| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology, University of Oslo, Institute of Basic Medical Sciences, Oslo, Norway
Submitted 20 December 2005; accepted in final form 7 August 2006
 |
ABSTRACT |
|---|
|
70 ms after stimulus onset. The changes in spatial resolution did not follow changes of firing rate; peak firing appeared earlier than the maximal spatial resolution. We suggest that the response initially conveys a strong but spatially coarse message that might have a detection and tune-in function, followed by transient transmission of spatially precise information about the stimulus. Experiments with spots presented inside the maximum but outside the minimum center width suggested a dynamic reduction in number of responding neurons during the stimulation; from many responding neurons initially when the field centers are large to fewer responding neurons as the centers shrink. Thereby, there is a change from coarse-to-fine also in the recruitment of responding neurons during brief static stimulation. |
|
INTRODUCTION |
|---|
|
; Knierim and Van Essen 1992; Suder et al. 2002; Wörgötter et al. 1998; Zisper et al. 1996). Reported changes include shrinkage of receptive field subregions during brief (300 ms) static spot stimulation (Suder et al. 2002; Wörgötter et al. 1998). Because width of the receptive field subregions is generally regarded as an indicator of the spatial resolution mediated by the neurons, this suggested a dynamic increase of spatial resolution during the brief spot stimulation (Wörgötter et al. 1998). Increasing spatial resolution during such short periods of stimulation was subsequently shown directly (Bredfeldt and Ringach 2002; Frazor et al. 2004). Frazor et al. (2004) presented static gratings of variable spatial frequency on the receptive field. They pointed out that their stimulation periods (200 ms) were similar to typical durations of fixation periods in natural saccadic inspections (Evinger and Fuchs 1978; Stryker and Blakemore 1972; Yarbus 1967), suggesting that dynamic increase of spatial resolution might occur during single fixations. Beside this coarse-to-fine dynamics of receptive field structure and spatial resolution, similar dynamics have been shown with respect to several other types of response selectivity in cortical neurons during brief static stimulation. In striate cortex, this includes increasing sharpness of disparity tuning (Menz and Freeman 2003, 2004), time-dependent changes in the contrast response function (Müller et al. 2001), increasing orientation discriminability (Zohary et al. 1990), and time-dependent changes in orientation selectivity (Ringach et al. 1997; Xing et al. 2005). In visual area, V2, Hegdé and Van Essen (2004) showed increased shape selectivity in single units of awake, fixating macaques during brief presentation of a series of different shape stimuli. The various types of response selectivities are generated in cortex, and it is generally assumed that the dynamics of these selectivities are generated by cortical circuits as well (Frazor et al. 2004; Suder et al. 2002; Wörgötter et al. 1998).
Temporal changes in the receptive field structure of lateral geniculate nucleus (LGN) neurons are less well studied. Early studies were mainly done with the response-plane technique (Stevens and Gerstein 1976). More recent studies have mainly used reverse correlation and spike-triggered averaging techniques capable of revealing linear aspects of stimulus-response coupling (DeBoer and Kuyper 1968). In these studies, small stimuli briefly flashed on different parts of the receptive field are used to approximate impulse stimuli needed for determination of the impulse-response function of the neuron. However, the previous studies have provided little information about dynamics of receptive field center width and spatial resolution of LGN neurons during short-lasting static stimulation for periods similar to those of natural fixations. The response-plane studies (Bullier and Norton 1979; Stevens and Gerstein 1976) focused mainly on classification of different types of LGN neurons, and the reverse correlation studies mainly on intrinsic temporal dynamics of the receptive field of the neurons and the question of space-time separability (Cai et al. 1997; Eckhorn et al. 1993; Golomb et al. 1994; Menz and Freeman 2004; Reid et al. 1997; Wolfe and Palmer 1998). Menz and Freeman (2004) estimated the changes in receptive field center width during static stimulation, but changes could only be estimated within a narrow range of time delays because of the short duration of the initial phase of the response in their conditions.
We studied the temporal response pattern and dynamics of receptive field structure in single LGN neurons using static spot stimuli flashed on the receptive field for 400–500 ms. We estimated spatial receptive field parameters from spatial summation curves determined for successive 5-ms intervals throughout the stimulus period. Thereby we could study dynamics of the response properties during periods similar to those in natural fixations, and with a method that, contrary to the reverse correlation methods (DeBoer and Kuyper 1968), does not presuppose a linear system. The results showed pronounced changes in the receptive field structure during the spot presentation. Initially, the neurons had wide receptive field centers. The center rapidly shrank to a minimum that occurred on average 70 ms after stimulus onset whereupon the center widened slightly. Thus the maximum spatial resolution occurred in a brief time window after onset of stimulation. In parallel, the center-surround antagonism increased. The changes in spatial resolution did not follow the changes of firing rate. The initial strong burst of action potentials appeared earlier than the maximal spatial resolution. These results are consistent with the hypotheses that the firing pattern of the neurons during brief static stimulation initially mediates a strong but spatially coarse message to cortex that gradually changes into a weaker, but spatially more precise message. This property of the LGN neurons may at least partly be the basis for the dynamics of feature selectivity in cortical neurons.
|
|
METHODS |
|---|
|
). Atropine and phenylephrine were applied to both eyes to dilate the pupils and retract the nictitating membranes. Contact lenses were fitted to focus the eyes on a video monitor 1.14 m in front of the cat's eyes.
