Receptive field properties and sensitivity to edges defined by motion in the postero-lateral lateral suprasylvian (PLLS) area of the cat

BR A IN RE S EA RCH 1 1 87 ( 20 0 8 ) 8 2 –94

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s

Research Report

Receptive field properties and sensitivity to edges defined by motion in the postero-lateral lateral suprasylvian (PLLS) area of the cat N. Robitaillea , F. Leporea , B.A. Bacona,b , D. Ellemberga,c , J.-P. Guillemot a,d,⁎ a

Centre de Recherche en Neuropsychologie et Cognition, Département de Psychologie, Université de Montréal, Montréal, Québec, Canada Department of Psychology, Bishop's University, Sherbrooke, Québec, Canada c Département de Kinésiologie, Université de Montréal, Québec, Canada d Département de Kinanthropologie, Université du Québec à Montréal, Montréal, Canada b

A R T I C LE I N FO

AB S T R A C T

Article history:

The present study investigated the spatial properties of cells in the postero-lateral lateral

Accepted 10 October 2007

suprasylvian (PLLS) area of the cat and assessed their sensitivity to edges defined by motion.

Available online 22 October 2007

A total of one hundred and seventeen (117) single units were isolated. First, drifting sinusoidal gratings were used to assess the spatial properties of the cells' receptive fields and

Keywords:

to determine their spatial frequency tuning functions. Second, random-dot kinematograms

Extrastriate

were used to create illusory edges by drifting textured stimuli (i.e. a horizontal bar) against a

Kinematogram

similarly textured but static background. Almost all the cells recorded in PLLS (96.0%) were

PLLS

binocular, and a substantial majority of receptive fields (79.2%) were end-stopped. Most units

Spatial frequency

(81.0%) had band-pass spatial frequency tuning functions and responded optimally to low

Texture

spatial frequencies (mean spatial frequency: 0.08 c./degree). The remaining units (19.0%) were low-pass. All the recorded cells responded vigorously to edges defined by motion. The vast majority (96.0%) of cells responded optimally to large texture elements; approximately half the cells (57.3%) also responded to finer texture elements. Moreover, 38.5% of the cells were selective to the width of the bar (i.e., the distance between the leading and the trailing edges). Finally, some (9.0%) cells responded in a transient fashion to leading and to trailing edges. In conclusion, cells in the PLLS area are low spatial frequency analyzers that are sensitive to texture and to the distance between edges defined by motion. © 2007 Elsevier B.V. All rights reserved.

1.

Introduction

The visual system can quickly detect and identify objects in natural scenes, even though these scenes are usually composed of varied, complex and often degraded visual information.

Perceiving an object as a distinct entity requires that the visual system segregate that object from the rest of the scene by defining its boundaries or edges. For example, an object may be segregated from its background based on differences in luminance, color, spatial disparity, texture, and direction of motion

⁎ Corresponding author. Département de Kinanthropologie, Université du Québec à Montréal, C.P. 8888, Succ. Centre-Ville, Montréal, Canada H3C 3P8. Fax: +1 514 343 5787. E-mail address: [email protected] (J.-P. Guillemot). Abbreviations: c./degree, cycle per degree; cd/m2, Candela per meter squared; Imp./s, impulses per second; MT, middle temporal area; PLLS, postero-lateral lateral suprasylvian; PMLS, postero-medial lateral suprasylvian; RF, receptive field; S.D., standard deviation 0006-8993/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2007.10.031

BR A IN RE S E A RCH 1 1 87 ( 20 0 8 ) 8 2 –9 4

(Julesz, 1971; Nothdurft, 1993). One of the most salient cues for figure–ground segregation is relative motion between an object and its background (Nothdurft, 1993). This is demonstrated by the fact that several species of prey have evolved a “freezing” strategy so as to eliminate this cue and escape the detection of predators. In cats, the cortical areas surrounding the lateral suprasylvian sulcus are known to be involved in the processing of motion information (Lomber et al., 1996; Spear, 1991). A number of studies suggest that these areas are implicated in attention shifts (Ogasawara et al., 1984; Hardy and Stein, 1988; Payne et al., 1996), speed discrimination (Pasternak et al., 1989), the integration of complex motion (Rudolph and Pasternak, 1996), and the detection of forms that are in motion (Kiefer et al., 1989; Krüger et al., 1993). The lateral suprasylvian cortex consists of six areas, each containing an independent retinotopic map of the visual field (Palmer et al., 1978). Research on the lateral suprasylvian cortex has largely focused on the postero-medial lateral suprasylvian area (PMLS). The receptive field properties of cells in PMLS have been extensively studied (Camarda and Rizzolatti, 1976; Hubel and Wiesel, 1969; Rauschecker et al., 1987a,b; Spear and Baumann, 1975; von Grünau et al., 1987; Zumbroïch et al., 1986), and the spatio-temporal properties of these cells have been well defined (Morrone et al., 1986; Zumbroïch and Blakemore, 1987; Zumbroïch et al., 1988; Guido et al., 1990). The role of PMLS cells with regards to motion perception has been well established (Morrone et al., 1986; Blakemore and Zumbroïch, 1987; Rauschecker et al., 1987a; von Grünau and Frost, 1983; Yin and Greenwood, 1992) and this area is usually considered homologous to area V5 (MT) of the macaque brain, the acknowledged motion center in primates (Payne, 1993). Much less is known about the postero-lateral lateral suprasylvian cortex (PLLS). It is generally accepted that this area, like PMLS, contributes to motion analysis but the specific responses of cells in this area and the exact role they play in this process remains poorly defined. Zumbroïch et al. (1986) highlighted an important difference between PLLS and PMLS. In the PLLS area, the upper visual field is over-represented and a large proportion of neurons extend their receptive fields into the ipsilateral visual field (up to 28 degrees from the vertical meridian). Furthermore, Rauschecker et al. (1987a) reported a centrifugal–centripetal organization of the receptive fields' optimal direction in PLLS; they link this type of organization with the analysis of expanding stimuli. At the neuroanatomical level, the PLLS area receives both direct and indirect Y inputs from the dorsal lateral geniculate nucleus (Rauschecker et al., 1987b; Raczkowski and Rosenquist, 1983). This Y pathway is associated with the detection of motion (Stone et al., 1979; Khayat et al., 2000) and the coarse analysis of form (Stone, 1983). PLLS neurons also share strong reciprocal connections with the anterior ectosylvian area (Scannell et al., 1995), a higher-order cortical region that contributes to the analysis of motion (Benedek et al., 1988; Scannell et al., 1996). Furthermore, PLLS cells send outputs to neurons in the superficial, the intermediate and the deep layers of the superior colliculus (Kawamura and Hashikawa, 1978; Niida et al., 1997; Brecht et al., 1998), a structure associated with ocular movements, fixation, and orienting behavior (Roucoux and Crommelinck, 1976; Stein, 1978). Moreover, PLLS cells discharge after voluntary ocular movements (Komatsu et al., 1983), and

