Abstract
Purpose: :
Infants with PMG without involvement of the occipital lobe show detectable VEP response but fail to show any visual orienting behavior. The VEP waveform is abnormally distorted and prolonged to several stimuli. We hypothesize that averaging of the VEP during periods of poor fixation contribute to the abnormal responses. This study examined the transient VEP in the Fourier domain with subsequent averaging of epochs with consistent amplitude and phase across trials.
Methods: :
Subjects were 5 infants with PMG (0.4 - 1.6 yrs age) and 5 age matched controls. PMG was documented by MRI (coronal, sagittal and axial T1, T1 inversion recovery, T2, FLAIR, coronal T2, diffusion, gradient echo sequences). Transient VEPs were recorded under binocular viewing to 163’ reversing checks and onset-offset of 0.5 cycle/degree gratings. Amplitude and latency was measured for the P1 and N2 peaks. Individual epochs were plotted in amplitude/phase plot at 1 - 13 low frequency harmonics. Clusters of data points at set magnitude/phase relationships (i.e., forming a circular cloud) of could be selectively averaged.
Results: :
Both controls and infants with PMG had distinct clusters that followed systematic phase shifts with increasing harmonics (however scatter was greater in infants with PMG). Clusters could still be detected in both infants with PMG and no visual behaviors. Selecting epochs that made a cluster tighter for one harmonic had the same effect on the remaining harmonics. Compared to standard artifact rejection our analysis consistently increased P1 and N2 amplitudes without any effect on latency. For controls, P1 and N2 amplitude increased by 21% (range 4 - 55%) and 65% (range 11- 311%), respectively. Latency differences were 0.7 and 7.3 msec, respectively. For infants with PMG our analysis had a greater increase in P1 and N2 amplitude (P1 increase = 240%, range 1 - 580%; N2 increase = 337%, range 1-1000%). Latency differences between analyses were -1.20 and 8.8 msec, respectively. Waveforms always appeared similar between analyses but with larger amplitude showing that consistency in magnitude/phase were related to the final averaged VEP.
Conclusions: :
VEP amplitude can be improved in Fourier domain by selective averaging of epochs with consistent amplitude and phase across trials. Infants with PMG and no visual behaviors show abnormal VEPs that can be improved in amplitude but not in waveform distortion or latency. The presence of variance in the VEP could result from poor fixation to the stimulus, increased variation in feedback mechanisms from association cortex and dysfunctional striate cortex not observed on MRI.
Keywords: visual development: infancy and childhood • electrophysiology: clinical • neuro-ophthalmology: cortical function/rehabilitation