Extracellular recordings of action potentials or combined recordings of action potentials and S-potentials (Bishop et al. 1958; Cleland et al. 1971; Hubel and Wiesel 1961; Kaplan and Shapley 1984; Ruksenas et al. 2000) from single units in the A-laminae of LGN were made with glass-insulated tungsten electrodes (Levick 1972; exposed tip, 6–10 µm) or with glass pipettes filled with 0.9% NaCl (15–25 M
in vivo) inserted through a craniotomy over the left hemisphere at H-C coordinates anterior 6.0 mm, lateral 15.0 mm, and with an angle of 32° from the vertical in a coronal plane. After isolation of action potentials from a single neuron, its receptive field center was plotted with hand-held stationary or moving light and dark spots, as well as grating stimuli. The neurons were classified as X or Y and lagged or nonlagged as described previously (Hartveit and Heggelund 1993).
For quantitative studies, we recorded responses to visual stimuli presented on a computer-controlled and gamma-corrected video monitor in front of the cat's eyes. First, the center of the receptive field was determined with a narrow, flashing slit (bright slits for on-center cells and dark slits for off-center cells) presented in different positions across the receptive field along the horizontal and the vertical axis. With this centering we presented a series of circular spot stimuli of stepwise increasing diameters; each spot was presented for 400–500 ms. The spot stimuli were luminance increments above (on-center cells) or decrements below (off-center cells) a constant, uniform background. The diameter of the spots varied from smaller than the receptive field center to wider than the whole receptive field. The contrast was fixed throughout the recordings for a given neuron, and contrast and background luminance were adjusted for each neuron such that the spots evoked a clear response but with maximal response well below response saturation for the neuron. Contrast was defined as (Lspot – Lbkg)/(Lspot + Lbkg), where Lspot is the luminance of the spot, and Lbkg is the luminance of the background. The contrast range for on-center cells was from 0.04 to 0.43. For off-center cells, the range was from –0.11 to –0.5 except for three cells (–0.67, –0.8, and –0.9). The range of background luminance was 10–65 cd/m2 except for recordings from one of the neurons (95 cd/m2). The various spots were presented interleaved such that each spot in the series was presented once whereafter the whole series of spots was repeated 
200 times. Before presentation of each spot, there was a 250-ms period for recording of spontaneous activity, and after each spot presentation, there was a pause of 1,500 ms to avoid sequence effects. A peristimulus time histogram (PSTH) with 5-ms bin width was determined for the response to each spot size. To estimate temporal changes in the receptive field, we made a time slice through the corresponding bins of all histograms for each 5-ms interval after stimulus onset (Fig. 1A). From the set of response versus spot width values in each time slice, we plotted a spatial summation curve (Fig. 1, B and C; Ruksenas et al. 2000). After smoothing the curve with equally moving average through three adjacent points, three spatial receptive field parameters were estimated from the curve: the width of the spot that elicited maximum response was taken as estimate of the center width; the width of the spot just large enough to give minimum response was taken as estimate of the surround width; and the difference of response to the spot that just filled the center and the one that just filled the whole receptive field was used to estimate center-surround antagonism (Fig. 1C). Center-surround antagonism was defined as the ratio between this difference and the center response (Fjeld et al. 2002; Ruksenas et al. 2000). The dynamics of the parameters were determined from their changes during the series of time slices throughout the stimulus period.
|
At the end of the experiment, the animal was deeply anesthetized with pentobarbital sodium (50 mg/kg, iv) and perfused transcardially with saline followed by 4% formaldehyde in saline. Brain blocks were removed, sectioned, and Nissl stained for histological verification of electrode positions.
|
|
RESULTS |
|---|
|
). For each neuron, we recorded the responses to the series of spots (light for on- and dark for off-center cells) in which spot width was stepwise increased from considerably smaller than the receptive field center to larger than the whole receptive field and where each spot was presented on the receptive field for 400–500 ms. For all neurons we made quantitative analyses of temporal changes of response properties during the spot presentation based on the time slicing across the histograms for the response to the different spot sizes and the spatial summation curve for each time slice (see METHODS and Fig. 1). Gradual increase in selectivity for spot size during brief spot stimulation
For all the nonlagged cells, there was a pronounced change in the selectivity for spot size during the spot-on period. At the beginning of the period, the neurons responded well to a broad range of spot sizes, but subsequently, the response was restricted to a gradually narrower range of the smaller spots. This is shown in Fig. 2A by a color map image of the response (z-axis) of a nonlagged cell to the set of spot widths (y-axis) plotted against time after stimulus onset (x-axis). Notice that initially this neuron gave a clear response even to the largest spot (14°), but at 60 ms, the response was mainly limited to a narrow range of small spots. Moreover, for the smaller spots, the latency to peak response increased with decreasing spot size. This gradual change in the response pattern is more clearly seen in Fig. 2B, where the results are shown on a finer time scale. Similar results from four other nonlagged cells are shown in Fig. 3.
|
|
12° occurred with slightly shorter latency than the response to the largest spot tested (14°). For the neurons shown in Figs. 3, B–D, the shift in latency occurred at 10°. The neuron shown in Fig. 3A was exceptional in this respect because it hardly responded to the larger spots at all. Dynamics of spatial resolution during single-unit response to brief spot stimuli
The quantitative estimates of the width of the receptive field center revealed pronounced changes during the stimulus presentation in all the nonlagged cells, in particular during the first 150 ms after spot onset. The receptive field center was initially wide, but rapidly shrank to a minimum. In the majority of neurons, the center thereafter widened again such that the minimal size occurred only briefly. This is shown by the plots of receptive field center width against time after stimulus onset in Figs. 2 and 3.