83

electrical stimulation of these cells provokes convergent ocular movements (Toda et al., 2001). One study systematically investigated the spatial and temporal properties of cells in area PLLS (Zumbroïch and Blakemore, 1987). Based on a very small sample (13 cells), they concluded that PLLS cells are spatially similar to those of PMLS, but that they tend to prefer higher temporal frequencies. Recently, Li et al. (2000) have shown that the vast majority (90.0%) of PLLS cells responded to optic flow patterns, although only 20–25% of the cells were selective to certain types of optic flow stimuli (i.e., translation, rotation, or expansion–contraction). This is consistent with the report that the majority of cells in the lateral suprasylvian cortex respond preferentially to optic flow movies rather than to equivalent texture movies (Kim et al., 1997). Moreover, Sherk et al. (1997) have shown that cells in the PLLS area respond preferentially to objects moving against an optic flow movie rather than to a bar moving against a uniform background. Together, these findings suggest that the PLLS area constitutes an intermediate stage of processing for optic flow fields. There is also indirect evidence suggesting that the PLLS cortex is involved in figure–ground segmentation. The PLLS/ PMLS border receives dense heterotopic callosal connections from the 17/18 border (Innocenti et al., 1995; Bressoud and Innocenti, 1999), and according to Innocenti et al. (1995), these callosal connections play a role in figure–ground segregation. The contribution of motion-cues to figure–ground segregation has mainly been investigated with two types of texturebased stimuli. The first, a static form, is defined solely by the motion of a subset of elements (visual noise or textons) within fixed spatial coordinates (Frost, 1985; Gulyas et al., 1987, 1990). The second, similar to the stimuli used in the present study, consists of a subset of elements that move coherently within a background of static elements. In this case, if the elements become static, the form disappears and a completely uniform field of dots is perceived. This type of motion-defined form can be detected by neurons in the cat's dorsal lateral geniculate nucleus (Mason, 1976) and area 17 (Hammond and MacKay, 1977), as well as in the monkey primary visual cortex (Bourne et al., 2002) and MT area (Olavarria et al., 1992; Marcar et al., 1995). A recent study from our laboratory demonstrated that almost all cells in area 19 of the cat respond to a motion-defined bar or to its edges (Khayat et al., 2000). Our results also showed that texture density has an influence on figure–ground segmentation: cell responses increased as dot density decreased. The first objective of the present study was to assess the spatial properties of the receptive fields of a large sample of cells in area PLLS and to determine their spatial frequency tuning functions. The second and main objective was to investigate the role of PLLS cortex in figure–ground analysis based on motion cues. Random-dot kinematograms were used to create drifting edges consisting of a textured form (i.e., a horizontal bar) against a similarly textured but static background. The response rates of PLLS neurons to moving bars of different widths were measured as a function of dot size.

2.

Results

One hundred and seventeen cells were isolated in area PLLS of the cat. Cells presenting unstable or erratic responses were

84

BR A IN RE S EA RCH 1 1 87 ( 20 0 8 ) 8 2 –94

5 degrees from the midline. The limit of the most distant RF in the ipsilateral hemifield was 17.3 degrees from the vertical meridian. The relationship between the sizes of the RFs of the dominant eye and their eccentricity on the azimuth is plotted in Fig. 1B. The receptive fields were quite large (mean: 175.3 degrees2, S.D.: 272.4 degrees2), ranging from 5.0 to 735 degrees2, although for nearly three-quarters (74.2%) of the cells, they were smaller than 200 degrees2. Of these relatively small RFs, 89% were within 10 degrees of azimuth. However, RFs larger than 400 degrees2 were found at various eccentricities (0.8–20.2 degrees), and a few cells having smaller RFs (≤200 degrees2) had their centers as far as 15 degrees from the vertical meridian. Nonetheless, there was a significant (r = 0.4, p ≤0.01) relationship between RF size and azimuthal eccentricity: they tended to increase in size as distance from the vertical meridian increased.

Fig. 1 – (A) Spatial distribution of receptive fields mapped in the PLLS area. Each point corresponds to the center of the receptive field of the dominant eye. Most receptive fields were located in the upper quadrant of the contralateral visual field but several straddled the ipsilateral visual field. (B) Relationship between receptive field size (degree2) and the azimuth position for 101 cells. Receptive field size shows a significant positive correlation (r = 0.4, p < 0.001) to azimuth position, as illustrated by the scatter plot and the regression line.

excluded from the analysis. The position, the limits, and the spatial organization of the receptive fields of 101 cells were carefully determined using light/dark bars and drifting sinusoidal gratings. Moreover, sensitivity to edges defined by motion was successfully tested in all these cells.

2.1.

Receptive field properties

Fig. 1A shows the azimuth and the elevation positions of the centers of the receptive fields (RFs) defined through the dominant eye for 101 cells. Most (82.5%) of the RF centers were located in the upper visual field. Although all penetrations were performed in the right hemisphere, 25.7% of these RF centers were located in the ipsilateral hemifield. Interestingly, in 59.4% of the cells, the RFs defined through the dominant eye touched or straddled the vertical meridian. Nearly half (49.5%) of the RF centers were located at an azimuth eccentricity of less than

Fig. 2 – (A) Classification by receptive fields types of cells recorded in area PLLS. Most receptive fields had a spatial organization of the end-stopped complex (Ch) and complex (C) type. A small proportion of cells (15.8%) had receptive fields of the end-stopped simple (Sh) type. No simple (S) receptive fields were found. (B) Ocular dominance distribution of PLLS cells. Most cells were binocular and there was a strong bias in favour of the contralateral eye.

BR A IN RE S E A RCH 1 1 87 ( 20 0 8 ) 8 2 –9 4

85

monocular cells could only be driven through the contralateral eye. Similarly, most of the binocular cells (88.2%) preferred contralateral stimulation.

2.2.

Spatial frequency tuning functions

Spatial frequency tuning functions were determined for 73 cells by drifting sinusoidal gratings across the receptive field of the dominant eye. The majority of cells (81.0%) were classified as band-pass. These types of cells responded most strongly or optimally to a given spatial frequency and presented a drastic attenuation in their response rates at higher and lower spatial frequencies. For example, as can be seen in Fig. 3A, cell 50 has a very low optimal spatial frequency (0.02 c./

Fig. 3 – Six typical examples of spatial frequency tuning functions, established by stimulation of the dominant eye, of cells in the PLLS area. The stimuli were high contrast (50%) sinusoidal gratings drifting in the optimal direction. Each point corresponds to the mean firing rate (minus the baseline rate) of the cell in response to a particular spatial frequency. The three cells in (A) are band-pass while the three cells in (B) are low-pass.

Fig. 2A presents the distribution of the spatial organization of the RFs of 101 cells. They were assessed through the dominant eye and classified as complex, end-stopped complex, simple, end-stopped simple using the criteria of Hubel and Wiesel (1962, 1965), Henry et al. (1967), Henry (1977) and Skottun et al. (1991). The results show that most (63.4%) had a spatial organization of the end-stopped complex type and 20.8% of the complex type. A smaller proportion of cells (15.8%) had RFs of end-stopped simple type, and no simple types were encountered. Therefore, more than three-quarters (79.2%) of the cells showed clear end-stopped inhibition. Fig. 2B shows the distribution of ocular dominance. As can be seen, almost all cells (96.0%) were binocular and the few

Fig. 4 – (A) Distribution of optimal spatial frequencies estimated from the spatial frequency tuning functions of the 59 band-pass cells recorded in area PLLS. Almost all the cells responded optimally to relatively low spatial frequencies. (B) Distribution of spatial bandwidths of these 59 band-pass cells; bandwidths approximate the range of spatial frequencies to which the cells were selective. The distribution is quite broad and it is centered at approximately 2 octaves. Note that a non-negligible proportion of cells (15.3%) showed a very narrow spatial bandwidth (≤1 octave).

