Search

To understand visual functions mediated by intrinsically photosensitive melanopsin-expressing retinal ganglion cells (mRGCs), it is important to elucidate axonal projections from these cells into the brain. Initial studies reported that melanopsin is expressed only in retinal ganglion cells within the eye. However, recent studies in Opn4-Cre mice revealed Cre-mediated marker expression in multiple brain areas. These discoveries complicate the use of melanopsin-driven genetic labeling techniques to identify retinofugal projections specifically from mRGCs. To restrict labeling to mRGCs, we developed a recombinant adeno-associated virus (AAV) carrying a Cre-dependent reporter (human placental alkaline phosphatase) that was injected into the vitreous of Opn4-Cre mouse eyes. The labeling observed in the brain of these mice was necessarily restricted specifically to retinofugal projections from mRGCs in the injected eye. We found that mRGCs innervate multiple nuclei in the basal forebrain, hypothalamus, amygdala, thalamus and midbrain. Midline structures tended to be bilaterally innervated, whereas the lateral structures received mostly contralateral innervation. As validation of our approach, we found projection patterns largely corresponded with previously published results; however, we have also identified a few novel targets. Our discovery of projections to the central amygdala suggests a possible direct neural pathway for aversive responses to light in neonates. In addition, projections to the accessory optic system suggest that mRGCs play a direct role in visual tracking, responses that were previously attributed to other classes of retinal ganglion cells. Moreover, projections to the zona incerta raise the possibility that mRGCs could regulate visceral and sensory functions. However, additional studies are needed to investigate the actual photosensitivity of mRGCs that project to the different brain areas. Also, there is a concern of "overlabeling" with very sensitive reporters that uncover low levels of expression. Light-evoked signaling from these cells must be shown to be of sufficient sensitivity to elicit physiologically relevant responses.

Melanopsin-expressing retinal ganglion cells (mRGCs) in the eye play an important role in many light-activated non-image-forming functions including neonatal photoaversion and the adult pupillary light reflex (PLR). MRGCs rely on glutamate and possibly PACAP (pituitary adenylate cyclase-activating polypeptide) to relay visual signals to the brain. However, the role of these neurotransmitters for individual non-image-forming responses remains poorly understood. To clarify the role of glutamatergic signaling from mRGCs in neonatal aversion to light and in adult PLR, we conditionally deleted vesicular glutamate transporter (VGLUT2) selectively from mRGCs in mice. We found that deletion of VGLUT2 in mRGCs abolished negative phototaxis and light-induced distress vocalizations in neonatal mice, underscoring a necessary role for glutamatergic signaling. In adult mice, loss of VGLUT2 in mRGCs resulted in a slow and an incomplete PLR. We conclude that glutamatergic neurotransmission from mRGCs is required for neonatal photoaversion but is complemented by another non-glutamatergic signaling mechanism for the pupillary light reflex in adult mice. We speculate that this complementary signaling might be due to PACAP neurotransmission from mRGCs.

Melanopsin-expressing intrinsically photosensitive retinal ganglion cells (ipRGCs) are the only functional photoreceptive cells in the eye of newborn mice. Through postnatal day 9, in the absence of functional rods and cones, these ipRGCs mediate a robust avoidance behavior to a light source, termed negative phototaxis. To determine whether this behavior is associated with an aversive experience in neonatal mice, we characterized light-induced vocalizations and patterns of neuronal activation in regions of the brain involved in the processing of aversive and painful stimuli. Light evoked distinct melanopsin-dependent ultrasonic vocalizations identical to those emitted under stressful conditions, such as isolation from the litter. In contrast, light did not evoke the broad-spectrum calls elicited by acute mechanical pain. Using markers of neuronal activation, we found that light induced the immediate-early gene product Fos in the posterior thalamus, a brain region associated with the enhancement of responses to mechanical stimulation of the dura by light, and thought to be the basis for migrainous photophobia. Additionally, light induced the phosphorylation of extracellular-related kinase (pERK) in neurons of the central amygdala, an intracellular signal associated with the processing of the aversive aspects of pain. However, light did not activate Fos expression in the spinal trigeminal nucleus caudalis, the primary receptive field for painful stimulation to the head. We conclude that these light-evoked vocalizations and the distinct pattern of brain activation in neonatal mice are consistent with a melanopsin-dependent neural pathway involved in processing light as an aversive but not acutely painful stimulus.