Figure 4 summarizes the data for all nonlagged cells. In Fig. 4A, the maximum (initial) center width for each neuron is plotted against the minimum center width. For all neurons, there was a dynamic shrinkage of the center. On average, the initial field center was 4.8 ± 3.3 (SD) times wider (P < 0.001) than the minimum center width. The timing of minimum center width varied between the neurons as shown in Fig. 7A. On average, the minimum occurred 69 ± 15 ms after stimulus onset. For the majority of neurons, the center widened again toward the more or less steady-state tail response. In Fig. 4B, the center width when a steady-state structure was reached (250 ms) is plotted against the minimum center width for each neuron. The average of the ratios between the width at 250 ms and at minimum width was 2.0 ± 1.1 (P < 0.001). This ratio was larger for Y-cells (2.5 ± 1.3) than for X-cells (1.6 ± 0.6, P = 0.008).
|
|
In addition to the dynamics of the center width, we attempted to estimate the dynamics of the width of the receptive field surround and the center-surround antagonism (see METHODS). However, in many neurons, it was difficult to determine the surround width with sufficient accuracy because of little change in the response with increasing spot width for the larger spot sizes and weak response to the larger spots in many cases. We were therefore unable to verify whether or not there was shrinkage in the surround width during the stimulus presentation. The problems with accurate estimation of surround width also led to a higher uncertainty for the estimates of center-surround antagonism. Nevertheless, it was clear that the center-surround antagonism in the majority of neurons was clearly present already at the start of the visual response and that the antagonism increased during the initial response in parallel with the decreasing width of the receptive field (Fig. 4C).
By defining spatial resolution as the inverse of the center width (Livingstone and Hubel 1981), our results showed that the spatial resolution conveyed to visual cortex by the single neuron improved markedly during the stimulus period. To check the changes of spatial resolution during the early visual response in a more traditional way, we recorded (n = 4) the response to stationary, flashing (1 cycle/s) square-wave gratings differing in spatial frequency (0.23–1.7 cycles/°). At the lowest frequency tested, the bar width of the grating was close to the maximum center width determined in the spot experiment, and at the highest frequency, the bar width was smaller than the minimum center width. The time-course of the response varied with changes of spatial frequency (Fig. 5) and, as in cortical neurons (Bredfeldt and Ringach 2002; Frazor et al. 2004), the latency to peak response generally increased with increasing spatial frequency. These results are consistent with the hypothesis that the spatial resolution rapidly changes from coarse to fine during the visual response. When the response to the coarsest grating (Fig. 5B) peaked, the response to the finest grating (Fig. 5E) had just started to appear, and at the time of peak response to the finest grating, the response to the coarsest grating had dropped to less than half-maximal response.
|
) are inhibited during the early period when the transient response to a spot stimulus occurs in nonlagged cells (Heggelund and Hartveit 1990). This is the period when the major shrinkage of the receptive field center occurred in the nonlagged cells. Accordingly, one might expect a less pronounced center shrinkage in lagged than in nonlagged cells. This was confirmed by results from the four lagged cells recorded in this study. As shown by the example in Fig. 6, the lagged cells showed no clear dynamic changes in the spatial structure of the receptive field. When the response appeared (Fig. 6C), the neuron already had a small receptive field center.
|
The changes in the firing rate during the spot stimulation of the nonlagged cells had a time-course that differed from the changes of the center width (Figs. 2 and 3). In the beginning of the visual response, when the firing rate was low, the center width was maximal, and as the firing rate increased in the initial transient burst of action potentials, the center width decreased. Furthermore, in nearly all neurons, the firing rate peaked before the minimal receptive field center was reached (Figs. 2 and 3). The average of the time differences between the occurrence of peak firing rate and minimum center width was 19 ± 15 ms (P < 0.001, paired t-test). On average, the field center at peak firing rate was 1.9 ± 0.8 times wider than the minimum center width (P < 0 001; cf. Fig. 7B). The average response dropped from a peak of 249 ± 121 to 138 ± 75 spikes/s (P < 0.001, paired t-test) at the time of minimum center width. This dissociation between the changes of firing rate and center width shows that the changes of center width were not simply "iceberg effects" of changing response intensity.
