Purpose: The objective detection of local visual field defects using multi-
focal pattern visual evoked potentials (VEP) has recently been described. T
he individual waveforms show variable polarity in different parts of the vi
sual field due to underlying cortical convolutions. Normal trace arrays wer
e examined to determine if certain areas of similar waveform could be group
ed for analysis, while minimising cancellation of data. Method: The VEP was
assessed using multi-focal pseudo-randomly alternated pattern stimuli whic
h were cortically scaled in size. Bipolar occipital electrodes were used fo
r recording. Waveforms were compared for different locations within the fie
ld up to 25 degrees of eccentricity. Analysis of sectors showing similarly
shaped waveforms was performed. Twelve normal subjects were studied. Result
: Grouping waveforms by sectors of similar waveform increased the total cal
culated upper hemifield amplitude by 60%, compared with simple summations o
f responses for the whole hemifield. The inferior hemifield showed more con
sistent waveforms throughout, with the amplitude only increasing by 11% wit
h sectoral summation. Intra-subject variability (10.6%) is less for sectors
than for individual points (17.3%). Inter-subject amplitude differences ar
e high, calculated at 56% for individual points and 45% for sectors. Conclu
sions: Due to differences in waveform as a result of underlying cortical an
atomy, individual VEP responses from multifocal recordings should be groupe
d as sectors along the vertical meridian and above and below the horizontal
, rather than by hemifields or quadrants. This finding is significant if on
e is considering within-field grouping strategies similar to the glaucoma h
emifield test used in conventional perimetry, or reporting derived overall
VEP amplitudes and latencies from a multifocal recording. Large amplitude v
ariations between individuals and small signals from horizontal and upper f
ield seen in single channel recording, still limit the application of this
technique as a form of objective perimetry.