1. In an anesthetized, paralyzed in vivo preparation, we recorded extracellular responses of 61 geniculate neurons (2 W, 25 X, 33 Y, and 1 mixed) to drifting sine-wave gratings of various spatial frequency, temporal frequency, and contrast. Our goal was to study the differential contributions to these visual responses of bursting caused by voltage dependent, low-threshold (LT) Ca2+ spikes and of purely tonic responses unrelated to LT spikes. Cells responding with LT spikes are said to be in the burst firing mode and those responding in a purely tonic fashion to be in the relay or tonic firing mode. We separated the total visual response into LT burst and tonic components by use of the empirical criteria set forth in our intracellular study described in the previous paper (Lu et al. 1992). A response component was considered to be an LT burst if its action potentials displayed interspike intervals less-than-or-equal-to 4 ms and if the first spike in the burst episode occurred after a silent period of greater-than-or-equal-to 100 ms (or greater-than-or-equal-to 50 ms when the neuron responds to visual stimuli at temporal rates greater-than-or-equal-to 8 Hz). All other activity is considered to be part of the tonic response. 2. In addition to LT bursts, we recognized another type of burst response, the high-threshold (HT) burst. These also have clusters of action potentials with interspike intervals less-than-or-equal-to 4 ms. However, HT bursts, unlike LT bursts, lack a preburst silent period. HT bursts are part of the tonic response component and merely reflect the gradual decrease in interspike intervals that occurs as the cell becomes more depolarized and thus more responsive. Thus interspike interval is a necessary but insufficient criterion to identify LT bursts. 3. Visually evoked LT bursts were recorded among W, X, and Y cells. When evoked, LT bursts occurred in phase with drifting sine-wave grating stimuli at a rate never exceeding one per stimulus cycle. In response to individual cycles of the visual stimulus, LT bursts could comprise the total response, a tonic component could comprise the total response, or an LT burst and tonic component could be mixed. When a stimulus evoked a mixture of LT bursts and tonic response components, LT bursts were always the first response. 4. Of the 61 cells tested with grating stimuli, 47 exhibited LT bursts and 14 did not. Those that did exhibited varying amounts of burstiness. We occasionally recorded individual cells for a sufficiently lengthy period to observe them switch between the burst and tonic firing modes. We conclude that most cells have membrane potentials close to the level needed for LT spike deinactivation and that small fluctuations in membrane voltage can switch them between these response modes. We also occasionally recorded several cells simultaneously and noted that different cells could be in different modes. This implies that whatever afferent pathways control response modes need not act globally on all geniculate neurons. 5. Two indexes were used to express the variation of LT burstiness. For the first index, an LT burst ratio was determined for each cell by separating the total visual response into LT burst and tonic components and calculating the fundamental (F1) Fourier response amplitude for each. Once separated, we divided the F1 amplitude of the LT burst component by the sum of F1 amplitudes of the LT burst and tonic components. The other index was the percentage of stimulus cycles that elicit LT bursts. This percentage was computed by inspecting the response to each stimulus cycle and determining the presence or absence of an LT burst. The indexes were highly correlated (r = 0.93), and both showed that LT bursts represent a variable proportion of the total response from barely detectable for some cells to the vast majority of the response for others. The proportion of cells exhibiting LT bursts was slightly higher for Y cells (88%) than for X cells (64%). 6. Inspection of average response histograms indicated that the LT burst component added a substantial nonlinearity to the visual response. Is a measure of nonlinearity, we used the second Fourier harmonic (F2) component of the response and computed an F2-to-F1 ratio. For every neuron tested, this ratio was larger for the LT burst than the tonic response component. Thus LT bursts seem to distort visual responses by making them less linear. 7. To test for the possibility that the amount of LT bursting may vary with stimulus strength, we generated response-versus-spatial frequency functions for 35 neurons and response-versus-contrast response functions for 15 neurons. The extent of LT bursting seen in cells did not vary significantly with spatial frequency or contrast; that is, the proportion of LT bursting in the total response was no more prevalent at one spatial frequency or contrast than another. Thus LT bursts do not appear to provide a selective nonlinear response amplification for weak or less salient stimuli. 8. To test for the possibility that LT bursting may vary with temporal frequency, we generated response-versus-temporal frequency functions (at optimal spatial frequency) for 17 neurons. We found a progressive increase in the LT burst ratio with increasing temporal frequency. 9. As noted, we found that LT bursts occurred earlier in the response cycle than did the tonic response components. We used the response-versus-temporal frequency functions of 10 cells to examine whether these temporal differences reflect phase and/or latency differences. The latencies of the two components were virtually identical, but the LT bursts were evoked by an earlier phase of the stimulus than was the tonic response component. We thus conclude that the temporal difference between the LT burst and tonic response components was due primarily to phase and not to latency. 10. We conclude that LT spikes contribute to the transfer of visual signals through the lateral geniculate nucleus to visual cortex. They do not represent an obligatory disconnection of thalamic relay cells from their sensory inputs. Instead, they can provide a nonlinear amplification that permits hyperpolarized relay cells to signal cortex about the presence of a salient stimulus.