Sample of activated neurons shrinks during brief stimulation periods
The initially wide receptive field centers suggest that a small spot will initially activate a sample of neurons with receptive field centers distributed over a relatively large retinal area and that the sample of neurons, and consequently the retinal area covered by their field centers, will be gradually reduced as the field centers shrink during the stimulation. Thus a small spot will initially activate many neurons, most of them only transiently, followed by a sustained response in a selection of neurons that still have the stimulus within their field center (schematically shown in Fig. 8A). To test this hypothesis, we stimulated neurons (n = 4) with spots that were not centered on the receptive field. This is presumably the stimulus conditions for neurons with receptive fields located adjacently to the receptive field of the recorded neuron. We used a spot (1.9° wide) positioned at three different locations outside the minimum receptive field center (radius 0.8° for the neuron of Fig. 8). For all three positions, there was an initial transient response, but the subsequent sustained response that occurs to a small spot in the minimum receptive field center was lacking in these cases (Fig. 8, B–D).
|
|
Dynamics of receptive field structure in the input from retina
To compare the dynamics of receptive field structure between LGN neurons and their retinal input, we recorded S-potentials in addition to action potentials for 16 nonlagged cells. S-potentials are putatively postsynaptic potentials in the relay neurons generated by single action potentials in the retinal afferents, and by comparing the frequency of action potentials with the frequency of S-potentials, it is possible to estimate the changes of response that occur at the retinogeniculate relay (Cleland et al. 1971; Hubel and Wiesel 1961; Kaplan and Shapley 1984; Ruksenas et al. 2000). The initial receptive field center of the retinal input to a LGN neuron has to be at least as wide as the initial center of the LGN neuron, but could be even wider. Previous studies have shown that there are only minor differences in the center width between an LGN neuron and its retinal input when the response is summated over several hundred milliseconds (Hartveit et al. 1993; Mastronarde 1987b; Ruksenas et al. 2000). Thus the center width at steady state is unlikely to differ between a LGN neuron and its retinal input, but the minimum center width during the dynamic changes might differ.
We found no significant differences between the LGN neurons and their retinal input with respect to initial, minimum, or steady-state width of the receptive field center. However, the reduction of center size occurred faster in the LGN neuron than in its retinal input, consistent with the notion that the temporal response pattern is sharpened at the retinogeniculate relay. Furthermore, the two components in the initial wide-range response, which differ with respect to latency, were seen in the retinal input as well.
Figure 10 shows results from one neuron. The color map images in Fig. 10A show results for the S-potentials and the ones in Fig. 10D show results for action potentials. Clearly the fast initial shrinkage of the center was present already in the retinal input. Figure 10, B and E, shows that the degree of shrinkage was about the same for the retinal input and the LGN neuron. For this neuron, the initial wide-range response in the retinal input consisted of two parallel bands (Fig. 10A). In the LGN-neuron response, this initial wide-range response was strongly attenuated (Fig. 10D), and the band with the longer latency was presumably attenuated to the degree that it was not apparent in the plot. Moreover, for this neuron, the increase of center width toward the steady-state response was more pronounced in the retinal input (Fig. 10B) than in the LGN neuron response (Fig. 10E), but this was not representative for all neurons (Fig. 11B). Figure 10, C and F, shows the firing rate to the optimal spot sizes as function of time during stimulation. The peak firing rate of the retinal input (Fig. 10C) and the neuron response (Fig. 10F) was about the same, but the well-known lower sustained response in LGN neurons compared with the retinal input (Cleland et al. 1971; Hubel and Wiesel 1961) was clearly noticeable.
|
|
|
|
DISCUSSION |
|---|
|
70 ms. Thereafter, the center size increased in most neurons such that minimum center size occurred only briefly. Parallel to the decreasing center size, the center-surround antagonism increased, and in most neurons maximal antagonism was attained when the center size had reached minimum. Because the width of the receptive field center presumably is inversely related to the spatial resolution (Livingstone and Hubel 1981; Wörgötter et al. 1998), there was a qualitative change from coarse-to-fine spatial resolution in the information transmitted to cortex by the single neuron. The major part of the initial transient response occurred before the maximal spatial resolution, and at peak firing rate, the resolution was still clearly suboptimal. Thus the major part of the initial transient response contains visual information of low spatial resolution, and near the end of this transient, the response contains spatial information with maximal resolution. Because our spot stimuli were centered on the receptive field, it could be argued that we overestimated the width of the field centers. In particular, it could be argued that the outer parts of the initially wide centers were not real parts of the center in the sense that suprathreshold responses could be elicited without simultaneous stimulation of the core part of the center. The experiments with ring stimuli and eccentric spots, which selectively stimulated these outer parts, showed that this was not the case. On the contrary, the results showed that a fast response was elicited at these outer parts and that the response stopped as the field center shrank.
The fact that the neurons initially respond to stimuli relatively far outside their steady-state receptive field center implies that a small spot may initially activate neurons with receptive fields centered over a relatively large part of retina. Thus in the initial part of the stimulation, there is not only a change from coarse-to-fine in the information transmitted to cortex by the single neuron, but also in the sample of neurons responding to the stimulus. The sample of responding neurons will rapidly shrink, such that most of the neurons give only a brief transient response. Only a selection of these neurons, which have properly centered receptive field, will continue to respond, and they will transmit information with maximal spatial resolution.
