Dipole source modelling of the magnetoencephalogram to stereopsis, binocular fusion and rivalry.

Abstract

 

Purpose: To characterise magnetoencephalographic activity (MEG) during normal binocular fusion, binocular rivalry and stereopsis using red/green stereograms. Methods: Subjects viewed red/green random dot stereo pairs. A horizontal grating (0.2 cycles/degree) was detectable only when stereo-pairs contained horizontal disparity. Similar vertical disparity produced binocular rivalry. Zero disparity was the control condition for flat fusion. MEG signals were recorded from 9 normal subjects using a helmet-shaped 306-channel detector array, and band-passed from 0.03 to 200 Hz with a sampling rate of 600 Hz excluding blink artefacts. Visually evoked fields (VEFs) time-locked to each stimulus pair were modelled as equivalent current dipoles whose 3-D locations, orientations and strengths were found using signals from several sensors over the maximum response area. Results: Maximal activity in the VEF was present at about 150 ms and again between 250 and 300 ms. A 3-dipole model was developed consisting of left and right occipital dipoles and a single parietal dipole. This model accounted for over 80% of the MEG activity at the maxima. The maxima at 150 ms had contributions from all three dipoles. After 250 ms the activity could be represented by a single parietal dipole which was strongest during horizontal disparity and was weaker during vertical disparity. Conclusion: Stimuli containing stereoscopic depth selectively increase MEG activity in the parietal cortex while activity localised to the occipital areas is relatively constant for stimulus patterns containing fusion, rivalry and stereoscopic depth cues.

 

Footnotes

 Support: European Union under the NeuroBirch III Initiative, Helsinki, Finland