86

BR A IN RE S EA RCH 1 1 87 ( 20 0 8 ) 8 2 –94

Fig. 5 – (A and B) Peristimulus time histograms representing the responses of two cells to kinematograms as a function of pixel size (texture density) and bar width. These two cells vigorously responded to bars composed of larger pixels and their response rate decreased (cell 5; B) or even disappeared (cell 23; A) when smaller pixels were used, regardless of bar width. In (B), even for the smallest pixel size, two peaks are discernable in the peristimulus time histograms for the larger bar widths (4 and 8 degrees). The first and second peaks correspond to responses to the leading edge and to the trailing edge, respectively. PSTH duration: 2 s.

degree) while cells 104 and 115 have higher optimal spatial frequencies (0.08 and 0.15 c./degree, respectively). The distribution of optimal spatial frequency responses for the 59 bandpass cells is shown in Fig. 4A. and ranged from 0.017 to 0.33 c./ degree. The mean optimal spatial frequency was very low (0.08 c./degree; S.D.: 0.05 c./degree), with only 5.1% of the cells having a frequency higher than 0.16 c./degree. The optimal spatial frequency was negatively and significantly correlated with both RF azimuth eccentricity (r = −0.31, p ≤ 0.05) and size (r = −0.28, p ≤ 0.05). That is, the optimal spatial frequencies were lower for large receptive fields situated at high eccentricities. To assess the selectivity of the 59 band-pass cells, spatial bandwidths were calculated at half-amplitude of the optimal response. The bandwidths were relatively large, ranging from 0.66 to 3.59 octaves, as shown by the typical examples presented in Fig. 3A (cell 104: 1.34 octaves, cell 115: 1.7 octaves, cell 50: 1.8 octaves). The spatial bandwidth distribution of the entire group of band-pass cells is presented in Fig. 4B. The mean bandwidth was relatively large (1.82 octaves; S.D.: 0.78 octaves), but similar to what was found in other extrastriate visual areas (area 19: Tanaka et al., 1987; Bergeron et al., 1998; Tardif et al., 1997; area PMLS: Zumbroïch and Blakemore, 1987). Interestingly, the cells (15.3%) that had narrow bandwidths (≤1 octave) showed the strongest end-stopped inhibition. The spatial bandwidths were significantly correlated with eccentricity (r = 0.4, p ≤ 0.001) and receptive field size (r = 0.27, p ≤ 0.05). Indeed, they were lower for small receptive fields situated near the vertical meridian. Zumbroïch and Blakemore (1987) found, in the PMLS area, a relationship between optimal spatial frequency and bandwidth. However, the present results do not show such a relationship (r = −0.2, p N 0.05).

The remaining cells (19%) had spatial frequency tuning functions that were classified as low-pass. These cells showed an attenuation of their response rates at high spatial frequencies and little or no attenuation at the lower end of the spatial frequency spectrum (for example, see Fig. 3B, cells 87 and 88). However, it is not impossible that some of these cells could have been classified as band-pass. For example, cell 95 (Fig. 3B) was classified as low-pass because its response rate at the lowest spatial frequency was higher than half of its response at the optimal spatial frequency. However, it is possible that its response could have undergone a greater attenuation if lower spatial frequencies had been tested.

2.3.

Response to edges defined by motion

All 101 cells responded vigorously to at least one of the 20 conditions tested (4 bar widths × 5 pixel sizes). Fig. 5 presents the responses of two cells to these different conditions. Fig. 5A shows that cell 23 responded vigorously to bars made of larger pixels and that the response rate of this cell decreased, and even disappeared, when smaller pixels were used to form the edge defined by motion, regardless of the width of the bar. This pattern of response was found in almost all of the cells in area PLLS. Cell 5 (Fig. 5B) also showed an increasing response rate when larger pixels were used, regardless of the width of the bar. However, in contrast to cell 23, this cell also responded to kinematograms composed of finer (0.06 and 0.12 degrees) pixels. Furthermore, as shown by the shape of the peristimulus time histogram (PSTH), a particular pattern of response to larger bars (4 or 8 degrees) was observed for the cell shown in Fig. 5B. Even for the smallest pixel size, two peaks in the peristimulus time

BR A IN RE S E A RCH 1 1 87 ( 20 0 8 ) 8 2 –9 4

87

Fig. 6 – Three-dimensional response areas reconstructed from cell responses to kinematograms. (A–F) Typical examples of cells classified according to their selectivity to bar width (four conditions) and pixel size (five conditions). The abscissa represents pixel size and the ordinate represents bar width. The Z axis shows the normalized response rate (see text) for each of the 20 conditions. The cell in (A) exemplifies the most common pattern (35.6%) of response. This category of cells responded vigorously to all bar widths and their response rates increased with increasing pixel size. The remainder of the cells (38.5%) showed responses that were tuned to bar width (B–F). The cell in (B) only responded to bar widths larger than 2 degrees. A third category (C) of cells only responded to bar widths that were smaller than 4 degrees. Another category of cells (D) responded preferentially bar of intermediate widths (2 and 4 degrees). The two last categories (E, F) consist of cells that were highly selective to widths: In (E) the cell responded exclusively to the largest widths while in (F) it responded only to the smallest widths.

histograms are discernable. The first peak corresponds to an increase in firing rate in response to the leading edge of the kinematogram, and the second corresponds to an increase in firing rate in response to its trailing edge. For the smaller bar widths (1 and 2 degrees), only one peak is discernable. This type of response was encountered in 8.9% of the cells. A majority of cells (74.1%) could be classified as belonging to one of six categories based on their response to the distance between the edges or, in other words, their response to bar width (Fig. 6). All of these six categories of cells responded mainly to kinematograms that were composed of the largest pixels (N0.5 degrees). Figs. 6A to F present normalized responses for each of the 20 conditions. The abscissa represents pixel size and the ordinate represents bar width. The Zaxis shows the normalized response rate. The normalized response rate for each condition was calculated by dividing response rate for that particular condition by the maximal firing rate of the cell to motion defined edges. Results are reported as a percentage of the maximal response based on a color chart (on the right of Fig. 6). Fig. 6A shows the most common pattern of responses. More than a third (35.6%) of the total cell sample vigorously responded to all bar widths. The response rate of these cells increased with increasing pixel size. The remainder of the cells that could be classified (38.5% of the total cell sample)

showed responses that were specific, or tuned, to bar width (Figs. 6B to F). Fig. 6B shows a cell (representing 7.9% of the sample) that only responded to edges that were wider than 2 degrees. Other cells (7.9%) only responded to bar widths that were smaller than 4 degrees (Fig. 6C). Cells represented in Figs. 6B and C responded preferentially to kinematograms composed of larger pixels when the stimuli had optimal bar width. This pattern was also found for cell 117 (Fig. 6D; representing 8.9% of the sample), except that this cell responded preferentially to bars having an intermediate width (2 and 4 degrees). The last two categories of cells (Figs. 6E and F) consist of cells that were highly selective to bar width. Fig. 6E shows cells (representing 9.9% of the sample) that responded exclusively to the largest bar width, and Fig. 6F represents a small proportion (3.9%) that responded exclusively to the smallest bar width (Fig. 6F). The condition that elicited the strongest firing rate in a given cell represents that cell's optimal response profile. Fig. 7A shows the distribution of pixel sizes that produced optimal responses. Nearly three-quarters (72.0%) of the cells maximally fired when stimuli were made of the largest pixels (1 degree). In addition, more than one-quarter of the cells (26.7%) responded preferentially to stimuli made of the second-largest pixel size (0.5 degrees). A single cell responded maximally to stimuli made of the intermediate (0.25 degrees)

88

BR A IN RE S EA RCH 1 1 87 ( 20 0 8 ) 8 2 –94

correlation between bar width and optimal spatial frequency was not significant (r = −0,085, p N 0.60). The correlation between bar width and receptive field size was not significant (r = − 0.037, p N 0.75). The correlation between bar width and bandwidths of band-pass cells was not significant (r = −0.076, p N 0.64). Therefore, the present findings indicate that receptive field sizes and spatial frequency tunings were not related to the spatial characteristics of optimal edges defined by motion. More than one-quarter (25.9%) of the cells were excluded from the classification into six categories because they did not show a distinct pattern of response to bar width and/or pixel size. These cells either fired at a similar rate in nearly all conditions or rather presented a seemingly random pattern of response across the different conditions. Fig. 8 presents the response profile of such a cell. First, this cell had a clear bandpass spatial frequency tuning function (Fig. 8A), with an optimal response at 0.05 c./degree and a large bandwidth (3.35 octaves). Fig. 8B presents the cell's peristimulus time histograms to the different kinematograms. No particular pattern of response appears, either to bar width or to pixel size. This is more clearly illustrated in Fig. 8C, where a high rate of response is seen for most conditions.