The lagged cells lack the initial transient response component (Mastronarde 1987a) because the initial transient in the retinal input is blocked by intrageniculate feedforward inhibition (Heggelund and Hartveit 1990). Here we showed that lagged cells have maximal spatial resolution already from the beginning of their response. This is consistent with the hypothesis that lagged cells are specifically linked to pattern analysis, whereas nonlagged cells also seem to play a key role in detection of targets (Hartveit and Heggelund 1992). This hypothesis was suggested because lagged cells lack of the fast and strong initial transient response and respond in a graded manner to contrasts over a wider range than nonlagged cells (Hartveit and Heggelund 1992).
Reported dynamics of receptive field size in striate cortex during static stimulation (Suder et al. 2002; Wörgötter et al. 1998) were ascribed to mechanisms at the cortical level because only minor changes were seen in LGN. In cortex, the width of subregions in the receptive fields shrank by an average of 2° during 300-ms static spot stimulation, whereas in LGN, an average shrinkage of only 0.2° was found (Suder et al. 2002). It was suggested that the dynamic sizing of receptive field subregions in cortex was caused by synaptic facilitation of geniculocortical input during the initial transient response of the LGN neurons. Thereby, otherwise subthreshold excitatory inputs in peripheral parts of the cortical receptive field could become suprathreshold for the duration of the transient response, causing a brief widening in the field (Suder et al. 2002). Our data, which were obtained with higher temporal and spatial resolution than those of Suder et al. (2002), show changes in the width of receptive field centers in LGN that are large enough to explain the changes observed in cortex (Suder et al. 2002; Wörgötter et al. 1998). In fact, the average differences between the maximum and the minimum width we found (2.9 ± 2.2°, P < 0.001, paired t-test) was slightly larger than the average shrinkage of the subfields in cortex; a discrepancy that could be caused by differences of eccentricity of the sampled neurons. Furthermore, in LGN, we found the widest field centers at the beginning of the response when the firing rate was low, rather than at the peak firing rate. Therefore we suggest that the dynamics of receptive field width in cortex reflect the dynamic shrinkage of field centers in LGN neurons rather than a brief widening caused by the high firing rate in the transient response component in the geniculate input. Moreover, the time-course of the shrinkage in the cortical receptive fields (fast shrinkage after the 1st 50 ms of the response; Wörgötter et al. 1998) is similar to the time-course we found in LGN.
The shrinkage of the receptive field center and increase of spatial resolution during static stimulation were not mentioned in most of the previous studies in LGN (Bullier and Norton 1979; Cai et al. 1997; Eckhorn et al. 1993; Golomb et al. 1994; Reid et al. 1997; Stevens and Gerstein 1976; Wolfe and Palmer 1998). Nevertheless, shrinkages of receptive field center can be seen in figures in some of the papers (Stevens and Gerstein 1976), suggesting that the phenomenon may not have been noticed or considered to be without importance. However, Menz and Freeman (2004) addressed this phenomenon and estimated the degree of shrinkage. Because they could only estimate these changes within a rather short time interval, it is difficult to compare the values they found with our results.
The spatiotemporal response characteristics of visual neurons have been extensively studied (for reviews, see Albrecht et al. 2003; Frishman et al. 1987; Shapley and Lennie 1985). In most studies, the stimulus was modulated on the receptive field for longer periods (typically several seconds) to approximate a steady-state condition and minimize initial transient effects. However, in natural saccadic inspections transient response components presumably play a key role. Accordingly, Frazor et al. (2004) used static grating stimuli presented briefly (200 ms) on the receptive field of neurons in striate cortex to study dynamics of optimal spatial frequency during periods similar to those of natural fixations. They found increasing optimal spatial frequency during the stimulation, consistent with the dynamics of receptive field width (Suder et al. 2002; Wörgötter et al. 1998). Frazor et al. (2004) suggested that the changes of optimal spatial frequency were generated in cortex. However, the changes of optimal spot diameter we found, as well as our experiments with static gratings, suggest that the dynamics of spatial resolution in cortex is caused by dynamics in the geniculate input. Frazor et al. (2004) found a gradual increase of optimal spatial frequency over 30 ms after response onset, and the response latency seemed to be 30 ms. This is about the same time-course as we found for the decrease in optimal spot size. Moreover, we found a similar increase of latency to peak response with decreasing spot width (Figs. 2, A and B, and 3, A–D) and increasing spatial frequency (Fig. 4) as Frazor et al. (2004) found in cortex.
The dynamics of the more specialized response properties generated in the cortical circuits might also to a certain degree be caused by the coarse-to-fine changes in the geniculate input to cortex. Although differences of stimulus conditions, experimental designs, and differences of species studied complicate a direct comparison, there are examples of striking similarities between the dynamics at the geniculate and cortical levels. Orientation selectivity in V1 appears after a delay of 30–45 ms (Ringach et al. 1997), and maximum selectivity is reached after 60 ms (Xing et al. 2005). In area V2, of awake, fixating macaques Hegdé and Van Essen (2004) showed increased shape selectivity in single units during brief (300 ms) presentation of different shape stimuli. The neurons responded unselectively to most shapes in an early transient response 40–60 ms after stimulus onset, whereas the subsequent, weaker response showed clear shape selectivity. In general, most of the dynamic coarse-to-fine changes in cortex seem to occur within the first 150 ms after stimulus onset (Müller et al. 2001), like the dynamics we found for spot width selectivity and spatial resolution in LGN.
