HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 93/6/3537    most recent 00847.2004v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (14) Citing Articles via Google Scholar Google Scholar Articles by Weng, C. Articles by Alonso, J.-M. Search for Related Content PubMed PubMed Citation Articles by Weng, C. Articles by Alonso, J.-M.

Receptive Field Size and Response Latency Are Correlated Within the Cat Visual Thalamus

¹Department of Biological Sciences, State University of New York–State College of Optometry, New York, New York; and ²Department of Psychology, University of Connecticut, Storrs, Connecticut

Submitted 17 August 2004; accepted in final form 2 December 2004

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Each point in visual space is encoded at the level of the thalamus by a group of neighboring cells with overlapping receptive fields. Here we show that the receptive fields of these cells differ in size and response latency but not at random. We have found that in the cat lateral geniculate nucleus (LGN) the receptive field size and response latency of neighboring neurons are significantly correlated: the larger the receptive field, the faster the response to visual stimuli. This correlation is widespread in LGN. It is found in groups of cells belonging to the same type (e.g., Y cells), and of different types (i.e., X and Y), within a specific layer or across different layers. These results indicate that the inputs from the multiple geniculate afferents that converge onto a cortical cell (approximately 30) are likely to arrive in a sequence determined by the receptive field size of the geniculate afferents. Recent studies have shown that the peak of the spatial frequency tuning of a cortical cell shifts toward higher frequencies as the response progresses in time. Our results are consistent with the idea that these shifts in spatial frequency tuning arise from differences in the response time course of the thalamic inputs.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
The cat lateral geniculate nucleus (LGN) is organized into channels of information processing that have different temporal and spatial properties. On average, Y geniculate cells respond faster to visual stimuli and have larger receptive fields than X geniculate cells at the same eccentricity (Saul and Humphrey 1990; Sestokas and Lehmkuhle 1986; So and Shapley 1979; Troy and Lennie 1987; Yeh et al. 2003), and the same is true for Magnocellular and Parvocellular cells in primate (Croner and Kaplan 1995; Levitt et al. 2001; Maunsell et al. 1999; Usrey and Reid 2000; Xu et al. 2001). However, the differences in response latency between X and Y cells (or Magno and Parvo cells) are far from being clear-cut. For example, Troy and Lennie (1987) found that, although the average Y geniculate cell had a shorter latency (36.6 ms) than that of the average X geniculate cell (37.8 ms), the latency range within each category was very large (about 25 ms). Similarly, Maunsell et al. (1999) found that the distribution of response latencies for Parvocellular and Magnocellular neurons in primate overlapped extensively.

The differences in receptive field size between X and Y cells are more pronounced than the differences in response latency. So and Shapley (1979) indicated that, at a given eccentricity, it was almost always possible to categorize a cell as X or Y based on the size of its receptive field center (although the best categorization method was the linearity test; Shapley and Hochstein 1975). Specifically, the highest spatial frequency of a drifting grating that can effectively drive an X cell (modulate the first Fourier harmonic) was shown to be 2–3 times higher than the one that drives a Y cell (So and Shapley 1979), and the frequency ranges did not overlap much, at least within 5 deg of the area centralis [X cells: 0.3–1.8 cycles/deg; Y cells: 0.1–0.5 cycles/deg (Derrington and Fuchs 1979)]. [It should be noticed, however, that X and Y cells have the same spatial resolution when measured with contrast reverse gratings (comparing resolution of first Fourier harmonic in X cells to second Fourier harmonic of Y cells; So and Shapley 1979) because the X cell center size and the Y cell subunit size are both determined by the spatial resolution of the bipolar cell (Demb et al. 2001).] Hand-plotting measurements also indicate that the receptive field sizes of Y cells are 1.5–2 times larger than those of X cells [see Sherman and Spear 1982 for review; Saul and Humphrey (1990) estimated that the average receptive field diameter is 0.7 deg for an X cell (0.2–2.0 deg) and 1.58 deg for a Y cell (0.3–3.5 deg) within 35 deg of the area centralis].

Most measurements of receptive field size and response latency are somewhat limited by the fact that cells are usually recorded at different times, under different conditions, and in different animals. The level of anesthesia/arousal, eccentricity, individual variations like age, and the properties of the stimulus projected on the retina could affect the measurements (Cleland et al. 1979; Kremers et al. 2001; Maunsell et al. 1999; Swadlow and Weyand 1985; Worgotter et al. 1998). This variability is not trivial and has been demonstrated by some careful studies in the past. For example, by recording from 42 ON-center retinal ganglion cells in a single animal, Cleland et al. (1979) demonstrated that the variability in receptive field size was much smaller than previously thought. Similarly, Maunsell et al. (1999) showed that response latencies in LGN could significantly vary across individuals in relation with age and size (see also Maunsell and Gibson 1992 for V1 cells).

