Exp Brain Res (2006) 170: 457–463 DOI 10.1007/s00221-005-0229-1 R ES E AR C H A RT I C L E Alexandra Séverac Cauquil Æ Yves Trotter Margot J. Taylor At what stage of neural processing do perspective depth cues make a difference? Received: 19 January 2005 / Accepted: 13 September 2005 / Published online: 24 November 2005 Springer-Verlag 2005 Abstract The present study investigated the cortical processing of three-dimensional (3D) perspective cues in humans, to determine how the brain computes depth from a bidimensional retinal image. We recorded visual evoked potentials in 12 subjects in response to flat and in-perspective stimuli, which evoked biphasic potentials over posterior electrodes. The first, positive component (P1, at 90 ms) was not sensitive to perspective, while the second, negative peak (N1 at 150 ms) was significantly larger for 3D stimuli, regardless of attention. The amplitude increase due to perspective was seen on all posterior electrodes, but was largest over the right hemisphere, particularly at parietal sites. Source modeling low-resolution electromagnetic tomography (LORETA) confirmed that among the different areas participating in two- and three-dimensional stimuli processing, the right parietal source is the most enhanced by perspective depth cues. We conclude that the extraction of depth from perspective cues occurs at a second level of stimulus processing, by increasing the activity of the regions involved in 2D stimuli processing, particularly in the right hemisphere, possibly through feedback loops from higher cortical areas. These modulations would participate in the fine-tuned analysis of the 3D features of stimuli. Keywords Depth perception Æ Human electrophysiology Æ Linear perspective Æ Source localization Abbreviations 2D: Two-dimensional Æ 3D: Three-dimensional Æ CIP: Caudal part of the lateral bank of the intraparietal sulcus Æ LORETA: Low-resolution electromagnetic tomography Æ RDS: A. Séverac Cauquil (&) Æ Y. Trotter Æ M. J. Taylor Centre de Recherche Cerveau et Cognition, UMR 5549 UPS/CNRS, Université Paul Sabatier Toulouse III, 133 route de Narbonne, 31062 Toulouse cedex 4, France E-mail: [email protected] Tel.: +33-5-62172835 Fax: +33-5-62172809 Random-dot stereogram Æ STS: Superior temporal sulcus Æ VEPs: Visual evoked potentials Introduction Three-dimensional (3D) vision relies on monocular cues such as perspective, shading and texture as well as binocular cues; understanding the chronology and location of 3D processing is a major research area in visual science. The cortical processing of monocular depth cues has been less studied when compared with the processing of binocular cues such as disparity. Yet monocular depth cues are basic and critical to normal vision, for example, we still have efficient depth perception if we close one eye. As shown in monkeys, neurons sensitive to retinal disparity are found from the striate cortex (Trotter 1995; Gonzalez and Perez 1998; Cumming and Parker 2000) to the frontal eye fields (Ferraina et al. 2000), and most parts of the dorsal pathway are involved in depth perception, based on both monocular and binocular cues (Tsutsui et al. 2001, 2002; Uka and DeAngelis 2003). The ventral pathway is also involved in disparity coding (Janssen et al. 1999, 2000; Uka et al. 2000). In human fMRI studies, areas in the intra-parietal sulcus were found activated by stimuli with monocular depth cues such as shading (Taira et al. 2001) and shading plus perspective (Shikata et al. 2001). In a study of Necker cube perception, Inui et al. (2000) suggested that ‘‘monocular stereopsis’’ activates the bilateral occipito-temporal gyrus as well as premotor and lateral parietal cortices. Moore and Engel (2001) found that the lateral occipital complex (LOC), the lateral and ventral region of the human occipital cortex, is responsive to both the knowledge-based and the image-based monocular components of a 3D object structure. They presented strictly identical stimuli, which gave rise to a significantly stronger activity when, owing to gray-scale primes, the stimuli were perceived as a volume rather than as a flat object. Thus, the LOC could be a region where information from different 458 sources is integrated to support the subjective experience of perceiving volume. Another fMRI study demonstrated an increase of activity in this ventral region (plus parietally) for the presentation of 3D versus 2D shape stimuli (Murray et al. 2003). In contrast, Kourtzi and Kanwisher (2000) failed to find a stronger activity in the LOC, when processing 3D rather than 2D objects, and claimed that the region was more involved in processing shape than depth. For