Effect of varying skin surface electrode position on electroretinogram responses recorded using a handheld stimulating and recording system

Abstract

Purpose: A handheld device (the RETeval system, LKC Technologies) aims to increase the ease of electroretinogram (ERG) recording by using specially designed skin electrodes, rather than corneal electrodes. We explored effects of electrode position on response parameters recorded using this device.

Methods: Healthy adult twins were recruited from the TwinsUK cohort and underwent recording of light-adapted flicker ERGs (corresponding to international standard stimuli). In Group 1, skin electrodes were placed in a "comfortable" position, which was up to 20 mm below the lid margin. For subsequent participants (Group 2), the electrode was positioned 2 mm from the lid margin as recommended by the manufacturer. Amplitudes and peak times (averaged from both eyes) were compared between groups after age-matching and inclusion of only one twin per pair. Light-adapted flicker and flash ERGs were recorded for an additional 10 healthy subjects in two consecutive recording sessions: in the test eye, electrode position was varied from 2 to 10-20 mm below the lid margin between sessions; in the fellow (control) eye, the electrode was 2 mm below the lid margin throughout. Amplitudes and peak times (test eye normalised to control eye) were compared for the two sessions.

Results: Including one twin per pair, and age-matching yielded 28 individuals per group. Flicker ERG amplitudes were significantly lower for Group 1 than Group 2 participants (p = 0.0024). However, mean peak times did not differ between groups (p = 0.54). For the subjects in whom electrode position was changed between recording sessions, flash and flicker amplitudes were significantly lower when positioned further from the lid margin (p < 0.005), but peak times were similar (p > 0.5).

Conclusions: Moving the skin electrodes further from the lid margin significantly reduces response amplitudes, highlighting the importance of consistent electrode positioning. However, this does not significantly affect peak times. Thus, it may be feasible to adopt a more comfortable position in participants who cannot tolerate the recommended position if analysis is restricted to peak time parameters.

Keywords: Electrode; Electroretinogram; Retina; Retinal function.

Fig. 1

Two different electrode positions in a study participant. The skin sensor electrodes are placed 2 mm below the lid margin (manufacturer’s recommended position) in the upper photograph, and in an inferior position 10–20 mm below lid margin (lower photograph). Permission granted for use of photograph

Fig. 2

Boxplots comparing the distribution of ERG parameters between participants in Groups 1 and 2. In Group 1, the electrode was in a comfortable position (up to 20 mm from the lid margin); in Group 2, the electrode was placed in the recommended position of 2 mm below the lid margin. Median, upper and lower quartiles (limits of box) are shown. Squares plot means; whiskers span the 5th–95th centiles; and crosses plot minimum and maximum data points. Amplitudes are plotted in a and peak times in b

Fig. 3

Sample traces from one participant. a, b Averaged light-adapted ERGs to a flash obtained from right and left eyes, respectively. In the right eye, the skin electrode was placed 2 mm from the lid margin. In the left eye, the electrode was 15 mm below the lid margin. c, d Light-adapted responses to the flicker stimulus obtained with electrodes placed as in a, b, respectively. Green and orange traces represent averaged recordings obtained from consecutive runs of stimulus presentations

Fig. 4

ERG parameters for different electrode positions in participants in the third group (n = 10; data normalised to control eye). a, b, c Boxplots showing a-wave, b-wave and flicker ERG amplitudes: differences were significant between the two positions (p = 0.004, 0.002 and 0.002, respectively, Wilcoxon signed rank test). d, e, f, Boxplots showing a-wave, b-wave and flicker ERG peak times: differences were not significant (p = 0.85, 0.57 and 0.56, respectively). Boxes show median and upper and lower quartiles; squares plot mean values; whiskers extend to maximum and minimum values. Note that the y-axis scales differ in the lower panels and have been expanded considerably in F, where the actual range between maximum and minimum values is less than 0.1