The variability in measuring conditions can be greatly diminished by comparing the response properties of pairs of X and Y geniculate cells that are simultaneously recorded and have overlapping receptive fields. Measurements of these X–Y cell pairs demonstrate that the average Y geniculate cell has a response latency 3.9 ms faster and a receptive field center 1.6 times larger than that of the average neighboring X geniculate cell (Yeh et al. 2003). However, even neighboring cells that are simultaneously recorded show great variation in receptive field size and response latency. What could be the reason for such variability? If the variability in receptive field size and response latency were correlated (i.e., the larger the receptive field, the faster the response), the cortex would receive initially a fast, low-resolution image that would become increasingly sharper over time. In this scenario, the variability in receptive field size and timing could help to smooth the transition from low to high spatial-resolution images. The results presented here are consistent with this hypothesis and also support recent findings showing that the peak of the spatial frequency tuning of a cortical cell shifts toward higher frequencies as the response progresses in time (Bredfeldt and Ringach 2002; Frazor et al. 2004; see also Mazer et al. 2002). Recently Frazor et al. (2004) modeled this shift in spatial frequency with a feedforward circuit in which X and Y cells (in cats) or Parvocellular and Magnocellular cells (in primates) converge onto the same cortical cell. The results presented here provide evidence for this model in both cortical cells that receive mixed X–Y inputs and those that receive input from only one cell type. Preliminary results have appeared in abstract form (Weng et al. 2002).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Surgery and preparation

Cats were initially anesthetized with ketamine (10 mg/kg, intramuscular) followed by thiopental sodium [20 mg/kg, intravenous (iv), supplemented as needed during surgery; and at a continuous rate of 1–2 mg kg^–1 h^–1, iv, during recording]. Lidocaine was administered topically or injected subcutaneously at all incisions or points of pain and pressure. The animal was then intubated and placed in a stereotaxic apparatus. We monitored the following vital signs: electrocardiogram, electroencephalogram, oxygen in blood (measured by pulse oximetry), expired carbon dioxide, heart rate, blood pressure (measured with a pressure cuff), and rectal temperature. A craniotomy and duratomy were made to introduce recording electrodes into LGN (anterior, 5.5; lateral, 10.5). Animals were paralyzed with Norcuron (0.2 mg kg^–1 h^–1, iv) to minimize eye movements and artificially ventilated to keep the expired CO₂ between 28 and 33 mmHg. Pupils were dilated with atropine sulfate (1%) and nictitating membranes retracted with Neo-Synephrine (2.5%). The eyes were covered with contact lenses to protect the corneas. The optic disk and the area centralis were projected onto a tangent screen located 114 cm from the cat (Pettigrew et al. 1979). All surgical and experimental procedures were performed in accordance with U.S. Department of Agriculture guidelines and were approved by the Institutional Animal Care and Use Committee at the University of Connecticut and at the State University of New York, State College of Optometry.

Electrophysiological recordings and data acquisition

A matrix of 7 independently moveable electrodes arranged circularly was introduced into the LGN (Eckhorn and Thomas 1993). The electrodes were very thin (80 µm rod; 25 µm at the shaft) and had impedances of 3–6 M (System Eckhorn Thomas Recordings, Marburg, Germany). A glass guide tube with an inner diameter of about 300 µm at the tip was attached to the shaft probe of the multielectrode to reduce the interelectrode distances to about 80–300 µm. The matrix of electrodes was then lowered into the brain, leaving the tip of the guide tube about 3 mm above the LGN. Each electrode was moved independently within the LGN and positioned either within layer A or layer C. The angle of the multielectrode was adjusted precisely for each experiment (about 25–30° anterior–posterior; about 2–5° lateral–central) to simultaneously record from cells with spatially overlapping receptive fields in both layers. All cells were recorded within 5–10° of the area centralis. Recorded voltage signals from all 7 electrodes were conventionally amplified, filtered, and passed to a computer running the Discovery software package (Datawave System, Longmont, CO). For each cell, spike waveforms were identified initially during the experiment and verified carefully off-line by spike-sorting analysis (Datawave System). Visual stimuli were generated with an AT-vista graphics card (Truevision, Indianapolis, IN) and shown on a 20-in. monitor (Nokia 445Xpro, Salo, Finland; frame rate: 128 Hz).

Receptive field mapping

Geniculate receptive fields were mapped by reverse correlation with white noise. The white-noise stimulus was derived from a binary m-sequence (Reid et al. 1997; Sutter 1987), and spatially consisted of a 16 x 16 grid of black or white squares (pixels). Each pixel in the white-noise stimulus was 0.81 deg² in size (0.9° x 0.9°) and each frame was presented for 15.5 ms (the entire white-noise sequence lasted about 508 s). From all the frames presented, we selected those that preceded a spike within a specific time delay. Then the selected frames were averaged and represented as a contour plot smoothed with a cubic spline, each contour line showing from center to periphery 100 to 20% of the maximum response (Matlab, The MathWorks, Natick, MA). The receptive field size was quantified as the number of contiguous pixels within the 20% contour line at the time delay that contained the maximum response (15 ms earlier than the latency of the surround response). The 20% contour line was used because it defines quite precisely the size of the geniculate center (measurements <20% are less accurate because they include responses from the surround and background noise). It should be emphasized that the receptive field size in this study refers exclusively to the receptive field center and not the surround. In all figures, ON-center receptive fields are shown as continuous lines and OFF-center receptive fields as discontinuous lines.

