothesis that eruption volume varies inversely with mid ocean ridge spreading rates.