The morphology of dendrites constrains and reflects the nature of synaptic inputs to neurons. The visual system has served as a useful model to show how visual function is determined by the arborization patterns of neuronal processes. In retina, light ON and light OFF responding ganglion cells selectively elaborate their dendritic arbors in distinct sublamina, where they receive, respectively, inputs from ON and OFF bipolar cells. During neonatal maturation, the bilaminarly distributed dendritic arbors of ON-OFF retinal ganglion cells (RGCs) are refined to more narrowly localized monolaminar structures characteristic of ON or OFF RGCs. Recently, brain-derived neurotrophic factor (BDNF) has been shown to regulate this laminar refinement, and to enhance the development of dendritic branches selectively of ON RGCs. Although other related neurotrophins are known to regulate neuronal process formation in the central nervous system, little is known about their action in maturing retina. Here, we report that overexpression of neurotrophin-3 (NT-3) in the eye accelerates RGC laminar refinement before eye opening. Furthermore, NT-3 overexpression increases dendritic branch number but reduces dendritic elongation preferentially in ON-OFF RGCs, a process that also occurs before eye opening. NT-3 overexpression does affect dendritic maturation in ON RGCs, but to a much less degree. Taken together, our results suggest that NT-3 and BDNF exhibit overlapping effects in laminar refinement but distinct RGC-cell-type specific effects in shaping dendritic arborization during postnatal development.

In the few days prior to eye-opening in mice, the excitatory drive underlying waves switches from cholinergic to glutamatergic. Here, we describe the unique synaptic and spatiotemporal properties of waves generated by the retina's glutamatergic circuits. First, knockout mice lacking vesicular glutamate transporter type 1 do not have glutamatergic waves, but continue to exhibit cholinergic waves, demonstrating that the two wave-generating circuits are linked. Second, simultaneous outside-out patch and whole-cell recordings reveal that retinal waves are accompanied by transient increases in extrasynaptic glutamate, directly demonstrating the existence of glutamate spillover during waves. Third, the initiation rate and propagation speed of retinal waves, as assayed by calcium imaging, are sensitive to pharmacological manipulations of spillover and inhibition, demonstrating a role for both signaling pathways in shaping the spatiotemporal properties of glutamatergic retinal waves.

In mammals, retinal ganglion cell (RGC) projections initially intermingle and then segregate into a stereotyped pattern of eye-specific layers in the dorsal lateral geniculate nucleus (dLGN). Here we found that in mice deficient for ephrin-A2, ephrin-A3 and ephrin-A5, eye-specific inputs segregated but the shape and location of eye-specific layers were profoundly disrupted. In contrast, mice that lacked correlated retinal activity did not segregate eye-specific inputs. Inhibition of correlated neural activity in ephrin mutants led to overlapping retinal projections that were located in inappropriate regions of the dLGN. Thus, ephrin-As and neural activity act together to control patterning of eye-specific retinogeniculate layers.

Synaptically localized calcium channels shape the timecourse of synaptic release, are a prominent site for neuromodulation, and have been implicated in genetic disease. In retina, it is well established that L-type calcium channels play a major role in mediating release of glutamate from the photoreceptors and bipolar cells. However, little is known about which calcium channels are coupled to synaptic exocytosis of glycine, which is primarily released by amacrine cells. A recent report indicates that glycine release from spiking AII amacrine cells relies exclusively upon L-type calcium channels. To identify calcium channel types controlling neurotransmitter release from the population of glycinergic neurons that drive retinal ganglion cells, we recorded electrical and potassium evoked inhibitory synaptic currents (IPSCs) from these postsynaptic neurons in retinal slices from tiger salamanders. The L-channel antagonist nifedipine strongly inhibited release and FPL64176, an L-channel agonist, greatly enhanced it, indicating a significant role for L-channels. omega-Conotoxin MVIIC, an N/P/Q-channel antagonist, strongly inhibited release, indicating an important role for non-L channels. While the P/Q-channel blocker omega-Aga IVA produced only small effects, the N-channel blocker omega-conotoxin GVIA strongly inhibited release. Hence, N-type and L-type calcium channels appear to play major roles, overall, in mediating synaptic release of glycine onto retinal ganglion cells.

