Purchase this article with an account.
B.G. Borghuis, N.K. Dhingra, R.G. Smith, P. Sterling; Neighboring Retinal Ganglion Cells Provide Independent Channels for Contrast Detection . Invest. Ophthalmol. Vis. Sci. 2005;46(13):2278.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: Threshold for detecting a small, brief stimulus is the same for a human and a brisk–transient ganglion cell, about 3% contrast. We asked whether psychophysical detection might depend on the response of a single ganglion cell with no loss along the visual pathway, or on the responses of several neighboring ganglion cells with significant losses. Methods: Recording simultaneously from pairs of neighboring brisk–transient ganglion cells, we used an ideal observer to compute detection thresholds. We recorded spatio–temporal receptive field characteristics with a white–noise stimulus and the response to a brief (100 ms) spot of optimal diameter, centered on either cell or halfway in between. Results: Average distance between neighboring receptive field centers was 208 ± 92 µm (ON/ON) and 137 ± 42 µm (OFF/OFF). This is equivalent to 1.9 ± 0.4 σ (ON/ON) and 1.3 ± 0.3 σ (OFF/OFF) of a difference–of–Gaussian fit to the spatial receptive fields and reflects a physiological coverage factor of about 5. When a spot of medium contrast (14 %) was centered on one cell, the neighbor also responded. However, when a spot of low contrast was presented, the neighbor ceased to respond and therefore did not contribute to detection. In a simulation, combining responses of up to 6 neighbors lowered contrast threshold only slightly. When a spot was positioned between two cells, threshold approximated the average for the cells measured separately to a spot centered on their receptive fields. Thus, for a small spot contrast threshold is constant over space and about equal to single cell threshold. Such constancy is obtained with the fewest cells (in simulation) when receptive field center spacing is about 2 σ, which approximates the recorded value. Cross–correlations between nearest neighbors were mostly stimulus–evoked and were strongly reduced for fine spatial stimuli. Conclusions: For a small stimulus near threshold, ganglion cells signal contrast independently. Thus, psychophysical detection of a small stimulus can be mediated by a single cell. Furthermore, the observed receptive field spacing is near optimal for creating an even sensitivity surface with the fewest cells.
This PDF is available to Subscribers Only