3.

Fig. 7 – Distribution of the optimal pixel size (A) and of the optimal bar widths (B) of PLLS cells stimulated by edges defined by motion. Almost all cells responded optimally to the larger pixel sizes (0.5 and 1 degrees). A large proportion of the cells (42.7%) responded maximally to the largest (8 degrees) bar width but it is interesting to note that nearly one-quarter of the cells (24.0%) responded maximally to the smallest (1 degree) bar width.

pixel size, and one cell (shown in Fig. 8B) responded weakly to kinematograms made of the two smallest pixel sizes (0.06 and 0.12 degrees). This distribution highlights the pixel size effect previously described. A simple regression analysis of the relationship between pixel size and either optimal spatial frequency or receptive field size did not reveal any significant correlations (r = −0.004, p N 0.90; r = −0.11, p N 0.3, respectively). The distribution of responses to bar width is shown in Fig. 7B. Although a large proportion of cells (42.7%) responded maximally to the largest (8 degrees) width, nearly one-quarter of them (24%) also responded maximally to the smallest width (1 degree). The remaining cells (33.3%) optimally responded to the intermediate widths (2 and 4 degrees). The relationships between the preference for a particular bar width and other receptive field characteristics of the cells were assessed. The

Discussion

The main objective of the present study was to investigate whether PLLS neurons can be driven by stimuli composed of edges that only become apparent when the constituting elements are moving. The results indicate that these neurons respond very well to edges defined by motion, that they respond more vigorously when the texture composing the edges is made of larger elements, and that a number of cells are selective to the spatial properties of the stimuli subtended by the edges (i.e., the distance between the edges or the width of the horizontal bar). A corollary objective was to provide an extended description of the receptive field properties and of the spatial frequency tuning functions of PLLS cells.

3.1.

Receptive field properties

In agreement with previous findings (Zumbroïch et al., 1986; Rauschecker et al., 1987a), the results of the present study indicate that area PLLS shows an overrepresentation of the upper visual field and a substantial representation of the ipsilateral visual field. However, there are some important differences between our findings and those of Zumbroïch et al. (1986). The present study reports a lower mean RF size and a stronger relationship between size and azimuth eccentricity. One likely explanation for these differences is that nearly half of our RF centers were located within 5 degrees of the vertical meridian, a region that generally contains smaller receptive fields. We found a clear predominance of complex receptive fields, particularly of the end-stopped complex types. This finding is consistent with that of von Grünau et al. (1987). We also found in the present study that the majority of PLLS cells were binocular. This is in agreement with the results found in a number of other studies (Rauschecker et al., 1987a,b; von Grünau et al., 1987).

BR A IN RE S E A RCH 1 1 87 ( 20 0 8 ) 8 2 –9 4

3.2.

Spatial frequency tuning functions

The only other study that looked into the spatial frequency tuning of PLLS cells investigated the response properties of 13 neurons (Zumbroïch and Blakemore, 1987). In comparison to their findings, the present results were computed on 101 neurons and they show a slightly lower mean optimal spatial frequency (0.08 c./degree compared to 0.17 c./degree), a smaller number of low-pass units (19% compared to 31%), and a relatively narrower mean bandwidth (1.8 octaves com-

89

pared to 2.2 octaves). Although it was not explicitly stated, the lowest spatial frequency tested by Zumbroïch and Blakemore (1987) appears to be 0.03 c./degree (derived from the data shown in Figs. 2–4, and 10). The use of a lower spatial frequency in the present study could account for the discrepancies just noted, since some cells classified as low-pass in their study would likely have been classified as band-pass had they been tested with our protocol. In addition, their small sample size could also account for these differences. Both reports lead to the conclusion that PLLS cells respond to lower spatial frequencies than cells in most visual areas in the cat. Based solely on the mean optimal spatial frequency, PLLS area could be considered, along with area 21b (Tardif et al., 2000), as a very low spatial frequency analyzer. In comparison, extrastriate areas 18, 19, 21a, and PMLS (Movshon et al., 1978; Tardif et al., 1996; Tardif et al., 1997; Khayat et al., 2000) could be considered mid-range spatial frequency analyzers, and area 17, of course, being the high spatial frequency analyzer (Ikeda and Wright, 1975; Tardif et al., 2000). Furthermore, based on the mean spatial bandwidth, area PLLS is more similar to areas 19, 21b, and PMLS, because they have broader spatial bandwidths, than areas 17, 18, and 21a (Tardif et al., 2000). Zumbroïch and Blakemore (1987) found a significant correlation between bandwidth and optimal spatial frequency in area PMLS. The present study did not find such a relationship in the PLLS area. This suggests that area 17 is more strongly related to area PMLS than to area PLLS. At the neuroanatomical level, there is indeed evidence for stronger connections between areas 17 and PMLS then between areas 17 and PLLS (Scannell et al., 1995, 1999). The significant correlations between azimuthal eccentricity and spatial bandwidth and between azimuthal eccentricity and optimal spatial frequency in PLLS cells have important implications for the role of this area in perception. Cells with RFs near the area centralis respond optimally to higher spatial frequencies than cells with RFs further from this region. This is consistent with the hypothesis that the PLLS area is involved, at least to some degree, in the processing of optic flow fields (Li et al., 2000).

3.3.

Responses to illusory edges defined by motion

Each of the single cells recorded in area PMLS responded to edges defined by motion. Most of them were also sensitive to Fig. 8 – (A) Spatial frequency tuning function established for the dominant eye of an end-stopped complex cell. The stimulus was a high-contrast (50%) sinusoidal grating drifting in the optimal direction. Each point corresponds to the mean firing rate of the cell at a given spatial frequency. (B) Peristimulus time histograms representing the same cell's responses to kinematograms as a function of pixel size and bar width. (C) These responses are represented in a three dimensional plot to better depict the relationship between response rate, pixel size and bar width. Raw data of the response changes for pixel size and width are normalized to the maximum response. This cell vigorously responds to kinematograms but shows no coherent relationship between response rate and pixel size, regardless of bar width. PSTH duration: 2 s.