Time course of the visual response

The time course of the visual response was also calculated from responses to white-noise stimuli by reverse correlation. The impulse response was defined as the time course of the response evoked by the most effective stimulus pixel within the receptive field center (the pixel that generated the maximum response). This pixel is ideal for our measurements because it is located at the middle of the receptive field center and its response is least contaminated by surround inhibition. Impulse responses were normalized and smoothed with a cubic spline to compare timing differences between simultaneously recorded cells (Alonso et al. 2001; Yeh et al. 2003). A not-a-knot end condition was chosen to interpolate bins of 15.5 ms, from the –46.5-ms bin [–61 ms, –46.5 ms] to the 186-ms bin [170.5 ms, 186 ms] (spline function, Matlab). With respect to the figures shown herein, impulse responses are shown from the 0-ms bin [–15.5 ms, 0 ms] to the 186-ms bin [170.5 ms, 186 ms]. Most impulse responses were biphasic. For example, the impulse response of an ON-center geniculate cell had a positive first phase followed by a negative second phase. By convention, the first phase is positive for ON-center cells and negative for OFF-center cells. The peak time (maximum of the first phase) was measured from the smoothed impulse response. This method gave a reliable estimate of the response latency for each cell (i.e., latency of the cell's maximum response). We estimated the error in the measurements of peak time by calculating the impulse responses of 10 cells in 3 consecutive repetitions of white-noise stimulation. In 6 of these cells the values of peak time were identical in the 3 repetitions. In 4 other cells there was a difference of 1.55 ms between 2 of the repetitions. This measurement error is consistent with previous results from our laboratory. Yeh et al. (2003) was able to demonstrate that the average peak time of Y cells recorded from layer C is 2.5 ms faster than the average peak time of Y cells recorded from layer A. This 2.5-ms difference in peak time was highly significant (P < 0.001, Wilcoxon test).

Rank-order analysis

We used a rank-order analysis to measure the correlation between receptive field size and response latency for the entire population of X and Y cells recorded in this study. This analysis aimed to minimize the variability in receptive field size and response latency caused by the different experimental conditions in which the cells were recorded (i.e., different eccentricities and different animals). In this analysis, each cell within a simultaneously recorded group was assigned a rank value based on its receptive field size (the rank values ranged from 1 to n, where n is the number of simultaneously recorded cells). The cell with the largest receptive field was assigned a rank value of 1 and the cell with the smallest receptive field a rank value of n. Cells with the same receptive field sizes were treated as follows. First, they were assigned temporary numbers as if they had different sizes (e.g., 1 and 2 for two cells with the same receptive field size that was the largest in a group). They were then given a rank value obtained from the average of the temporary numbers (e.g., 1.5 and 1.5 for two cells with the same receptive field size that was the largest in a group). Because the value of n varies (e.g., some groups have 4 cells and others have 9 cells), we normalized the ranks by subtracting the mean rank value [(n + 1)/2] from each group. For example, a group with the rank values 1, 2, 3, 4, and 5 would be normalized as –2, –1, 0, 1, and 2, respectively.

Similarly, each cell within a simultaneously recorded group was assigned a rank value based on the peak time of the impulse response, which was 1 for the cell with the fastest response. For example, in a group of 3 cells, the cell with the shortest peak time had a rank of 1, the one with the longest peak time a rank of 3, and the one with the intermediate peak time a rank of 2. When 2 cells had impulse responses with the same peak times, we used the zero-crossing values to determine the rank (zero-crossing: zero value between first and second phases of impulse response). If the zero-crossing values were also the same, we used the rebound times (rebound: maximum of the second phase of the impulse response in absolute value), and if all these values were the same we used the same method described above for receptive field size. As for receptive field size, the ranks were normalized by subtracting the mean rank value.

Monte Carlo simulations were done to estimate the significance of the correlations obtained with the normalized rank-ordered data. As a first approach, we ran 1,000 simulations using a bootstrap method. In each bootstrap simulation, we took the entire sample of cells (e.g., 144 Y cells) and replaced random paired values of peak time/receptive field size with other paired values already present in the data sample. As a second approach, we ran 1,000 Monte Carlo simulations by randomly choosing n values in each simulation from the entire sample of cells recorded (n = 150 for all cells, n = 100 for Y cells, n = 60 for X cells). The correlation coefficients between peak time and receptive field size were calculated for each simulation.