ON and OFF pathways separately relay increment and decrement luminance signals from retinal bipolar cells to cortex. ON-OFF retinal ganglion cells (RGCs) are activated via synaptic inputs onto bistratified dendrites localized in the ON and OFF regions of the inner plexiform layer. Postnatal maturational processes convert bistratifying ON-OFF RGCs to monostratifying ON and OFF RGCs. Although visual deprivation influences refinement of higher visual centers, no previous studies suggest that light regulates either the development of the visual-evoked signaling in retinal ON and OFF pathways, nor pruning of bistratified RGC dendrites. We find that dark rearing blocks both the maturational loss of ON-OFF responsive RGCs and the pruning of dendrites. Thus, in retina, there is a previously unrecognized, pathway-specific maturation that is profoundly affected by visual deprivation.

Ganglion cells responding only transiently to changes in illumination are found in many different vertebrate retinas. The interactions underlying formation of these transient responses are still poorly understood. Two recently proposed hypotheses are (1) functional inhibitory pathways are necessary for transient response production, and (2) direct inhibition of the ganglion cell has little effect on its output. Here, we examine these conclusions by using cell-attached patch-clamp recordings of spiking, whole-cell recordings of synaptic currents, and computer modeling. We found that picrotoxin (a GABA(A) and GABA(C) receptor antagonist), bicuculline (a GABA(A) receptor antagonist), and strychnine (a glycine receptor antagonist), applied either singly or in combination, always failed to convert transient responses to sustained responses. Application of the GABA(B) antagonist CGP35348 in the presence of picrotoxin and strychnine also failed to convert transient responses into sustained responses. Whole-cell recordings of synaptic currents at various holding potentials indicated that direct inhibitory inputs to ganglion cells limit the duration of net excitation, implying that direct inhibition does act to truncate the ganglion cell spiking response. Computer simulations using spiking and synaptic data from combined cell-attached and whole-cell recordings supported this interpretation. We conclude that inhibitory pathways are not required for generation of transient responses, but these pathways do serve to modulate transient ganglion cell spiking responses. We find that this modulation occurs, in part, via inhibitory inputs directly to the ganglion cell.

Rapid and reliable detection of randomly occurring small amplitude synaptic events resulting from activation of different classes of ligand-gated receptors is a difficult task. Here, we describe and characterize an amplitude threshold algorithm, written as an IGOR Pro procedure, which detects events as well as characterizes their amplitudes and kinetics. The program was developed to analyze recording traces that each contained both excitatory (glutamate-mediated) and inhibitory (GABA and glycine-mediated) events. By using differences in kinetics, the program could discriminate between the two different classes of events. In summary, the program has the following strengths: (1) it is generally applicable to circumstances in which different populations of elementary events occur concurrently, a drawback of methods that employ matched filtering techniques, (2) it is relatively insensitive to drifts in baseline, and (3) it generates user-accessible arrays of the timing, amplitude and kinetic parameters of the detected events, making customized statistical analysis of event characteristics easily executable.

Spontaneous inhibitory and excitatory postsynaptic currents (sIPSCs and sEPSCs) were identified and characterized with whole cell and perforated patch voltage-clamp recordings in adult mouse retinal ganglion cells. Pharmacological dissection revealed that all cells were driven by spontaneous synaptic inputs mediated by glutamate and gamma-aminobutyric acid-A (GABAA) receptors. One-half (7/14) of the cells also received glycinergic spontaneous synaptic inputs. Both GABAA and glycine receptor-mediated sIPSCs had rise times (10-90%) of < 1 ms. The decay times of the GABAA receptor-mediated sIPSCs were comparable with those of the glycine receptor-mediated sIPSCs. The average decay time constant for monoexponentially fitted sIPSCs was 63.2 +/- 74.1 ms (mean +/- SD, n = 3278). Glutamate receptor-mediated sEPSCs had an average rise time of 0.50 +/- 0.20 ms (n = 109) and an average monoexponential decay time constant of 5.9 +/- 8.6 ms (n = 2705). Slightly more than two-thirds of the spontaneous synaptic events were monoexponential (68% for sIPSCs and 76% for sEPSCs). The remainder of the events was biexponential. The amplitudes of the spontaneous synaptic events were not correlated with rise times, suggesting that the electrotonic filtering properties of the neurons and/or differences in the spatial location of synaptic inputs could not account for the difference between the decay time constants of the glutamate and GABAA/glycine receptor-mediated spontaneous synaptic events. The amplitudes of sEPSCs were similar to those recorded in tetrodotoxin (TTX), consistent with the events measured in control saline being the response to the release of a single quantum of transmitter. The range of the sIPSC amplitudes in control saline was wider than that recorded in TTX, consistent with some sIPSCs being evoked by presynaptic spikes having an average quantal size greater than one. The rates of sIPSCs and sEPSCs were determined under equivalent conditions by recording with perforated patch electrodes at potentials at which both types of event could be identified. Two groups of ganglion cell were observed; one group had an average sEPSCs/sIPSCs frequency ratio of 0.96 +/- 0.77 (n = 28) and another group had an average ratio of 6.63 +/- 0.82 (n = 7). These findings suggest that a subset of cells is driven much more strongly by excitatory synaptic inputs. We propose that this subset of cells could be OFF ganglion cells, consistent with the higher frequency of spontaneous action potentials found in OFF ganglion cells in other studies.

