Abstract
Abstract: :
Purpose: To optimize the information obtained from multifocal visual evoked potentials (mfVEP) and to evaluate the quality of the records obtained. Methods: Monocular mfVEPs were obtained from each eye of 30 control subjects and 49 patients with open-angle glaucoma using VERIS 4.3 (EDI). The patients had mild visual field defects (MD<-8 dB on 24-2 Humphrey fields). The display was a pattern-reversal dartboard array (VERIS), 44.5 deg. in dia., which contained 60 sectors. Recording electrodes were placed at the inion (I) and I+4 cm, and also at two lateral locations up 1 cm and over 4 cm from I. A signal-to-noise ratio (SNR) was determined for each response by dividing the root mean square (RMS) of a signal window by the mean RMS of a noise window[1]. Using the data from the control subjects, an analysis of false negatives (signal window) and false positives (noise window) was performed. Optimization and evaluation analyses were performed on the records. Optimization analyses: 1. SNRs were determined for low pass filtering from 100 to 10 Hz using a Fast Fourier Transformation and a sharp cutoff. 2. The size and location of the signal and noise windows were systematically varied and the RMS and SNR values determined. Evaluation analyses: An analysis of the frequency content in the noise windows provided a measure of the overall noise as well as the level of alpha contamination. The results for the patients were expressed as noise and alpha indices in z-scores. Results: Optimization analyses: The mean SNR reached a maximum for a low-pass of 20 Hz. However, a cutoff of 35 Hz provided minimal distortion of the waveforms and essentially identical detection of defects in terms of false positive and false negative rates. The optimal window for analysis was 105 ms in length and started at 45 ms for the signal window and at more than 300 ms for the noise window. Evaluation analyses: The alpha and noise indices were slightly higher for the patients’ records; 4% of the z-scores of both indices fell beyond a z-score of 1.96 (2.5% of normals). Conclusions: The records from VERIS must be analyzed offline to optimize the ability of the mfVEP to detect glaucomatous defects. Patients’ records should be evaluated to see if there is excessive alpha or noise as high levels of noise can lead to false positives[2]. 1. Zhang et al (2002) Doc Ophth 2. Hood et al (2003) J. Glau.
Keywords: visual fields • electrophysiology: clinical • clinical research methodology