Regression analysis

The strength of the correlation between peak time and receptive field size was measured by linear regression analysis. Statistical significance was assessed with a Pearson test in all cases except for the rank-order analysis in which we used a Spearman test. Multiple regression analysis was used to estimate the possible contribution from other variables to the correlations between peak time and receptive field size. The multiple regression analysis was done as follows. First, we selected all cell groups in which we were able to demonstrate a correlation between peak time and receptive field size and treated each cell as if it was independently recorded. Then, we used the following model (y = b₁x₁ + b₂x₂ + error) to estimate whether the value of peak time (y) could be predicted by taking into account both the receptive field size (x₁) and another additional variable (x₂). The additional variables tested were: amplitude of the first phase of the impulse response, mean firing rate under white-noise stimulation, layer location of the cell recorded, contrast polarity of the receptive field (ON- or OFF-center), and cell type (X or Y). We also did multiple regression analysis in each of the 10 cell groups that showed the strongest correlations between peak time and receptive field size (average r: 0.95).

Classification of geniculate cells

All geniculate cells recorded from layers A and C were classified as X or Y based on the linearity of spatial summation (Enroth-Cugell and Robson 1966; Hochstein and Shapley 1976; Shapley and Hochstein 1975). The linearity of spatial summation was measured with contrast reverse sinusoidal gratings. We used at least 2 different spatial frequencies that were higher than the optimal, usually 0.55 and 1.1 cycle/deg. Because some Y cells can generate linear responses when tested with very low spatial frequencies, high spatial frequencies were used to unambiguously classify groups of X and Y geniculate cells that were simultaneously recorded (Hochstein and Shapley 1976; So and Shapley 1979). Each spatial frequency was tested at 8 different phases. The gratings were presented at 0.4 Hz and repeated 8 times at each spatial phase. The Y/X identification was always made from the response to the highest spatial frequency that generated a significant response (5 spikes/50-ms bin). Cells that responded poorly (<5 spikes/50-ms bin) were labeled as unclassified. This study includes cells recorded in layer A and the first few hundred microns of layer C. A few cells recorded deep in layer C (>500 microns below the transition A1–C) had impulse responses with very long latencies and were not included in this study (Stanford et al. 1983; Sur and Sherman 1982; Wilson et al. 1976). Overall, we recorded from a total of 248 cells from layer A (104 X_A, 87 Y_A) and layer C (57 Y_C). Our matrix of 7 independently movable electrodes (Eckhorn and Thomas 1993) allowed us to record from clusters of 4–9 cells. These clusters were divided in groups based on cell type (X, Y) and layer location of the cells recorded (layer A or C). For example, a simultaneous recording from 4 X_A, 4 Y_A cells, and 1 Y_C cell yielded 4 groups: one X_A group (n = 4 cells), one Y_A group (n = 4 cells), one X_AY_A group (n = 8 cells), and one all-cells group (n = 9 cells).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
A multielectrode matrix was used to simultaneously record from several geniculate cells with overlapping receptive fields in cat LGN, either from the same or different layers. The electrodes were positioned very close to each other (80–300 microns between electrodes) and were independently moved (Eckhorn and Thomas 1993). Figure 1 shows an example of 4 cells that were simultaneously recorded within layer A of the LGN (3Y and 1X). The receptive field centers, mapped with white noise by reverse correlation, are shown in Fig. 1A as contour plots (shown separately on the right and superimposed on the left). For clarity, only the 20% contour lines are represented when showing superimposed receptive fields (in all figures ON-center cells are shown in continuous lines and OFF-center cells in discontinuous lines). The 4 cells shown in Fig. 1A had overlapping receptive fields that differed substantially in center size. In this example, the receptive field was increasingly larger for the cells represented, respectively, in blue (X cell), orange (Y cell), green (Y cell), and red (Y cell). Note that the contour plots represent the total receptive field center and not the Y cell subunit size (So and Shapley 1979).

View larger version (19K): [in this window] [in a new window]   FIG. 1. Receptive field size and response latency are correlated. A, left: receptive fields of 4 neighboring cells recorded within layer A of lateral geniculate nucleus (LGN): an X cell (in blue) and 3 Y cells (in red, green, and orange). For clarity, only the most external contour lines (20% of maximum response) are represented when showing superimposed receptive fields. All cells had OFF-center receptive fields (in all figures, OFF-center receptive fields are shown in discontinuous lines and ON-center receptive fields in continuous lines). Right: receptive fields are shown separately for each cell. Y cell in red had the largest receptive field size and the X cell the smallest one. B: impulse responses of the 4 cells had different peak times (indicated by arrows). Y cell in red had the fastest response latency and the X cell in blue the slowest one. Left: superimposed impulse responses. Right: impulse response of each cell is shown separately both as binned data (bin size = 15.5 ms) and as a cubic spline. C: peak time and receptive field sizes of these 4 cells were strongly correlated (r = –0.965, P = 0.035). XA: X cell recorded from layer A. YA: Y cell recorded from layer A.