90

BR A IN RE S EA RCH 1 1 87 ( 20 0 8 ) 8 2 –94

specific properties of the stimuli, such as pixel size and distance between the edges (bar width). Almost every cell responded more vigorously to kinematograms made of larger pixels. This facilitating effect of increasing the size of the elements was also found in area 19 of the cat (Khayat et al., 2000), and in the striate cortex of the New World monkey (Bourne et al., 2002). Similarly, enlarging the size of texture elements making up a whole-field pattern increased cellular response in areas 17 (Casanova et al., 1995) and PMLS (Merabet et al., 2000) of the cat. Because pixels consist of squares with contours that are well defined and orthogonal to the direction of motion, it could be argued that orientation cues were also present. Thus, the processing of the edge defined by motion with these coarser texture elements could solicit orientationsensitive mechanisms in addition to motion-sensitive ones. However, this is unlikely for a number of reasons: first, because the texture elements are quite small with respect to RF size; second, because short oriented bars which do not extend to the suppressive regions of the RF even for area 17 do not show strong orientation selectivity (Henry et al., 1974; Chen et al., 2005); and finally, because these orientation cues were equally present in the static background and in the motiondefined bars. A large proportion of cells (74.1%) also modified their response rates as a function of the distance between the leading and trailing edges. This suggests that bars defined by motion can be integrated to form surfaces in much the same way as “real” bars. In addition, some cells (8.9%) responded in a transient manner to the leading and to the trailing edges of the bar, particularly when the distance between the two edges was large. In fact, the greater proportion of cells that responded optimally to larger distances between the edges could be accounted for by the summed responses evoked by these edges. Similar edge-detector profiles were also found in area 19 of the cat (Khayat et al., 2000). It is likely that the processing of edges defined by motion in area PLLS is largely dependent on the substantial Y-input to that area (Scannell et al., 1995, 1999).

4.

Experimental procedures .

Twelve cats of either sex that weighted from 2 to 4 kg were used in this study. All were in good health and there was no indication of malformation or pathology. Surgical interventions, manipulations, husbandry and all experimental protocols were carried out within the guidelines proposed by the Canadian Council of Animal Care and by the National Institutes of Health (USA). The experimental protocol was approved by the Animal Care Committee of the Université de Montréal and the animals came from a supplier approved by the University.

4.1.

Anaesthesia and surgery

The methods used for animal preparation, anaesthesia, surgery, optical preparation, and recording are described in detail in previous papers from the laboratory (Guillemot et al., 1993; Bergeron et al., 1998; Bacon et al., 2000; Khayat et al., 2000; Mimeault et al., 2002) and will only be briefly described herein. The day before the experimentation the cats received an intramuscular injection of dexamethasone (10 mg/kg; Veto-

quinol Canada Inc., Joliette, Canada) to reduce inflammation during surgery. On the day of recording, the cat was premedicated with a subcutaneous injection of acepromazine maleate (Atravet, 1.0 mg/kg) and an intramuscular injection of atropine (Atro-sol, 0.2 mg/kg; Ormond Veterinary Supply Ltd, Lancaster, Canada) to limit bronchial secretion. The anaesthesia was then induced by facemask inhalation of 5% isoflurane mixed with nitrous oxide and oxygen (N2O:O2, 70:30). The animal was then intubated with an endotracheal tube connected to a respiratory pump (Harvard, model 665). Respiratory rate and stroke volume were adjusted to maintain end tidal CO2 at a constant level (∼4.0%). During all surgical procedures, the animals were kept deeply anaesthetized by maintaining isoflurane levels between 1.0% and 2.0%. The animals were placed in a modified stereotaxic apparatus (Kopf Instruments, Tujunga, CA, USA) to avoid pressure on the eyes and to free the visual field of any obstruction. A small trepanation (20 mm2) was performed over the PLLS area between Horsley–Clarke coordinates A6-P4 and L10-20 (Palmer et al., 1978). The dura mater was folded back and an electrode was inserted latero-medially (penetration angle: 30°) in the central representation of the visual field of this area (Palmer et al., 1978). In order to reduce brain pulsation and protect the cortex from dehydration, the brain was covered with agar solution (4.0% agar in physiological saline). All pressure points and incision sites were routinely infiltrated with a local anaesthetic (Xylocaine 2.0%; Astra Pharma Inc., Mississauga, Canada). At the end of the surgery, the isoflurane level was progressively reduced (0.5% per 15 min) and throughout the recording session, the animals were maintained under anaesthesia (N2O:O2, 70:30; isoflurane, 0.5% of gaseous mixture). The absence of reflexes and a stable heart rate ensured that the level of anesthesia was sufficient. From that point on, the animals were paralyzed. The neuromuscular blockade of the extra ocular muscles was maintained by the continuous infusion, through a saphenous vein cannula, of a gallamine triethiodide (Flaxedil: 12.5 mg/kg/h; Rhône-Poulenc, Montréal, Canada) and D-tubocurarine (Tubarine: 1.3 mg/kg/h; Sigma, St Louis, MO, USA) mixture dissolved in a solution of lactated Ringer with dextrose (5.0%). During the recording session, the stability of the heart rate was constantly verified to ensure that the level of anaesthesia was sufficient. The electroencephalogram, which was verified regularly, showed slow-wave activity throughout the recording session. During neuromuscular blockade and throughout the recording session the animals were artificially ventilated. Body temperature was kept constant (38 °C) with the help of a heated water pad.

4.2.

Optical preparation and recording

The nictitating membranes were retracted by topical application of phenylephrine hydrochloride (Neo-synephrine, 0.1%; Winthrop Laboratories, Aurora, Canada). Pupils were routinely dilated by topical application (eye drops) of atropine (Atro-sol, 0.2 mg/kg; Ormond Veterinary Supply Ltd, Lancaster, Canada). To prevent eye dehydration and to improve image resolution, neutral contact lenses with a 3-mm artificial pupil were placed on each eye. Appropriate dioptric lenses were placed in front of the eyes of the animals, as determined by the direct

BR A IN RE S E A RCH 1 1 87 ( 20 0 8 ) 8 2 –9 4

ophthalmoscopic examination. To determine the relative position of the areae centrales, retinal landmarks (optic discs and blood vessels) were projected on a tangent screen located 57 cm from the nodal point of each eye (Fernald and Chase, 1971). The area centralis of each eye was considered to be located 16° medially and 7.5° below the iso-elevation line of the center of each optic disc (Bishop et al., 1962). The optical quality of the eyes was checked routinely before and after each quantitative protocol. Recording was performed with tungsten microelectrodes that had an impedance range of 1–3 MΩ measured at 1000 Hz. The neural spikes were conventionally amplified, displayed on an oscilloscope, isolated with the help of a time–amplitude discriminator, and then transferred to an audio monitor and a PC computer.

4.3.

Histology

Electrolytic lesions were made at each recording track. At the end of the experiment, the cat was deeply anaesthetized with 5% isoflurane and perfused through the heart with isotonic saline followed by formalin (10%). The brain was removed, placed in formalin, and prepared for histology. The blocks of tissue that contained the electrode tracks were sectioned coronally (40 μm) using a freezing microtome and then stained with cresyl violet. According to the maps of Palmer et al. (1978), all electrode penetrations were within the PLLS area.

4.4.