The passive electrical properties of 17 ON-OFF retinal ganglion cells were derived from electrophysiological recordings. The parameters for each cells' equivalent model were obtained from the transient current responses to small step changes in clamp potential. Thirteen of the cells could be adequately approximated by a spherical soma connected to an equivalent dendritic cable. Estimates for the cell input conductance (GN), membrane time constant (tau m), the dendritic-to-soma conductance ratio (rho), and the normalized electrotonic length (L) were obtained (mean +/- standard deviation, n = 13): GN = 580 +/- 530 pS, tau m = 97 +/- 72 ms, rho = 2.8 +/- 2.8, and L = 0.34 +/- 0.13. Series resistance averaged 32 +/- 11 M omega. The mean of the derived soma diameters was 18 +/- 6 microns and the mean diameter and length of the equivalent cables were 1.4 +/- 0.6 and 470 +/- 90 microns, respectively. The average of the specific membrane conductances, 1.67 +/- 1.08 S/cm2, corresponded to a membrane resistivity of 60 k omega. cm2. Computer simulations of synaptic inputs were performed on a representative model, with an electrode at the soma and using the worst-case configuration, in which all synaptic inputs were confined to the tips of the dendrites. We draw three conclusions from the modeling: (1) Under voltage clamp, fast spontaneous EPSCs would be significantly attenuated and slowed while the time course of the slower, light-evoked non-NMDA and NMDA EPSCs would be minimally distorted by dendritic filtering. (2) Excitatory synaptic reversal potentials can be accurately determined under voltage clamp. (3) In the absence of GABAergic and glycinergic inhibition, the efficacy at the soma of excitatory conductance changes is essentially independent of their dendritic location. The specific membrane resistivity appears to represent a good compromise between having a small membrane time constant and minimal EPSP attenuation.

1. Light-evoked input-output characteristics of ganglion cells in dark-adapted tiger salamander retina were studied in the slice preparation using patch-clamp techniques. Excitatory postsynaptic currents (EPSCs), isolated by blocking inhibitory inputs and evoked by a range of light stimulus intensities, were recorded under whole-cell voltage clamp. Spike responses, evoked by the same light intensities, were recorded extracellularly from the same cells using the cell-attached patch-clamp technique. 2. When N-methyl-D-aspartate (NMDA) receptor-mediated input was blocked by the competitive NMDA antagonist DL-2-amino-5-phosphonoheptanoate (AP7), light-evoked EPSC amplitude and peak firing rate were reduced at all light intensities. In both cases, the data obtained in the presence of AP7 scaled linearly to control data, indicating that NMDA and non-NMDA receptors are activated in the same proportions across the entire 2 log unit stimulus response range of these ganglion cells. 3. The relationship between light-evoked spike frequency and light-evoked EPSC amplitude was linear. The slope of the light-evoked synaptic current-spike frequency relationship was close to the slope of the injected current-spike frequency relationship, indicating that synaptic current and injected current drive spiking in a similar manner. The linearity of the synaptic current-spike frequency relationship was not compromised when NMDA input was blocked by AP7. 4. Light-evoked voltage responses, recorded under whole-cell current clamp, revealed that the average membrane potential during a spike response was depolarized only slightly with increased firing rate. Once the membrane potential surpassed spike threshold, it was maintained by the voltage-gated, spike-generating conductances at a depolarized plateau upon which action potentials were fired. The potential of this plateau varied only slightly with spike frequency. We conclude that the voltage control exerted by the spike-generating currents in ganglion cells prevents a substantial response-dependent decrease in the electrical driving force of the excitatory currents, obviating the need for the voltage-independent synaptic efficacy provided by the combination of NMDA and non-NMDA inputs.