View larger version (27K): [in this window] [in a new window]   FIG. 2. Space–time correlations are found among cells of the same type and of different types, within the same LGN layer (layer A), or across different layers. A, left: example from 5 XA cells with overlapping receptive fields that were simultaneously recorded. Each cell is represented in a different color. Five cells had overlapping receptive fields of different sizes. Continuous lines represent ON-center geniculate cells; discontinuous lines represent OFF-center geniculate cells. Middle: 5 cells had impulse responses with different peak times. First phase of the impulse response is positive for ON-center geniculate cells and negative for OFF-center cells (see METHODS). Right: receptive field size and response latency were strongly correlated (r = –0.948, P = 0.014). B: example from 5 YA cells. C: example from 7 cells of different type (XA, YA, YC) simultaneously recorded in different geniculate layers. D: example from 6 Y cells simultaneously recorded from layer C. No significant correlation was found between receptive field size and response latency in this example (r = 0, P = 1). XA: X cell recorded from layer A. YA: Y cell recorded from layer A. Yc: Y cell recorded from layer C.

View larger version (15K): [in this window] [in a new window]   FIG. 3. Probability of finding a significant space–time correlation depends on the number of simultaneously recorded cells and the range of receptive field sizes sampled (see text for details). A: probability is not significantly reduced as the number of simultaneously recorded cells increases but it is lowest for groups of 4 cells. B: probability depends on the receptive field size ratio (max/min; max and min are respectively the largest and smallest receptive field size within a group of simultaneously recorded cells) and the geniculate layer in which the cells are recorded (inset). Notice that a single cell can be part of 2 different groups. Therefore the total number of cells recorded (n = 248) is smaller than what would be obtained by counting all the cells from the 192 groups shown in this figure. See METHODS for detail.

View larger version (26K): [in this window] [in a new window]   FIG. 4. Space–time correlations can be demonstrated in large populations of X and Y cells by using data from simultaneous recordings. A: each cell in one simultaneously recorded group was rank-ordered according to its receptive field size and response latency. In each simultaneously recorded group, the cell with the biggest receptive field was ranked as 1 in "receptive field size" (top) and the cell with shortest response latency was ranked as 1 in "peak time" (middle). Rank values were normalized by subtracting the mean rank and then represented in a scatter plot (bottom). See METHODS for detail. B: when the entire cell population was rank-ordered by receptive field size and response latency, significant correlations were found within and across cell types. Circles of different sizes are used to represent the number of the cells with the same rank values. RF, receptive field.

View larger version (16K): [in this window] [in a new window]   FIG. 5. Monte Carlo simulations of the rank-ordered data demonstrate a positive correlation between peak time and receptive field size. Figure shows frequency histograms obtained from 1,000 simulations run for all the cells (A), Y cells (B), and X cells (C). Correlation coefficients obtained in these simulations were all positive. See METHODS for detail.

View larger version (32K): [in this window] [in a new window]   FIG. 6. Space–time correlations obtained with different stimulus contrasts. A: example from a group of simultaneously recorded cells (5 X and 4 Y cells from layer A) that showed a space–time correlation at 3 different contrasts. Stimulus was white noise (inset). B: distribution of correlation coefficients (r values) obtained from each group of simultaneously recorded cells. Distributions show only significant correlations (at low contrasts some space–time correlations were no longer significant). Dashed line indicates the average r-value. Peak time range (minimum peak time measured–maximum peak time measured) was largest at the lowest stimulus contrast. Receptive field size range (minimum receptive field size measured–maximum receptive field size measured) was similar under different stimulus contrasts. Top: high contrast (98%). Middle: medium contrast (45%). Bottom: low contrast (22%).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

Why were space–time correlations not reported before?

The receptive field size and response time course of LGN cells have been measured in the past with several methods, although no study did a systematic comparison between both parameters (for review see Sherman and Spear 1982; Stone et al. 1979; Van Hooser et al. 2003). Previous studies emphasized the high variability of the spatial and temporal properties of LGN cells and the importance of using similar measuring conditions to reduce such variability (Cleland et al. 1979; So and Shapley 1979). The use of simultaneous recordings greatly minimizes this variability and, in addition, allows a direct comparison of neighboring geniculate cells. For example, simultaneous recordings from 2 cells with the same electrode showed that a Y retinal cell could respond 10–15 ms faster to visual stimuli than its neighboring X retinal cell (Bolz et al. 1982). Similarly, simultaneous recordings within the LGN demonstrated that a Y geniculate cell could respond 10–20 ms faster than the neighboring X geniculate cell (Yeh et al. 2003).

It is well accepted that Y geniculate cells have, on average, larger receptive fields and respond faster to visual stimuli than X geniculate cells (Saul and Humphrey 1990; Sestokas and Lehmkuhle 1986; So and Shapley 1979; Troy and Lennie 1987; Yeh et al. 2003). However, this fact could reflect either differences between cell types (i.e., X, Y) or a more general correlation between receptive field size and response latency. Our results provide support for the second possibility. In other words, if 2 neighboring cells have different receptive field sizes, the one with the larger receptive field is more likely to respond faster to visual stimuli than the one with the smaller receptive field, even if the 2 cells are of the same type (i.e., Y cells).