A brief static stimulus has been regarded as a condition that in many respects resembles the stimulus conditions that occur naturally during fixations in saccadic vision (Frazor et al. 2004). In this perspective, it may seem puzzling that the response we found at the time with maximum spatial resolution was relatively weak in several of the neurons. However, in natural visual inspections, the visual response in LGN neurons to the brief visual stimulation in intersaccadic intervals seems to be modulated by saccade-related signals (Lee and Malpeli 1998; Reppas et al. 2002). Before a saccade, there is a suppression of the visual response that in the cat peaks 100 ms before the start of the saccade and smoothly reverse to facilitation by the end of the saccade. The facilitation peaks 70–130 ms after the end of the saccade (Lee and Malpeli 1998). This suggests that the strongest postsaccadic enhancement coincides with the time of maximal spatial resolution after stimulus onset. Thereby, the response of the LGN neurons at the time of maximal resolution will probably be stronger in normal viewing conditions than in our experimental conditions. Moreover, the suppression of visual response during preparation of the next saccade probably attenuates the response in the later part of a fixation period. That could mean attenuation in the part of the response pattern that appeared as a steady state in our experiments. This could further accentuate the signal during the time window with maximal spatial resolution.
Our experiments with combined S-potential and action potential recordings showed that the initial fast shrinkage of the receptive field center was present already in the retinal input to LGN neurons. The degree of shrinkage was similar for the retinal input and the LGN neuron, and apart from the faster shrinkage in the LGN neurons, the temporal pattern of the shrinkage was also similar. Mechanisms for generation of the dynamics of center width in retinal ganglion neurons are unclear. Several previous studies have shown that spatiotemporal characteristics of X-cell responses in the cat retina to drifting or sinusoidally modulated grating patterns can be fitted by the receptive field model of Rodieck (1965) by assuming that the surround mechanism has slightly longer latency (0–3.6 ms, Derrington and Lennie 1982; 1.2–7.7 ms, Enroth-Cugell et al. 1983) than the center mechanism. Corresponding modeling of Y-cells is severely limited by the pronounced nonlinearities of these neurons (Frishman et al. 1987). A lag of surround suppression in X-cells is one factor that could contribute to the dynamics of center width in these neurons, but the short duration of the estimated lag suggests that other factors are involved as well, such as differences between center and surround mechanisms with respect to time-course of response. Direct experimental studies of such possible factors are hampered by the fact that the neural circuits mediating the inhibitory surround are still not well characterized (Demb et al. 2001; Flores-Herr et al. 2001; Kamerans and Spekreijse 1999; McMahon et al. 2004). The two components we observed in the initial wide range response suggest that retinal mechanisms with different latencies are involved in the generation of the transient response. One possibility is that the fastest component with the narrower spatial range is related to lateral summation in the outer plexiform layer, whereas the slightly slower component with the wider spatial range is related to transient spread of excitation in the inner plexiform layer. However, our experiments on retinal input to LGN were based on recordings of S-potentials rather than on paired recordings from connected retinal ganglion cells and LGN neurons. Therefore we cannot exclude the possibility that the shrinking receptive field centers we found in the retinal input were at least partly caused by a change in the number of retinal afferents contributing to the LGN neuron response. Physiological data from paired recordings from a LGN neuron and synaptically connected retinal ganglion cells suggest that only a few retinal ganglion cells converge on a single LGN neuron, and some LGN neurons are dominated by input from a single retinal ganglion cell (Cleland and Lee 1985; Dubin and Cleland 1977; Mastronarde 1987b, 1992). This conclusion was based on cross-correlation of responses to a visual stimulation lasting several hundred milliseconds. Thus although the response of a LGN neuron was dominated by input from a few retinal ganglion cells through the major part of the stimulation period, it could well be that a somewhat higher number of inputs contributed in the initial part of the visual response. Anatomical data on degree of convergence of retinal fibers on single LGN neurons are sparse (Hamos et al. 1987; Robson 1993), but for Y-cells, data suggest convergence of multiple inputs (>10) on the single neuron (Robson 1993).
The initial strong transient firing rate that occur in the nonlagged cells may serve an essential function in fast and efficient detection of visual targets, but also in alerting and fine-tuning of potential neuronal circuits involved in the subsequent detailed analysis of the visual features of the target. In this connection, it gives sense to initially alert and tune in neurons over relatively large retinotopic areas before the neurons optimally suited for analysis of the target are selected. The early, strong response mediates spatial information with coarse resolution and is followed by weaker response that mediates spatial information with maximal resolution. Thus the transmission could be described as a dual-phase process: an initial strong alerting and preparing signal, followed by transmission of the precise message.
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
Present address of A. Bulatov: Kaunas Medical University, Department of Biology, LT44307 Kaunas, Lithuania.
|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: P. Heggelund, Univ. of Oslo, Inst. of Basic Medical Sciences, Dept. of Physiology, PO Box 1103 Blindern, N-0717 Oslo, Norway (E-mail: paul.heggelund{at}medisin.uio.no)
|
|
REFERENCES |
|---|
|
Bishop PO, Burke W, Davis R. Synapse discharge by single fibre in mam-malian visual system. Nature 182: 728–730, 1958.[Medline]
Bredfeldt CE, Ringach DL. Dynamics of spatial frequency tuning in macaque V1. J Neurosci 22: 1976–1984, 2002.
