Abstract
Abstract: :
Purpose: To determine if multifocal visual–evoked potential (mfVEP) can detect hemianopic visual field defects and if the signal pattern can be used to predict the location of the patient's lesion. Methods: Fifteen patients with hemianopic visual field defects and 46 control subjects were included if they had conventional automated perimetry (CAP) showing a homonymous hemianopia. MfVEPs were obtained using the Accumap (OPERA) system. With mfVEP, a hemifield was defined as abnormal if 3 or more contiguous sectors in a clinically suspicious area had either <90nV signal or p<0.01 on the amplitude probability plot. Abnormal fields were then determined if they were a hemianopic pattern. MRI's were reviewed for each hemianopia to determine lesion location. Results:Of the 46 normal subjects, 4 (9%) had at least one abnormal mfVEP using the <90nV standard, but none were hemianopic. Using the probability plot method, only 1 (2%) was abnormal and it was not hemianopic. While CAP showed hemianopic defects in all 15 patients, the mfVEP only showed hemianopic patterns in 10 (67%). In lesion location analysis, optic tract and chiasmal lesions resulted in mfVEP hemifields that respected the vertical midline and matched the CAP very closely. Optic radiation lesions had mfVEP hemifields that were less likely to respect the midline or match the CAP. When occipital cortex was involved, the mfVEP was even less accurate (3 of 6 respected the midline and matched CAP). One patient with cortical involvement had supernormal VEP amplitude traces in the spared regions of cortex. Conclusions: The mfVEP has suboptimal sensitivity and high specificity for hemianopias. It is more reliable in subcortical than cortical causes of hemianopia, perhaps due to the importance of cortex in generating the potentials.
Keywords: electrophysiology: clinical • perimetry • neuro–ophthalmology: diagnosis