Possible mechanisms

What could be the mechanism responsible for the space–time correlations described here? It is possible that the mechanism that makes receptive fields larger is closely related to the mechanism that makes responses faster. An interesting possibility is that both parameters depend on the number of convergent inputs that a cell receives (from the photoreceptor to the LGN). Increasing the number of convergent inputs should make geniculate receptive fields larger (more cones per geniculate cell) and visual responses faster [more synapses being simultaneously activated lead to a larger compound synaptic potential that reaches threshold faster (e.g., Andreasen and Lambert 1998)]. Data on X and Y geniculate cells provide support for this "convergence hypothesis." Y geniculate cells have larger receptive fields than X geniculate cells because they receive input from more cones (for review see Sterling 2004), and possibly from more retinal inputs (Yeh et al. 2003). In addition, Y geniculate cells respond faster to visual stimuli than X cells because they may receive input from larger and more numerous synaptic boutons and may be served by thicker/faster axons (Sur et al. 1987) [thicker axons may be necessary to maintain the higher metabolic demand of a larger volume of synaptic boutons (Sterling 2004)]. Summing a larger number of inputs also improves the signal-to-noise ratio and the contrast sensitivity. A low-contrast stimulus generates the fastest responses in cells with the highest contrast sensitivity, which usually have the largest receptive fields (Shapley and Victor 1978). Therefore cells with large receptive fields should generate faster responses than other cells because they have higher contrast gain. This mechanism clearly influences the space–time correlations reported here (particularly when using low-contrast stimuli). Although the strength of the correlation does not depend on the stimulus contrast (Fig. 6), the range of response latencies is reduced when the contrast increases.

The space–time correlations could also be a consequence of a common developmental mechanism of receptive field size and response latency. Just before eye opening, a geniculate cell receives weak synaptic contacts from many retinal ganglion cells; however, as the LGN matures most of the multiple inputs are eliminated (Chen and Regehr 2000; Mason 1982; Sretavan and Shatz 1986; Sur et al. 1984). The elimination of multiple retinal inputs could lead to both a reduction in receptive field size and a faster synaptic integration (because the remaining inputs become much stronger and therefore reach threshold faster).

Functional significance

Because of their space–time correlations, X and Y geniculate cells should provide the cortex with images whose spatial resolution becomes sharper over time. This finding is consistent with evidence indicating that low spatial frequencies generate faster visual responses than high spatial frequencies in cortical cells (Frazor et al. 2004; Mazer et al. 2002). It is also consistent with results in macaque showing that the peak of the spatial frequency tuning in visual cortical cells shifts toward higher frequencies as the response progresses in time (Bredfeldt and Ringach 2002; Frazor et al. 2004). Recently Frazor et al. (2004) modeled these spatial frequency shifts using convergent feedforward inputs. As indicated in their paper, the model applies to different types of thalamic neurons: Magnocellular and Parvocellular in the primate, X and Y in the cat, or any thalamic inputs with latencies that covary with spatial frequency. Here, we demonstrate that LGN inputs in the cat show space–time covariations that are independent of cell type. Therefore our results indicate that the shifts in the peak of the spatial frequency tuning of a cortical cell can be explained by the convergence of feedforward thalamic inputs—in both cortical cells that receive one type of input (X or Y) and those that receive mixed input.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
This research was supported by National Eye Institute Grant EY-05253 and The Research Foundation at the University of Connecticut and State University of New York.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
We thank J. B. Troy for guiding us on the literature of response timing in LGN. We also thank J. de la Torre and J.-A. Aguilar for helping in the development of analysis software and S. Bhaskar and J. Bachand for technical assistance.

Address for reprint requests and other correspondence: J.-M. Alonso, Department of Biological Sciences, State University of New York—State College of Optometry, 33 West 42nd Street, New York, NY 10036 (E-mail: jalonso{at}sunyopt.edu)

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Alonso JM, Usrey WM, and Reid RC. Rules of connectivity between geniculate cells and simple cells in cat primary visual cortex. J Neurosci 21: 4002–4015, 2001.[Abstract/Free Full Text]

Andreasen M and Lambert JD. Factors determining the efficacy of distal excitatory synapses in rat hippocampal CA1 pyramidal neurones. J Physiol 507: 441–462, 1998.[Abstract/Free Full Text]

Bolz J, Rosner G, and Wassle H. Response latency of brisk-sustained (X) and brisk-transient (Y) cells in the cat retina. J Physiol 328: 171–190, 1982.[Abstract/Free Full Text]

Bredfeldt CE and Ringach DL. Dynamics of spatial frequency tuning in macaque V1. J Neurosci 22: 1976–1984, 2002.[Abstract/Free Full Text]

Chen C and Regehr WG. Developmental remodeling of the retinogeniculate synapse. Neuron 28: 955–966, 2000.[CrossRef][Web of Science][Medline]

Cleland BG, Harding TH, and Tulunay-Keesey U. Visual resolution and receptive field size: examination of two kinds of cat retinal ganglion cell. Science 205: 1015–1017, 1979.[Abstract/Free Full Text]

Croner LJ and Kaplan E. Receptive fields of P and M ganglion cells across the primate retina. Vision Res 35: 7–24, 1995.[CrossRef][Web of Science][Medline]

Demb JB, Zaghloul K, Haarsma L, and 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.[Abstract/Free Full Text]

Derrington AM and Fuchs AF. Spatial and temporal properties of X and Y cells in the cat lateral geniculate nucleus. J Physiol 293: 347–364, 1979.[Abstract/Free Full Text]

Eckhorn R and Thomas U. A new method for the insertion of multiple microprobes into neural and muscular tissue, including fiber electrodes, fine wires, needles and microsensors. J Neurosci Methods 49: 175–179, 1993.[CrossRef][Web of Science][Medline]

Enroth-Cugell C and Robson JG. The contrast sensitivity of retinal ganglion cells of the cat. J Physiol 187: 517–552, 1966.