1. Spontaneous excitatory postsynaptic currents (sEPSCs) were recorded under voltage-clamp conditions. Consistent with activation of non-NMDA-type glutamate receptors, the sEPSCs reversed at potentials above 0 mV, were blocked by 1 microM CNQX and prolonged by 2 mM aniracetam. 2. The peak conductance of the averaged sEPSCs (n = 70-400) was 130 +/- 60 pS (mean +/- S.D.; 17 cells, ranging from 70 to 290 pS). Amplitude distributions were skewed towards larger amplitudes. 3. The decay of individual and mean sEPSCs was exponential with a mean time constant (tau d) of 3.75 +/- 0.84 ms (n = 13), which was voltage independent. The 10-90% rise time of the sEPSCs was 1.30 +/- 0.44 ms (n = 13). There was no correlation between sEPSC rise time and tau d suggesting that dendritic filtering alone did not shape the time course of sEPSCs. 4. Light-evoked EPSCs in these retinal ganglion cells are mediated by concomitant activation of NMDA and non-NMDA receptors; however, no NMDA component was discerned in the sEPSCs, even when recording at -96 mV in Mg(2+)-free solutions. The decay time course was not altered by 20 microM AP7, an NMDA antagonist, nor was an NMDA component unmasked by adding glycine or D-serine. These results suggest that NMDA and non-NMDA receptors are not coactivated by a single vesicle of transmitter during spontaneous release, and thus are probably not colocalized in the postsynaptic membrane at the sites of spontaneous release. 5. The sEPSCs were an order of magnitude faster than the non-NMDA receptor-mediated EPSCs evoked by light stimuli, and it is proposed that the EPSC time course is determined largely by the extended time course of release of synaptic vesicles from bipolar cells. The quantal content of a light-evoked non-NMDA receptor-mediated EPSC in an on-off cell is about 200 quanta.

To examine how light-evoked excitatory synaptic inputs to retinal ganglion cells are transformed into output patterns of activity, action potentials were recorded with cell-attached patch-clamp techniques, and then EPSCs and EPSPs were recorded from the same cell in the whole-cell configuration. AP7, an NMDA antagonist, reduced the light-evoked peak spike frequency 36% +/- 21% (mean +/- SD) and reduced the EPSC amplitude, indicating a major role for NMDA receptors in the light response. CNQX, a non-NMDA receptor antagonist, reduced the light-evoked peak spike frequency 28% +/- 22%. CNQX also caused a voltage- and magnesium-dependent delay in spike onset. AP7 and CNQX, however, did not differ significantly in their effect on the EPSC time course, indicating that postsynaptic cellular properties are responsible for the delay observed in the presence of CNQX. These results show that the NMDA receptor contribution to the excitatory response is increased as the cell is depolarized from rest by non-NMDA input.

1. Cells in the ganglion cell layer of salamander retinal slices were voltage clamped using patch pipettes. Light elicited transient excitatory postsynaptic currents (EPSCs) in on-off ganglion cells and sustained EPSCs in on ganglion cells. Light-evoked inhibitory postsynaptic currents in these cells could be blocked by 100 microM-bicuculline methobromide and 500 nM-strychnine. 2. In the presence of external Cd2+, at a concentration that blocked light-evoked synaptic inputs, N-methyl-D-aspartate (NMDA) and the non-NMDA-receptor agonists, quisqualate and kainate, gated conductances in both on-off and on ganglion cells. The current-voltage (I-V) curve for the conductance elicited by NMDA had a negative slope between -40 and -70 mV and a reversal potential near 0 mV. The I-V curves for the non-NMDA-receptor-mediated conductances were nearly linear and also had reversal potentials near 0 mV. 3. I-V curves were measured at an early time point near the peak of transient EPSCs and at a later time point during the decay phase of the responses. The late I-V curve had a negative slope below -40 mV. The early I-V curve had a positive slope over the entire voltage range but the slope was greater at positive than at negative potentials. The evoked current reversed near 0 mV at both time points. 4. The region of negative slope of the late I-V curve was eliminated when Mg2+ was removed from the external saline. A slowly decaying component of transient EPSCs was eliminated in 20 microM-DL-2-amino-7-phosphonoheptanoate (AP7), an NMDA-receptor antagonist. 5. Application of 1 microM-6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), a non-NMDA-receptor antagonist at this concentration, blocked a fast component of transient EPSCs. 6. Our results demonstrate that the synaptic inputs to on-off ganglion cells have two components: a slower NMDA-receptor-mediated component having a time-to-peak of 110 +/- 45 ms and an e-fold decay time of 209 +/- 35 ms at -31 mV (mean +/- S.D., n = 5), and a faster non-NMDA-receptor-mediated component having a time-to-peak of 28 +/- 10 ms and an e-fold decay time of 43 +/- 20 ms at -31 mV (n = 8). 7. A similar analysis of sustained EPSCs of on ganglion cells showed that these currents resulted from sustained activation of both NMDA and non-NMDA receptors.