91

Michelson (1927) formula: Contrast = (max. luminance − min. luminance/max. luminance+ min. luminance)× 100. Each trial began with a gradual increase in the contrast of the grating (from 0% to 50%) over a period of 100 ms. The cell's response was not recorded during this period in order to prevent a transient response from the cell. Subsequently, gratings drifted for 1 s and an inter-stimulus interval of 10–15 s was introduced between trials to minimize adaptation effects. Between each trial, the screen returned to mean luminance (contrast: 0). Discharge rate was computed for 1 s prior to each grating presentation, when the screen was still at mean luminance. During the presentation of the drifting grating, a peristimulus time histogram was constructed for a period of 1 s. A Fourier analysis was conducted to assess the modulation of the response at the first harmonic of the drifting grating. The criteria of Skottun et al. (1991) were used to classify the receptive field organization. Briefly, cells that showed a modulated response at the first harmonic for spatial frequencies higher than their optimal spatial frequency were classified as simple or end-stopped simple. Cells that showed an

Stimuli

The stimuli were generated with a G3 Macintosh computer using VPixx software (VPixx Technologies Inc., Longueuil, Canada). They were back-projected by means of a LCD projector (Mitsubishi LVP-X100A; refresh rate was 75 Hz), on a translucent screen located 57 cm in front of the animal's eyes. The mean luminance of the stimulation field was 19.9 cd/m2 and the resolution of the image was 11.9 pixels/degree. For each cell, the stimulation field (length and width) covered the whole RF and was precisely positioned at the center of the RF of the dominant eye. The optimal stimulus parameters (direction and velocity) were determined using sinusoidal gratings. The positions, the limits, and the spatial organization of each RF were carefully investigated. To test for end-stopping, the size of the stimulus was varied systematically. The principal inclusion criteria for end-stopping was that the cell preferred an oriented stimulus of a specific length, whereby extending the stimulus out of the boundaries of the RF caused a clear decrease (more than 80% as evaluated by ear) in response rate. For each cell, the spatial frequency tuning function was assessed for the dominant eye using drifting sinusoidal gratings (0.01–1.28 c./degree). The different points on the spatial frequency tuning function were separated by about half an octave. Each spatial frequency was presented at least ten times in a pseudo-random fashion. The cell's firing rate was recorded using optimal parameters for direction and temporal frequency (2, 4, or 6 Hz), as determined by ear. In order to obtain the highest response rate, special care was taken to adjust the size (width and length) of the gratings to the size of the dominant eye's RF. Contrast was kept constant at 50% and was defined using the

Fig. 9 – (A–D) Schematized representations of typical kinematograms used in the experiment. The diameter of the actual stimuli subtended 60 degrees of visual angle; black and white textured elements (ratio 1:1) subtended 0.06, 0.12, 0.25, 0.5, or 1 degrees of visual angle. Shaded areas have been superposed onto the figure to delimit the pixels forming the moving bars, and therefore the edges defined by motion. In the actual display, bar and background were identical in every way so that static bars could not have been perceived. Bars always drifted at the optimal speed in the cells' preferred direction. The bars in (A, B) represent the smallest bar width and the bars in (C, D) represent the largest bar width (actual bar widths: 1, 2, 4, and 8 degrees). For the testing of complex cells, the bar covered the entire length of the background (B and D). For the testing of end-stopped simple and end-stopped complex cells, bar length was adjusted so as to obtain maximal responses (A and C).

92

BR A IN RE S EA RCH 1 1 87 ( 20 0 8 ) 8 2 –94

unmodulated response and an overall increase in their response component (mean firing rate) at the first harmonic were classified as complex or end-stopped complex. Each data point on the tuning functions of the complex and end-stopped complex cells correspond to the mean firing rate of the cell during stimulus presentation minus the mean firing rate during the blank presentation. For simple and endstopped simple cells, the data points on the tuning functions were plotted according to the fundamental Fourier response component from which the response during the blank presentation was subtracted. The tuning functions of all cells were fitted with either a Gaussian or a sixth-order polynomial function (Table Curve 2D, Systat Software Inc., Richmond, CA, USA). The optimal curve fit chosen had a strong correlation (r ≥ 0.9) with the data points and was used to evaluate spatial frequency and bandwidth. The spatial bandwidth of the tuning function was defined as the full width of the curve at half the amplitude for the optimal spatial frequency. Based on their spatial frequency tuning functions, cells were classified as either band-pass or low-pass. Band-pass cells responded most strongly or optimally to a given spatial frequency and presented a drastic attenuation in their response rates at higher and lower spatial frequencies. In contrast, low-pass cells showed attenuation in their response rates at high spatial frequencies, but not at low spatial frequencies. For each cell, the receptive field of the non-dominant eye was also tested at the optimal spatial frequency to evaluate ocular dominance. An ocular dominance index ranging from 1 (contralaterally driven) to 7 (ipsilaterally driven) was used. In order to classify the binocular cells among the five intermediate categories, an ocular dominance index was calculated for every cell using the following formula (ipsilateral / (ipsilateral + contralateral))× 100, in which ipsilateral is the response rate of the cell to the monocular stimulation of the ipsilateral eye and contralateral is the response rate of the cell to the monocular stimulation of the contralateral eye. The ocular dominance index thereby obtained allowed for classification using the following criteria: class 2: 1.0–20.0%, class 3: 21.0–40.0%, class 4: 41.0–60.0%, class 5: 61.0– 80.0%, class 6: 81.0–99.0%. Edges defined by motion were elicited by moving an oriented textured form (bar) against an identically oriented but static textured background (see Fig. 9). The texture of the oriented form and of the background consisted of randomly positioned light and dark squares (1:1 ratio of dark and light elements). The sizes of the textured elements were set at: 0.06 × 0.06 degrees, 0.12 × 0.12 degrees, 0.25 × 0.25 degrees, 0.5 × 0.5 degrees, or 1 × 1 degree. For simplicity, these patterns are referred to in the text as 0.06 degrees, 0.12 degrees, 0.25 degrees, 0.5 degrees, and 1 degree, respectively. Both the bar and background elements had the same mean luminance (19.9 cd/m2) and contrast (90%). Because the texture elements making up the oriented form and the background had the same orientation and density, the edges of the oriented textured form could only be defined by motion and were otherwise camouflaged. Thus, the edges of the bar could only be perceived when the pixel polarities were coherently moving in the same direction. The highest texture density pattern was similar to the one used by Casanova et al. (1995) and revealed, through the 2D Fourier power spectrum analyses, that all spatial frequencies and all orientations had the same amount of power.

In this experiment, only the dominant receptive fields were tested. Schematized representations of the kinematograms used in the experiment are shown in Fig. 9. For all cells, the background subtended 60 degrees in diameter and four different textured form (bar) widths were used (1, 2, 4, and 8 degrees: see Figs. 9A and B for small bar widths and Figs. 9C and D for large bar widths). For simple and complex receptive fields, the length of the bar covered the whole field (60 degrees; see Figs. 9B and D) but for end-stopped receptive fields special care was taken to adjust the length of the bar in order to obtain the highest response rate (see Figs. 9A and C). The bar drifting speed (15– 30 degrees/s) at the cell's optimal direction was selected in order to elicit the highest response rate. Each of the 20 conditions (four bar widths × five pixel sizes) was tested 10 times in a pseudorandom fashion. An inter-stimulus delay of 10–15 s was used to avoid habituation. For each condition, a peristimulus time histogram having 500 bins and a binwidth of 4–6 ms was derived from the cellular responses.

Acknowledgments This study was supported by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) awarded to J.-P. Guillemot and F. Lepore and by a Canada Research Chair awarded to F. Lepore.

