Abstract
Purpose: :
Trials using novel therapeutical approaches will rely on the inclusion of patients with highly impaired retinal functions. This is a challenge as standard methods to assess visual functions tend to fail. We present a method to selectively assess heavily reduced rod and cone functions using refined ERG methods.
Methods: :
Recordings were performed on an Espion e2 system with a color dome LED stimulator. To identify very small rod and cone functions, a 3 step approach was used: 1) For isolation of rod dominated ERGs, blue (470nm) or green (513nm) short flashes (6-14ms) at a repetition frequency of 9Hz on a dark background were used in dark adapted state; for cone function isolation red 33Hz flicker stimuli (635nm, 8-14ms) were applied on a blue background of 2 cd/m2. 75 individual sweeps were averaged. 2) Offline analysis: De-trending using subtraction of Lorentzian minimization fit of 3rd order polynomials and consecutive Fourier analysis (Singleton’s Mixed Radix FFT) were used to extract the signal. 3) Signal to noise assessment: AR modeling of individual recordings was used to asses noise level at the expected frequency band of the signal. Peak-based critical limit significance levels of the extracted signal in respect to underlying noise were calculated using Monte Carlo approximation.
Results: :
By recording V log-I responses in 32 eyes of healthy volunteers a blue stimulus of 0.012 scot.cd*s/m2 and 10ms duration was found to yield maximum rod-pathway specific responses. For mixed L-, and M-cone specific responses intensities of 0.12 cd*s/m2 with 10 ms duration were determined. For the rod response, dark adaptation recovery after bleaching was performed using the same light stimulus repeated every 2 minutes, and showed regular adaptation kinetics, indicating that the stimulus does not critically interfere with the dark adapted state and addresses mainly rod function. In 128 of 207 recordings of patients with undetectable standard ISCEV rod response (visual acuity between light perception and 20/20) a 9Hz rod flicker response exceeding the 95% confidence limit against noise, could be measured. Responses of less than 0.5µV could be measured reliably.
Conclusions: :
This method allows for the detection of very small remaining retinal ERG responses with respective confidence limits even if standard clinical measures like visual field, visual acuity and ISCEV standard ERG protocols fail. This method may be a valuable tool in clinical trials where legally blind persons will be included.
Keywords: clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology • retinal degenerations: hereditary • electroretinography: clinical