Bullier J, Norton TT. X and Y relay cells in cat lateral geniculate nucleus: quantitative analysis of receptive field properties and classification. J Neurophysiol 42: 244–273, 1979.
Cai D, DeAngelis GC, Freeman RD. Spatiotemporal receptive field organization in the lateral geniculate nucleus of cats and kittens. J Neurophysiol 78: 1045–1061, 1997.
Cleland BG, Dubin MW, Levick WR. Sustained and transient neurones in the cat's retina and lateral geniculate nucleus. J Physiol 217: 473–496, 1971.
Cleland BG, Lee BB. A comparison of visual responses of cat lateral genic-ulate nucleus neurones with those of ganglion cells afferent to them. J Physiol 369: 249–268, 1985.
DeAngelis GC, Ohzawa I, Freeman RD. Spatiotemporal organization of simple-cell receptive fields in the cat striate cortex. 1. General characteristics of postnatal development. J Neurophysiol 69: 1091–1117, 1993.
DeBoer E, Kuyper P. Triggered correlation. Trans IEEE Biomed Eng 15: 169–179, 1968.
Demb JB, Zaghloul K, Haarsma L, Sterling P. Bipolar cells contribute to nonlinear spatial summation in the brisk-transient (Y) ganglion cell in mammalian retina. J Neurosci 21: 7447–7454, 2001.
Derrington AM, Lennie P. The influence of temporal frequency and adaptation level on receptive field organization of retinal ganglion cells in cat. J Physiol 333: 343–366, 1982.
Dubin MW, Cleland BG. Organization of visual inputs to interneurones of lateral geniculate nucleus of the cat. J Neurophysiol 40: 410–427, 1977.
Eckhorn R, Krause F, Nelson JJ. The RF-cinematogram - a cross-correlation technique for mapping several visual receptive fields at once. Biol Cybern 69: 37–55, 1993.[CrossRef][Web of Science][Medline]
Enroth-Cugell C, Robson JG, Schweizer-Tong DE, Watson AB. Spatio-temporal interaction in cat retinal ganglion cells showing linear spatial summation. J Physiol 341: 279–307, 1983.
Evinger C, Fuchs AF. Saccadic, smooth pursuit, and optokinetic eye movements of the trained cat. J Physiol 285: 209–229, 1978.
Fjeld IT, Ruksenas O, Heggelund P. Brainstem modulation of visual response properties of single cells in the dorsal lateral geniculate nucleus of cat. J Physiol 543: 541–554, 2002.
Flores-Herr N, Protti DA, Wässle H. Synaptic currents generating the inhibitory surround of ganglion cells in the mammalian retina. J Neurosci 21: 4852–4863, 2001.
Frazor RA, Albrecht DG, Geisler WS, Crane AM. Temporal dynamics of the spatial frequency response function. J Neurophysiol 91: 2607–2627, 2004.
Frishman LJ, Freeman AM, Troy JB, Schwitzer-Tong DE, Enroth-Cugell C. Spatiotemporal frequency responses of cat retinal ganglion cells. J Gen Physiol 89: 599–628, 1987.
Golomb D, Kleinfeldt D, Reid RC, Shapley RM, Shraiman BI. On temporal codes and the spatiotemporal response of neurons in the lateral geniculate nucleus. J Neurophysiol 72: 2990–3003, 1994.
Hamos JE, Van Horn SC, Raczkowski D, Sherman SM. Synaptic circuits involving an individual retinogeniculate axon in the cat. J Comp Neurol 259: 165–192, 1987.[CrossRef][Web of Science][Medline]
Hartveit E, Heggelund P. The effect of contrast on the visual response of lagged and nonlagged cells in the cat lateral geniculate nucleus. Vis Neurosci 9: 515–525, 1992.[Web of Science][Medline]
Hartveit E, Heggelund P. Brain-stem influence on visual response of lagged and nonlagged cells in the cat lateral geniculate nucleus. Vis Neurosci 10: 325–339, 1993.[Web of Science][Medline]
Hartveit E, Ramberg SI, Heggelund P. Brainstem modulation of spatial receptive field properties in the dorsal lateral geniculate nucleus of cat. J Neurophysiol 70: 1644–1655, 1993.
Hegdé J, Van Essen DC. Temporal dynamics of shape analysis in macaque visual area V2. J Neurophysiol 92: 3030–3042, 2004.
Heggelund P, Hartveit E. Neurotransmitter receptors mediating excitatory input to cells in the cat lateral geniculate nucleus. I. Lagged cells. J Neurophysiol 63: 1347–1360, 1990.
Hubel DH, Wiesel TN. Integrative action in the cat's lateral geniculate body. J Physiol 155: 385–398, 1961.
Kamermans M, Spekreijse H. The feedback pathway from horizontal cells to cones. Vision Res 39: 2449–2468, 1999.[CrossRef][Web of Science][Medline]
Kaplan E, Shapley RM. The origin of the S (slow) potential in the mammalian lateral geniculate nucleus. Exp Brain Res 55: 111–116, 1984.[Web of Science][Medline]
Knierim JJ, Van Essen DC. Neuronal responses to static texture patterns in area V1 of the alert macaque monkey. J Neurophysiol 67: 961–980, 1992.
