Abstract
Purpose:
To optimize the stimulus conditions for steady-state multifocal visual evoked potentials (ssmfVEP) and to compare the measurements with conventional mfVEP.
Methods:
All measurements used 4-channel VEP-recordings (best of 4 algorithms) with 58 field dartboard stimulation (Roland Consult, retiscan). M-sequence length (128, 255, 512 steps), the number of reversals (2, 3, 4, 5, 6, 7, 8), and temporal frequency (6, 7, 9, 10, 12, 15 Hz) were varied. Measurements were obtained from 9 normal subjects (mean age 39 years, 3 male, 6 female). All measurements were performed in 8 identical m-sequence cycles. FFTs were performed on each cycle and on averaged responses. Signal to noise ratio (SNR) was computed using the signal amplitude at the stimulus frequency and the noise defined as the averaged amplitudes at the neighbouring frequencies at which no response is elicited by the stimulus. All subjects additionally underwent conventional mfVEP measurements with a single pattern reversal per M-step with 300 ms signal length followed by a 200 ms noise window used for SNR calculation. Statistics: comparison of signal strength, SNR, and the correlation between local mfVEP as measured transiently and with steady-state stimulation.
Results:
SNR in the ssmfVEP was highest (3.1±1.7) at 9 Hz stimulation frequency and significantly (p<0.01, paired t-test) higher than SNR in conventional mfVEP (2.3±1.3). The time to perform one single ssmfVEP measurement at this frequency (4 reversals per M-step) was 123 s, compared to 141 s with conventional mfVEP. SNR and Amplitudes from ssmfVEP were significantly correlated with those from conventional mfVEP in 7 of the 9 subjects.
Conclusions:
The ssmfVEP may be useful for objective perimetry as spectrum analysis can be used for automated evaluation of responses leading to improved SNR. In comparison with conventional transient mfVEP, the recording time can be shorter because the total measurement period can be used for analysis and because no additional time is needed for noise estimation and SNR calculation.