1. The performance of ganglion cells in detecting flashes of light near the absolute threshold was studied in an isolated eye-cup preparation of toad retina. Retinal ganglion cells, through which all visual information from the rods must flow to the brain, are in a key position for evaluating the still unproven hypothesis that the absolute light sensitivity is limited by rod noise (Barlow, 1956). 2. The dark-adapted threshold intensity for these cells, which were selected on the basis of their high sensitivity, averaged 0.029 Rh* flash-1 (range 0.008-0.062), where Rh* signifies one photoisomerization per rod. On average, 46 photoisomerizations were needed per receptive field per flash to evoke a threshold response (range 16-84). 3. In the threshold region, frequency of responses versus mean flash intensity was determined. Threshold performance could be described by theoretical frequency of response curves, allowing intrinsic noise to be estimated in terms of an equivalent rate of photoisomerization-like (dark) events. In two completely characterized cells the rate of dark events corresponded to 0.03 and 0.06 Rh*DS-1, where Rh*D signifies one dark event per rod. 4. Threshold elevations produced by dim backgrounds were studied. The results of these experiments are consistent with a dark event rate equivalent to 0.046 Rh*DS-1, or 0.037 Rh*DS-1 after correcting for a probable decrease in summation time. 5. The rate of actual dark events (0.028 Rh*DS-1, 20 degrees C) measured in Bufo rods (Baylor, Lamb & Yau, 1980) is close to the equivalent rates determined here. Thus, for the ganglion cells signalling the dimmest lights, the dark events in rods appear to be the most significant intrinsic retinal noise source limiting detection.

Responses to flashes and steps of light were recorded intracellularly from rods and horizontal cells, and extracellularly from ganglion cells, in toad eyecups which were either dark adapted or exposed to various levels of background light. The average background intensities needed to depress the dark-adapted flash sensitivity by half in the three cell types, determined under identical conditions, were 0.9 Rh*s-1 (rods), 0.8 Rh*s-1 (horizontal cells), and 0.17 Rh*s-1 (ganglion cells), where Rh* denotes one isomerization per rod. Thus, there is a range (approximately 0.7 log units) of weak backgrounds where the sensitivity (response amplitude/Rh*) of rods is not significantly affected, but where that of ganglion cells (1/threshold) is substantially reduced, which implies that the gain of the transmission from rods to the ganglion cell output is decreased. In this range, the ganglion cell threshold rises approximately as the square root of background intensity (i.e. in proportion to the quantal noise from the background), while the maintained rate of discharge stays constant. The threshold response of the cell will then signal light deviations (from a mean level) of constant statistical significance. We propose that this type of ganglion cell desensitization under dim backgrounds is due to a post-receptoral gain control driven by quantal fluctuations, and term it noise adaptation in contrast to the Weber adaptation (desensitization proportional to the mean background intensity) of rods, horizontal cells, and ganglion cells at higher background intensities.

Responses to light were recorded from rods, horizontal cells, and ganglion cells in dark-adapted toad eyecups. Sensitivity was defined as response amplitude per isomerization per rod for dim flashes covering the excitatory receptive field centers. Both sensitivity and spatial summation were found to increase by one order of magnitude between rods and horizontal cells, and by two orders of magnitude between rods and ganglion cells. Recordings from two hyperpolarizing bipolar cells showed a 20 times response increase between rods and bipolars. At absolute threshold for ganglion cells (Copenhagen, D.R., K. Donner, and T. Reuter. 1987. J. Physiol. 393:667-680) the dim flashes produce 10-50-microV responses in the rods. The cumulative gain exhibited at each subsequent synaptic transfer from the rods to the ganglion cells serves to boost these small amplitude signals to the level required for initiation of action potentials in the ganglion cells. The convergence of rod signals through increasing spatial summation serves to decrease the variation of responses to dim flashes, thereby increasing the signal-to-noise ratio. Thus, at absolute threshold for ganglion cells, the convergence typically increases the maximal signal-to-noise ratio from 0.6 in rods to 4.6 in ganglion cells.