REFERENCES

Bacon, B.A., Lepore, F., Guillemot, J.-P., 2000. Neurons in the posteromedial lateral suprasylvian area of the cat are sensitive to binocular positional depth cues. Exp. Brain Res. 134, 464–476. Benedek, G., Mucke, L., Norita, M., Albowitz, B., Creutzfeldt, O.D., 1988. Anterior ectosylvian visual area (AEV) of the cat: physiological properties. Prog. Brain Res. 75, 245–255. Bergeron, A., Tardif, E., Lepore, F., Guillemot, J.-P., 1998. Spatial and temporal matching of receptive field properties of binocular cells in area 19 of the cat. Neuroscience 86, 121–134. Bishop, P.O., Kozak, W., Vakkur, G.J., 1962. Some quantitative aspects of the cat's eye: axis and plane of reference of visual field coordinates and optics. J. Physiol. Lond. 163, 466–502. Blakemore, C., Zumbroïch, T.J., 1987. Stimulus selectivity and functional organization in the lateral suprasylvian visual cortex of the cat. J. Physiol. (Lond.) 389, 569–603. Bourne, J.A., Tweedale, R., Rosa, M.G., 2002. Physiological responses of New World monkey V1 neurons to stimuli defined by coherent motion. Cereb. Cortex 12, 1132–1145. Brecht, M., Singer, W., Engel, A.K., 1998. Correlation analysis of corticotectal interactions in the cat visual system. J. Neurophysiol. 79, 2394–2407. Bressoud, R., Innocenti, G.M., 1999. Typology early differentiation and exuberant growth of a set of cortical axons. J. Comp. Neurol. 406, 87–108. Camarda, R., Rizzolatti, G., 1976. Visual receptive fields in the lateral suprasylvian area (Clare–Bishop area) of the cat. Brain Res. 101, 427–443. Casanova, C., Savard, T., Nordmann, J.P., Molotchnikoff, S., Minville, K., 1995. Comparison of the responses to moving texture patterns of simple and complex cells in the cat's area 17. J. Neurophysiol. 74, 1271–1286.

BR A IN RE S E A RCH 1 1 87 ( 20 0 8 ) 8 2 –9 4

Chen, G., Dan, Y., Li, C.Y., 2005. Stimulation of non-classical receptive field enhances orientation selectivity in the cat. J. Physiol. (Lond.) 564, 233–243. Fernald, R., Chase, R., 1971. An improved method for plotting retinal landmarks and focusing eyes. Vis. Res. 11, 95–96. Frost, B.J., 1985. Neural Mechanisms for detecting object motion and figure–ground boundaries contrasted with self-motion detecting systems. In: Ing, D., Jeannerod, M., Lee, D. (Eds.), Brain Mechanism of Spatial Vision. Martinus Nijhoft, Dordrecht, pp. 415–449. Guido, W., Tong, L., Spear, P.D., 1990. Afferent bases of spatial- and temporal-frequency processing by neurons in the cat's posteromedial lateral suprasylvian cortex: effects of removing areas 17, 18 and 19. J. Neurophysiol. 64, 1636–1651. Guillemot, J.-.P., Paradis, M.-.C., Samson, A., Ptito, M., Richer, L., Lepore, F., 1993. Binocular interaction and disparity coding in area 19 of visual cortex in normal and split-chiasm cats. Exp. Brain Res. 93, 405–417. Gulyas, B., Orban, G.A., Duysens, J., Maes, H., 1987. The suppressive influence of moving textured backgrounds on responses of cat striate neurons to moving bars. J. Neurophysiol. 57, 1767–1791. Gulyas, B., Spileers, W., Orban, G.A., 1990. Modulation by a moving texture of cat area 18 neuron responses to moving bars. J. Neurophysiol. 63, 404–423. Hammond, P., MacKay, D.M., 1977. Differential responsiveness of simple and complex cells in cat striate cortex to visual texture. Exp. Brain Res. 30, 275–296. Hardy, S.C., Stein, B.E., 1988. Small lateral suprasylvian cortex lesions produce visual neglect and decreased visual activity in the superior colliculus. J. Comp. Neurol. 273, 527–542. Henry, G.H., 1977. Receptive field classes of cells in the striate cortex of the cat. Brain Res. 133, 1–28. Henry, G.H., Bishop, P.O., Coombs, J.S., 1967. Inhibitory and subliminal excitatory receptive fields of units in cat striate cortex. Vis. Res. 9, 1289–1296. Henry, G.H., Bishop, P.O., Dreher, B., 1974. Orientation, axis and direction as stimulus parameters for striate cells. Vis. Res. 14, 767–777. Hubel, D.H., Wiesel, T.N., 1962. Receptive fields, binocular interactions and functional architecture in the cat's visual cortex. J. Physiol. (Lond.) 160, 106–154. Hubel, D.H., Wiesel, T.N., 1965. Receptive fields and functional architecture in two non-striate visual areas (18 and 19) of the cat. J. Neurophysiol. 28, 229–289. Hubel, D.H., Wiesel, T.N., 1969. Visual area of the lateral suprasylvian gyrus (Clare–Bishop area) of the cat. J. Physiol. (Lond.) 202, 251–260. Ikeda, H., Wright, M.J., 1975. The latency of visual cortical neurones in area 17 in the cat to visual stimuli with reference to the sustained (X) and transient (Y) and ‘simple’ and ‘complex’ cell classification. J. Physiol. 245, 114–115. Innocenti, G.M., Aggoun-Zouaoui, D., Lehmann, P., 1995. Cellular aspects of callosal connections and their development. Neuropsychologia 33, 961–987. Julesz, B., 1971. Foundations of Cyclopean Perception. University of Chicago Press, Chicago. Kawamura, K.F., Hashikawa, T., 1978. Cell bodies of origin of reticular projections from the superior colliculus in the cat: an experimental study with the use of horseradish peroxidase as a tracer. J. Comp. Neurol. 182, 1–15. Khayat, P.S., Saint-Amour, D., Molotchnikoff, S., Lepore, F., Guillemot, J.-P., 2000. Cellular response to texture and form defined by motion in area 19 of the cat. Eur. J. Neurosci. 12, 1727–1738. Kiefer, W., Kruger, K., Strauss, G., Berlucchi, G., 1989. Considerable deficits in the detection performance of the cat after lesion of the suprasylvian visual cortex. Exp. Brain Res. 75, 208–212. Kim, J.N., Mulligan, K., Sherk, H., 1997. Simulated optic flow and extrastriate cortex. J. Neurophysiol. 77, 554–561.

