Purchase this article with an account.
Manuel Spitschan, Long Luu, Ritobrato Datta, David H Brainard, Geoffrey K Aguirre; Melanopsin-driven responses in the human brain. Invest. Ophthalmol. Vis. Sci. 2014;55(13):5765.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Photopic vision arises from L, M and S cones as well as from the recently discovered intrinsically photosensitive retinal ganglion cells (ipRGCs) containing the photopigment melanopsin. Much is known about the brain pathways of signals originating from cones. Considerably less is known about the projections of the ipRGCs, in particular in the human brain. Here we used BOLD fMRI to measure neural responses in humans to spectral modulations that selectively target melanopsin while minimizing the responses of the cones, and compare these to responses generated using cone-targeted spectral modulations.
Four observers viewed large-field (27.5° diameter, central 5° obscured) sinusoidal flicker (1-16 Hz, log spaced) at 470 cd/m2 during BOLD fMRI. Using the method of silent substitution and a digital light synthesizer, spectral modulations were directed at melanopsin, L+M, L-M, or S cones. An isochromatic modulation (L+M+S+melanopsin) was also studied. Estimates of photopigment spectral sensitivities used to produce modulations accounted for observer age and stimulus extent. Anatomical regions of interest were defined for the lateral geniculate nucleus (LGN; using volumetric templates), for primary visual cortex (V1) and for extrastriate areas (V2, V3 and hV4) (Benson VSS2013). Cortical voxels were restricted to >5° eccentricity as an extra precaution against contamination from changes in cone spectral sensitivity at the fovea. Temporal transfer functions (TTFs), describing the BOLD response as a function of temporal frequency for each modulation direction, were averaged for the four observers.
We found robust melanopsin-driven responses in both LGN and V1-hV4. In LGN, the melanopsin response was bandpass, maximal at 8 Hz with no measurable response below 2 Hz. The melanopsin response in V1 and V2 through hV4 was also bandpass. Across areas the responses to cone-directed stimuli differed with the direction of the spectral modulation, and differed as well from the responses to melanopsin stimulation. S-cone responses were distinct from melanopsin responses in both LGN and cortex.
Signals originating in the melanopsin-containing ipRGCs produce robust responses in the LGN and visual cortex of the human brain. The TTFs for melanopsin-driven responses are bandpass in both LGN and visual cortex. The responses tomelanopsin-driven stimulation have a temporal signature distinct from that of cone-driven responses.
This PDF is available to Subscribers Only