Lee D, Malpeli JG. Effects of saccades on the activity of neurons in the cat lateral geniculate nucleus. J Neurophysiol 79: 922–936, 1998.
Levick WR. Another tungsten microelectrode. Med Biol Eng 10: 510–515, 1972.[CrossRef][Web of Science][Medline]
Livingstone MS, Hubel DH. Effects of sleep and arousal on the processing of visual information in the cat. Nature 291: 554–561, 1981.[CrossRef][Medline]
Mastronarde DN. Two classes of single-input X-cells in cat lateral geniculate nucleus. I. Receptive field properties and classification of cells. J Neurophysiol 57: 357–380, 1987a.
Mastronarde DN. Two classes of single-input X-cells in cat lateral geniculate nucleus. II. Retinal inputs and generation of receptive-field properties. J Neurophysiol 57: 381–413, 1987b.
Mastronarde DN. Nonlagged relay cells and interneurons in the cat lateral geniculate nucleus: receptive field properties and retinal inputs. Vis Neurosci 8: 407–441, 1992.[Web of Science][Medline]
McMahon MJ, Packer OS, Dacey DM. The classical receptive field surround of primate parasol ganglion cells is mediated primarily by a non-GABAergic pathway. J Neurosci 24: 3736–3745, 2004.
Menz MD, Freeman RD. Stereoscopic depth processing in the visual cortex: a coarse-to-fine mechanism. Nat Neurosci 6: 59–65, 2003.[CrossRef][Web of Science][Medline]
Menz MD, Freeman RD. Temporal dynamics of binocular disparity processing in the central visual pathway. J Neurophysiol 91: 1782–1793, 2004.
Müller JR, Metha AB, Krauskopf J, Lennie P. Information conveyed by onset transients in responses of striate cortical neurons. J Neurosci 21: 6978–6990, 2001.
Reid RC, Victor JD, Shapley RM. The use of m-series in the analysis of visual neurons: linear receptive field properties. Vis Neurosci 14: 1015–1027, 1997.[Web of Science][Medline]
Reppas RC, Usrey WM, Reid RC. Saccadic eye movements modulate visual responses in the lateral geniculate nucleus. Neuron 35: 961–974, 2002.[CrossRef][Web of Science][Medline]
Ringach DL, Hawken MJ, Shapley R. Dynamics of orientation tuning in macaque primary visual cortex. Nature 387: 281–284, 1997.[CrossRef][Medline]
Robson JA. Qualitative and quantitative analyses of the patterns of retinal input to neurons in the dorsal lateral geniculate nucleus of the cat. J Comp Neurol 334: 324–336, 1993.[CrossRef][Web of Science][Medline]
Rodieck RW. Quantitative analysis of cat retinal ganglion cell response to visual stimuli. Vision Res 5: 583–601, 1965.[CrossRef][Medline]
Rodieck RW, Pettigrew JD, Bishop PO, Nikara T. Residual eye movements in receptive field studies of paralysed cats. Vision Res 7: 107–110, 1967.[CrossRef][Web of Science][Medline]
Ruksenas O, Fjeld IT, Heggelund P. Spatial summation and center-surround antagonism in the receptive field of single units in the dorsal lateral geniculate nucleus of cat: comparison with retinal input. Vis Neurosci 17: 855–870, 2000.[CrossRef][Web of Science][Medline]
Shapley RM, Lennie P. Spatial frequency analysis in the visual system. Annu Rev Neurosci 8: 547–583, 1985.[CrossRef][Web of Science][Medline]
Stevens JK, Gerstein GL. Spatiotemporal organization of cat lateral geniculate receptive fields. J Neurophysiol 39: 213–238, 1976.
Stryker M, Blakemore C. Saccadic and disjunctive eye movements in cats. Vision Res 12: 2005–2013, 1972.[CrossRef][Web of Science][Medline]
Suder K, Funke K, Zhao Y, Kerscher N, Wennekers T, Wörgötter F. Spatial dynamics of receptive fields in cat primary visual cortex related to the temporal structure of thalamocortical feedforward activity. Experiments and models. Exp Brain Res 144: 430–444, 2002.[CrossRef][Web of Science][Medline]
Wolfe J, Palmer LA. Temporal diversity in the lateral geniculate nucleus of cat. Vis Neurosci 15: 653–675, 1998.[CrossRef][Web of Science][Medline]
Wörgötter F, Suder K, Zhao Y, Kerscher N, Eysel U, Funke K. State-dependent receptive field restructuring in the visual cortex. Nature 396: 165–168, 1998.[CrossRef][Medline]
Xing D, Shapley RM, Haeken MJ, Ringach DL. Effect of stimulus size on the dynamics of orientation selectivity in macaque V1. J Neurophysiol 94: 799–812, 2005.
Yarbus AL. Eye Movements and Vision. New York: Plenum Press, 1967.
Zisper K, Lamme VA, Schiller PH. Contextual modulation in primary visual cortex. J Neurosci 16: 7376–7389, 1996.
Zohary E, Hillman P, Hochstein S. Time course of perceptual discrimination and single neuron reliability. Biol Cybern 62: 475–486, 1990.[CrossRef][Web of Science][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
