Augmented Reality can induce illusory perceptions of heat when users experience virtual flames rendered on their hands. In this exploratory study, we used high-density electroencephalography to investigate the neural correlates associated with this experience. We recorded neural activity in a within-subjects design during three key audiovisual conditions: (A) lifelike virtual flames on the hand paired with a burning sound; (B) a non-fire-like control stimulus (green, downward-flowing visuals with a synthetic sound) on the hand; and (C) realistic flames presented between the hands. Our preliminary analyses revealed distinct electroencephalography signatures between the Fire-on-Hand condition and both control conditions. Our findings suggest that AR-induced thermal illusions may engage brain networks involved in somatosensation, multisensory body representation, and spatial attention. This top-down, predictive mechanism appears to recruit neural systems that overlap with those involved in the higher-order cortical processing of actual thermal stimuli.

Dual intracellular recording of nearby pairs of pyramidal cells in slices of rat visual cortex has shown that there are significant differences in functional connectivity between the superficial and deep layers (Mason et al. 1991; Nicoll and Blakemore 1993). For pairs of cells no farther than 300 μm apart, synaptic connections between layer 2/3 pyramidal neurons were individually weaker (median peak amplitude, A, of single-fiber excitatory postsynaptic potentials, EPSPs, = 0.4 mV) but more frequent (connection probability, p = 0.087) than those between layer 5 pyramidal neurons (mean A = 0.8 mV, p < 0.015). Taken in combination with plausible estimates of the density of pyramidal cells, the total numbers of synapses on them and the number of synapses formed on their intracortical axons, the present analysis of the above data suggests that roughly 70% of the excitatory synapses on any layer 2/3 pyramid, but fewer than 1% of those on a layer 5 pyramidal neuron, are derived from neighboring pyramidal neurons in its near vicinity. Even assuming very extreme values for some parameters, chosen to erode this difference, the calculated proportion of "local synapses" for layer 5 pyramids was always markedly lower than for layer 2/3 pyramidal neurons. These results imply that local excitatory connections are much more likely to provide significant "intracortical amplification" of afferent signals in layer 2/3 than in layer 5 of rat visual cortex.