Purchase this article with an account.
Brian Kunzer, Zahra Derafshi, Hadi Tajalli, John R Hetling; Corneal ERG Topography: Visualizing Multi-Electrode Electroretinogram (meERG) Data Using a Three-Dimensional Surface Spline. Invest. Ophthalmol. Vis. Sci. 2014;55(13):5120.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The impact of meERG recording lies in the ability to observe and interpret spatial differences in corneal potentials (ERG topography) following a light stimulus. An initial step is to interpolate between the meERG measurement locations to form a smooth topographic map, analogous to high-density EEG mapping. A MATLAB-based tool was developed that performs the interpolation at each time step in the ERG response. The visualization tool is used to differentiate healthy-eye responses from those obtained from eyes with retinal lesions. Robustness of the spline-interpolated maps was evaluated by calculating error introduced by missing values (due to noisy or inoperable electrode channels) in the meERG data set.
meERG data sets were obtained from Long Evans rat eyes using a contact lens electrode array. Photocoagulation lesions were created adjacent to the optic disk but restricted to one hemisphere. A three-dimensional spline interpolation approach was adapted to the meERG data structure (25 measurement locations) to create a smooth corneal potential map. Potential maps were converted to a dimensionless quantity, standard deviations from the spatial mean, in order to facilitate pooling and comparison of responses from different experiments. Robustness of the interpolation was evaluated by calculating an average percent difference between measured values and interpolated values when individual channels, or groups of channels, were not included in interpolation calculations.
The interpolation approach provides the first ERG topographic maps derived from meERG data sets. Topographic maps of eyes with retinal lesions are visually and quantitatively distinguishable from healthy eye responses. Interpolation based on incomplete data sets (< 25 channels) results in mean error levels of less than 10% until ten or more channel values are removed, though maximum error rate is sensitive to the position of missing channels.
The spline interpolation approach was successfully adapted to the meERG data structure, and proved robust in the case of moderate numbers of missing channels. The resulting corneal ERG topographic color maps can present absolute amplitudes or relative spatial differences, and can be used to visually and quantitatively compare results from healthy and unhealthy eyes.
This PDF is available to Subscribers Only