93

Komatsu, Y., Shibuki, K., Toyama, K., 1983. Eye movement-related activities in cells of the lateral suprasylvian cortex of the cat. Neurosci. Lett. 41, 271–276. Krüger, K., Kiefer, W., Groh, A., Dinse, H.R., von Seelen, W., 1993. The role of the lateral suprasylvian visual cortex of the cat in object–background interactions: permanent deficits following lesions. Exp. Brain Res. 97, 40–60. Li, B., Li, B.W., Chen, Y., Wang, L.H., Diao, Y.C., 2000. Response properties of PMLS and PLLS neurons to simulated optic flow patterns. Eur. J. Neurosci. 12, 1534–1544. Lomber, S.G., Payne, B.R., Cornwell, P., Long, K.D., 1996. Perceptual and cognitive visual functions of parietal and temporal cortices in the cat. Cereb. Cortex 6, 673–695. Mason, R., 1976. Responses of cells in the dorsal lateral geniculate complex of the cat to textured visual stimuli. Exp. Brain Res. 25, 323–326. Marcar, V.L., Xiao, D.K., Raiguel, S.E., Maes, H., Orban, G.A., 1995. Processing of kinetically defined boundaries in the cortical motion area MT of the macaque monkey. J. Neurophysiol. 74, 1258–1270. Merabet, L., Minville, K., Ptito, M., Casanova, C., 2000. Responses of neurons in the cat posteromedial lateral suprasylvian cortex to moving texture patterns. Neuroscience 97, 611–623. Michelson, A.A., 1927. Studies in Optics. University of Chicago Press, Chicago. Mimeault, D., Paquet, V., Lepore, F., Guillemot, J.-P., 2002. Phase-disparity coding in extrastriate area 19 of the cat. J. Physiol. (Lond.) 545, 987–996. Morrone, M.C., Di Stefano, M., Burr, D.C., 1986. Spatial and temporal properties of neurons of the lateral suprasylvian cortex of the cat. J. Neurophysiol. 56, 969–986. Movshon, J.A., Thompson, I.D., Tolhurst, D.J., 1978. Spatial and temporal contrast sensitivity of neurones in areas 17 and 18 of the cat's visual cortex. J. Physiol. (Lond.) 283, 101–120. Niida, T., Stein, B.E., McHaffie, J.G., 1997. Response properties of corticotectal and corticostriatal neurons in the posterior lateral suprasylvian cortex of the cat. J. Neurosci. 17, 8550–8565. Nothdurft, H.C., 1993. The role of features in preattentive vision: comparison of orientation motion and colour cues. Vis. Res. 33, 1937–1958. Ogasawara, K., McHaffie, J.G., Stein, B.E., 1984. Two visual corticotectal systems in cat. J. Neurophysiol. 52, 1226–1245. Olavarria, J.F., DeYoe, E.A., Knierim, J.J., Fox, J.M., Van Essen, D.C., 1992. Neural responses to visual texture patterns in middle temporal area of the macaque monkey. J. Neurophysiol. 68, 164–181. Palmer, L.A., Rosenquist, A.C., Tusa, R.J., 1978. The retinotopic organization of lateral suprasylvian visual areas in the cat. J. Comp. Neurol. 177, 237–256. Pasternak, T., Horn, K.M., Maunsell, J.H., 1989. Deficits in speed discrimination following lesions of the lateral suprasylvian cortex in the cat. Vis. Neurosci. 3, 365–375. Payne, B.R., 1993. Evidence for visual cortical area homologs in cat and macaque monkey. Cereb. Cortex 3, 1–25. Payne, B.R., Lomber, S.G., Geeraerts, S., van der Gucht, E., Vandenbussche, E., 1996. Reversible visual hemineglect. Proc. Natl. Acad. Sci. U. S. A. 93, 290–294. Raczkowski, D.F., Rosenquist, A.C., 1983. Connections of the multiple visual cortical areas with the lateral posterior–pulvinar complex and adjacent thalamic nuclei in the cat. J. Neurosci. 3, 1912–1942. Rauschecker, J.P., von Grünau, M.W., Poulin, C., 1987a. Centrifugal organization of direction preferences in the cat's lateral suprasylvian visual cortex and its relation to flow field processing. J. Neurosci. 7, 943–958. Rauschecker, J.P., von Grünau, M.W., Poulin, C., 1987b. Thalamo-cortical connections and their correlation with receptive field properties in the cat's lateral suprasylvian visual cortex. Exp. Brain Res. 67, 100–112.

94

BR A IN RE S EA RCH 1 1 87 ( 20 0 8 ) 8 2 –94

Roucoux, A., Crommelinck, M., 1976. Eye movements evoked by superior colliculus stimulation in the alert cat. Brain Res. 106, 349–363. Rudolph, K.K., Pasternak, T., 1996. Lesions in cat lateral suprasylvian cortex affect the perception of complex motion. Cereb. Cortex 6, 814–822. Scannell, J.W., Blakemore, C., Young, M.P., 1995. Analysis of connectivity in the cat cerebral cortex. J. Neurosci. 15, 1463–1483. Scannell, J.W., Sengpiel, F.F., Tovee, M.J., Benson, P.J., Blakemore, C., Young, M.P., 1996. Visual motion processing in the anterior ectosylvian sulcus of the cat. J. Neurophysiol. 76, 895–907. Scannell, J.W., Burns, G.A., Hilgetag, C.C., O'Neil, M.A., Young, M.P., 1999. The connectional organization of the cortico-thalamic system of the cat. Cereb. Cortex 9, 277–299. Sherk, H., Mulligan, K., Kim, J.N., 1997. Neuronal responses in extrastriate cortex to objects in optic flow fields. Vis. Neurosci. 14, 879–895. Skottun, B.C., De Valois, R.L., Grosse, D.H., Movshon, J.A., Albrecht, D.G., Bonds, A.B., 1991. Classifying simple and complex cells on the basis of response modulation. Vis. Res. 31, 1079–1086. Spear, P.D., 1991. Functions of extrastriate visual cortex in non-primate species. In: Leventhal, A.G. (Ed.), Vision and Visual Dysfunction. The Neural Basis of Visual Function, vol. 4. CRC Press, Boca Raton. Spear, P.D., Baumann, T.P., 1975. Receptive-field characteristics of single neurons in lateral suprasylvian visual area of the cat. J. Neurophysiol. 38, 1403–1420. Stein, B.E., 1978. Development and organization of multimodal representation in cat superior colliculus. Fed. Proc. 37, 2240–2245. Stone, J., 1983. Parallel Processing on the Visual System: The Classification of Retinal Ganglion Cells and its Impact on the Neurobiology of Vision. Plenum Press, New York. Stone, J., Dreher, B., Leventhal, A., 1979. Hierarchical and parallel mechanisms in the organization of visual cortex. Brain Res. 180, 345–394.

Tanaka, K., Ohzawa, I., Ramoa, A.S., Freeman, R.D., 1987. Receptive field properties of cells in area 19 of the cat. Exp. Brain Res. 65, 549–558. Tardif, E., Bergeron, A., Lepore, F., Guillemot, J.-P., 1996. Spatial and temporal frequency tuning and contrast sensitivity of single neurons in area 21a of the cat. Brain Res. 716, 219–223. Tardif, E., Richer, L., Bergeron, A., Lepore, F., Guillemot, J.-P., 1997. Spatial resolution and contrast sensitivity of single neurons in area 19 of split-chiasm cats: a comparison with primary visual cortex. Eur. J. Neurosci. 9, 1929–1939. Tardif, E., Lepore, F., Guillemot, J.-P., 2000. Spatial properties and direction selectivity of single neurons in area 21b of the cat. Neuroscience 97, 625–634. Toda, H., Yoshizawa, T., Takagi, M., Bando, T., 2001. The properties of convergence eye movements evoked from the rostral and caudal lateral suprasylvian cortex in the cat. Neurosci. Res. 39, 359–367. von Grünau, M., Frost, B.J., 1983. Double-opponent-process mechanism underlying RF-structure of directionally specific cells of cat lateral suprasylvian visual area. Exp. Brain Res. 49, 84–92. von Grünau, M.W., Zumbroïch, T.J., Poulin, C., 1987. Visual receptive field properties in the posterior suprasylvian cortex of the cat: a comparison between the areas PMLS and PLLS. Vis. Res. 27, 343–356. Yin, T.C., Greenwood, M., 1992. Visuomotor interactions in responses of neurons in the middle and lateral suprasylvian cortices of the behaving cat. Exp. Brain Res. 88, 15–32. Zumbroïch, T.J., Blakemore, C., 1987. Spatial and temporal selectivity in the suprasylvian visual cortex of the cat. J. Neurosci. 7, 482–500. Zumbroïch, T.J., Price, D.J., Blakemore, C., 1988. Development of spatial and temporal selectivity in the suprasylvian visual cortex of the cat. J. Neurosci. 8, 2713–2728. Zumbroïch, T.J., von Grünau, M., Poulin, C., Blakemore, C., 1986. Differences of visual field representation in the medial and lateral banks of the suprasylvian cortex (PMLS/PLLS) of the cat. Exp. Brain Res. 64, 77–93.