Abstract
Purpose: :
In our 2010 ARVO abstract, we reported that crossover inhibition from the ON pathway of the retina improves contrast detection by an OFF ganglion cell (GC). Here we ask if the effects of cross inhibition might extend to the postsynaptic lateral geniculate nucleus (LGN) relay cell.
Methods: :
We modeled a common arrangement: an OFF relay cell driven predominantly by an OFF GC. The probability that a retinal spike triggers an LGN spike increases exponentially with decreasing interspike interval (ISI) according to function e(i). Thus we took an extracellular recording of an OFF alpha GC responding to a full field dark flash or to a sham stimulus without flash (9 cells). For each GC spike, we measured the interval to the previous spike, then determined the probability of triggering from e(i). This probability p was compared to a random number r (0 < r < 1). If p > r, then the model generated a spike. The simulated relay cell spike train was subjected to a statistical analysis (Fisher Linear Discriminant) that predicted whether a flash or a sham stimulus had occurred.
Results: :
To simulate the effect of blocking the ON pathway of the retina, we based the model on an OFF alpha GC under control conditions or with the ON bipolar cell blocked with L-AP4. The result was that the relay cell reliably detected (90% accuracy) a Michelson contrast of 5.5 ± 0.7% under control conditions, which increased to 16.4 ± 3.2% with the ON bipolar blocked. Under control conditions, detectable contrast increased by 2.2 ± 0.5 units across the retinogeniculate synapse but during L-AP4 it increased by even more: 8.8 ± 2.1 units. To test the idea that L-AP4 increase detectable contrast across the synapse by changing the ISI distribution of the GC, we calculated the proportion of "clustered" spikes that had an ISI less than 5 ms. In confirmation, we found that L-AP4 reduced the percentage of clustered spikes from 50 ± 4% to 28 ± 6%. We also calculated the efficacy of the retinal-LGN synapse by dividing the number of LGN spikes by the number of GC spikes. L-AP4 reduced synaptic efficacy by 11 ± 3%.
Conclusions: :
Crossover inhibition 1) improves contrast detection by OFF GCs and 2) clusters the OFF GC’s spikes which may improve the efficacy of transmission across the retinogeniculate synapse. Both of these effects would improve contrast detection by OFF LGN relay cells.
Keywords: retinal connections, networks, circuitry • ganglion cells • detection