Frazor RA, Albrecht DG, Geisler WS, and Crane AM. Visual cortex neurons of monkeys and cats: temporal dynamics of the spatial frequency response function. J Neurophysiol 91: 2607–2627, 2004.[Abstract/Free Full Text]

Hochstein S and Shapley RM. Quantitative analysis of retinal ganglion cell classifications. J Physiol 262: 237–264, 1976.[Abstract/Free Full Text]

Kremers J, Silveira LC, and Kilavik BE. Influence of contrast on the responses of marmoset lateral geniculate cells to drifting gratings. J Neurophysiol 85: 235–246, 2001.[Abstract/Free Full Text]

Levitt JB, Schumer RA, Sherman SM, Spear PD, and Movshon JA. Visual response properties of neurons in the LGN of normally reared and visually deprived macaque monkeys. J Neurophysiol 85: 2111–2129, 2001.[Abstract/Free Full Text]

Mason CA. Development of terminal arbors of retino-geniculate axons in the kitten—II. Electron microscopical observations. Neuroscience 7: 561–582, 1982.[CrossRef][Web of Science][Medline]

Maunsell JH, Ghose GM, Assad JA, McAdams CJ, Boudreau CE, and Noerager BD. Visual response latencies of magnocellular and parvocellular LGN neurons in macaque monkeys. Vis Neurosci 16: 1–14, 1999.[CrossRef][Web of Science][Medline]

Maunsell JH and Gibson JR. Visual response latencies in striate cortex of the macaque monkey. J Neurophysiol 68: 1332–1344, 1992.[Abstract/Free Full Text]

Mazer JA, Vinje WE, McDermott J, Schiller PH, and Gallant JL. Spatial frequency and orientation tuning dynamics in area V1. Proc Natl Acad Sci USA 99: 1645–1650, 2002.[Abstract/Free Full Text]

Pettigrew JD, Cooper ML, and Blasdel GG. Improved use of tapetal reflection for eye-position monitoring. Invest Ophthalmol Vis Sci 18: 490–495, 1979.[Abstract]

Reid RC, Victor JD, and Shapley RM. The use of m-sequences in the analysis of visual neurons: linear receptive field properties. Vis Neurosci 14: 1015–1027, 1997.[Web of Science][Medline]

Saul AB and Humphrey AL. Spatial and temporal response properties of lagged and nonlagged cells in cat lateral geniculate nucleus. J Neurophysiol 64: 206–224, 1990.[Abstract/Free Full Text]

Sestokas AK and Lehmkuhle S. Visual response latency of X- and Y-cells in the dorsal lateral geniculate nucleus of the cat. Vision Res 26: 1041–1054, 1986.[CrossRef][Web of Science][Medline]

Shapley R and Hochstein S. Visual spatial summation in two classes of geniculate cells. Nature 256: 411–413, 1975.[CrossRef][Medline]

Shapley RM and Victor JD. The effect of contrast on the transfer properties of cat retinal ganglion cells. J Physiol 285: 275–298, 1978.[Abstract/Free Full Text]

Sherman SM and Spear PD. Organization of visual pathways in normal and visually deprived cats. Physiol Rev 62: 738–855, 1982.[Free Full Text]

So YT and Shapley R. Spatial properties of X and Y cells in the lateral geniculate nucleus of the cat and conduction velocities of their inputs. Exp Brain Res 36: 533–550, 1979.[Web of Science][Medline]

Sretavan DW and Shatz CJ. Prenatal development of retinal ganglion cell axons: segregation into eye-specific layers within the cat's lateral geniculate nucleus. J Neurosci 6: 234–251, 1986.[Abstract]

Stanford LR, Friedlander MJ, and Sherman SM. Morphological and physiological properties of geniculate W-cells of the cat: a comparison with X- and Y-cells. J Neurophysiol 50: 582–608, 1983.[Free Full Text]

Sterling P. How retinal circuits optimize the transfer of visual information. In: The Visual Neurosciences, edited by Chalupa LM and Werner JS. Cambridge, MA: MIT Press, 2004, p. 234–259.

