The three categories of gain calculation provided by the Interacoustics EyeSeeCam are:
There are in fact three instantaneous gain values provided for lateral impulses, all restricted to less than 100ms after the impulse commences. These gain values are the ratio of eye and head velocity. Values are provided at 40ms, 60ms and 80ms. Although emphasis is provided for the value at 60ms, the other values are provided because there is an indication that different (e.g. neurological) pathologies can lead to different gain values over the time course of the head impulse (Kremmyda et al., 2012).
In many (perhaps most) cases the instantaneous gain is both sufficient and ideal indication of VOR compensation for the head rotation. However, on occasions the goggle mounted camera can slip in relation to the pupil, causing “slippage” artefacts. Then the instantaneous gain can provide a misleading result.
In these cases we can turn to median gain and regression gain. Both of these values are also restricted to a time course of <100ms, which is important for providing an indication that is specific to VOR gain. These two gain calculations are averaged over the 100ms period, therefore should be less prone to artefactual anomalies present at any given moment in time (e.g. slippage, covert saccades, noise artefacts), although the dynamic information provided by instantaneous gain at multiple time points is lost. The regression gain and not the instantaneous gain would therefore be used routinely when performing RALP and LARP vHIT, since these procedures are more prone to google slippage.
Therefore, these various gain calculation methods are complimentary.
Kremmyda, O., Kirchner, H., Glasauer, S., Brandt, T., Jahn, K., & Strupp, M. (2012). False-positive head-impulse test in cerebellar ataxia. Frontiers in neurology, 3, 162. doi:10.3389/fneur.2012.00162