Stone J, Dreher B, and Leventhal A. Hierarchical and parallel mechanisms in the organization of visual cortex. Brain Res 180: 345–394, 1979.[Medline]

Sur M, Esguerra M, Garraghty PE, Kritzer MF, and Sherman SM. Morphology of physiologically identified retinogeniculate X- and Y-axons in the cat. J Neurophysiol 58: 1–32, 1987.[Abstract/Free Full Text]

Sur M and Sherman SM. Linear and nonlinear W-cells in C-laminae of the cat's lateral geniculate nucleus. J Neurophysiol 47: 869–884, 1982.[Abstract/Free Full Text]

Sur M, Weller RE, and Sherman SM. Development of X- and Y-cell retinogeniculate terminations in kittens. Nature 310: 246–249, 1984.[CrossRef][Medline]

Sutter EE. A practical non-stochastic approach to nonlinear time-domain analysis. In: Advance Methods of Physiological Systems Modeling, Workshop, Los Angeles, CA, University of Southern California, 1987, p. 1.

Swadlow HA and Weyand TG. Receptive-field and axonal properties of neurons in the dorsal lateral geniculate nucleus of awake unparalyzed rabbits. J Neurophysiol 54: 168–183, 1985.[Abstract/Free Full Text]

Troy JB and Lennie P. Detection latencies of X and Y type cells of the cat's dorsal lateral geniculate nucleus. Exp Brain Res 65: 703–706, 1987.[Web of Science][Medline]

Usrey WM and Reid RC. Visual physiology of the lateral geniculate nucleus in two species of new world monkey: Saimiri sciureus and Aotus trivirgatis. J Physiol 523: 755–769, 2000.[Abstract/Free Full Text]

Van Hooser SD, Alexander J, Heimel F, and Nelson SB. Receptive field properties and laminar organization of lateral geniculate nucleus in the gray squirrel (Sciurus carolinensis). J Neurophysiol 90: 3398–3418, 2003.[Abstract/Free Full Text]

Weng C, Stoelzel CR, Yeh CI, and Alonso JM. Receptive field size and response latency are correlated within the cat visual thalamus. Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience, 2002 CD-ROM Program No. 352.11, 2002.

Wilson PD, Rowe MH, and Stone J. Properties of relay cells in cat's lateral geniculate nucleus: a comparison of W-cells with X- and Y-cells. J Neurophysiol 39: 1193–1209, 1976.[Abstract/Free Full Text]

Worgotter F, Suder K, Zhao Y, Kerscher N, Eysel UT, and Funke K. State-dependent receptive-field restructuring in the visual cortex. Nature 396: 165–168, 1998.[CrossRef][Medline]

Xu X, Ichida JM, Allison JD, Boyd JD, and Bonds AB. A comparison of koniocellular, magnocellular and parvocellular receptive field properties in the lateral geniculate nucleus of the owl monkey (Aotus trivirgatus). J Physiol 531: 203–218, 2001.[Abstract/Free Full Text]

Yeh CI, Stoelzel CR, and Alonso JM. Two different types of Y cells in the cat lateral geniculate nucleus. J Neurophysiol 90: 1852–1864, 2003.[Abstract/Free Full Text]

This article has been cited by other articles:

				J.-M. Alonso My recollections of Hubel and Wiesel and a brief review of functional circuitry in the visual pathway J. Physiol., June 15, 2009; 587(12): 2783 - 2790. [Abstract] [Full Text] [PDF]

				C.-I Yeh, C. R. Stoelzel, C. Weng, and J.-M. Alonso Functional Consequences of Neuronal Divergence Within the Retinogeniculate Pathway J Neurophysiol, April 1, 2009; 101(4): 2166 - 2185. [Abstract] [Full Text] [PDF]

				J. W. Middleton, A. Longtin, J. Benda, and L. Maler Postsynaptic Receptive Field Size and Spike Threshold Determine Encoding of High-Frequency Information Via Sensitivity to Synchronous Presynaptic Activity J Neurophysiol, March 1, 2009; 101(3): 1160 - 1170. [Abstract] [Full Text] [PDF]

				D. J. Tolhurst, D. Smyth, and I. D. Thompson The Sparseness of Neuronal Responses in Ferret Primary Visual Cortex J. Neurosci., February 25, 2009; 29(8): 2355 - 2370. [Abstract] [Full Text] [PDF]

				J. Lee, T. Williford, and J. H. R. Maunsell Spatial Attention and the Latency of Neuronal Responses in Macaque Area V4 J. Neurosci., September 5, 2007; 27(36): 9632 - 9637. [Abstract] [Full Text] [PDF]

				B. J. Malone, V. R. Kumar, and D. L. Ringach Dynamics of Receptive Field Size in Primary Visual Cortex J Neurophysiol, January 1, 2007; 97(1): 407 - 414. [Abstract] [Full Text] [PDF]

				E. A. Allen and R. D. Freeman Dynamic spatial processing originates in early visual pathways. J. Neurosci., November 8, 2006; 26(45): 11763 - 11774. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 93/6/3537    most recent 00847.2004v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (14) Citing Articles via Google Scholar Google Scholar Articles by Weng, C. Articles by Alonso, J.-M. Search for Related Content PubMed PubMed Citation Articles by Weng, C. Articles by Alonso, J.-M.