binocular depth cues, both static and dynamic solid stereograms induce fMRI activity in the dorsal occipital portion and the superior parietal lobule with right hemisphere dominancy (Iwami et al. 2002). Random-dot stereogram (RDS) presentation induced a widespread activity within the extrastriate cortex, particularly area V3A (Backus et al. 2001). Very few studies have used visual evoked potentials (VEPs), which provide detailed information on the timing of cortical activation, to study monocular depth perception. Jeffreys (1996) found that the ‘‘later negative potential’’ present on occipital VEPs around 170 ms was sensitive to depth cues, particularly monocular depth cues: the activity recorded on the three occipital electrodes investigated was consistently enhanced (seen in 42 out of 49 subjects) although not statistically tested. The largest amplitude was recorded over the right electrode. Skrandies and Jedynak (1999) reported that in subjects learning to see sub-threshold 3D stimuli, the pattern of activation shifted towards the right hemisphere. Previous VEPs studies using RDS have noted the opposite (Manning et al. 1992) or the absence of (Lehmann and Julesz 1978) hemisphere dominancy. In summary, most of the aforementioned studies suggest that the exposure to visual stimuli with monocular perspective depth cues could increase cortical activity in occipital, temporal and parietal regions. Thus, we hypothesized that with 3D stimuli there would be increased activation in posterior temporal regions compared to 2D stimuli, and that this would occur early in processing. We investigated the precise timing and the organization of specific cortical network of 3D processing in humans using VEPs evoked by 2D and 3D shape stimuli (with or without perspective). This method allowed us to determine the extent of the activity and the time-structure of the monocular-based processing of 3D objects. Visual stimuli and procedure Figure 1 presents the procedure and the set of stimuli, 20 different monochromatic images: ten with perspective monocular cues (noted 3D; e.g. a cube) and ten matching without (noted 2D; e.g. a flattened cube), the 2D being the flat projection of the 3D. No shading or other monocular cues were present on the stimuli, which were monochromatic (bright green on a black background) and the same size across 2D and 3D sets (20 wide). They were presented binocularly one at a time on a computer screen 60 cm away from the seated subject. Subjects were instructed to maintain visual fixation on the central target during stimulus presentation. A session contained 12 blocks of 80 stimuli: the set of 20 stimuli (Fig. 1) presented four times in a pseudorandom order. Stimuli were presented for 200 ms, with an inter-stimulus interval varying between 700 and 1,400 ms. In order to control attention effects linked to the task, during the blocks, subjects were asked to respond by pressing the mouse button whenever they saw either a 2D (‘‘2D attended’’ condition) or a 3D stimulus (‘‘3D attended’’ condition). Therefore, during six blocks, subjects responded to 2D stimuli, and during six other blocks they responded to 3D stimuli; this order 200 ms + 700-1400 ms 200-ms stimulus presentation + x 40 = 1 block inter-stimulus interval varying between 700 and 1400 ms Materials and methods Experimental subjects Twelve adults were included in the study (38.1 years±2.8; four females); all were right-handed and had normal or corrected to normal vision, and gave informed consent. The experiments were performed in compliance with institutional guidelines, and the COPE ethics committee Centre National de la Recherche Scientifique (CNRS) approved the protocol. Fig. 1 Upper part: experimental procedure. Lower part complete set of the stimuli, ten 2D (upper two rows) and ten matched 3D images 459 was counterbalanced across subjects. Short pauses were given between blocks. VEP recording The VEPs were recorded with a NeuroScan system 4.1 (NeuroScan, Sterling, VA, USA, www.neuro.com) from 33 EEG electrodes applied with an EasyCap electrode cap (Falk Minow Services, Herrsching-Breitbrunn, Germany, www.easycap.de) according to the 10–10 system. Cz served as the reference lead during data collection. The EOG was monitored with three electrodes, at the outer canthi and the left supra-orbital ridge. The epochs lasted 1.1 s, with a 100 ms pre-stimulus baseline, and were recorded with a bandpass filter of 0.1–100 Hz. The EEG was digitally re-referenced offline to an average-reference montage. Trials were rejected for EOG or movement artifact (>±90 lV). For each trial, the baseline, estimated as the mean amplitude of the 100 ms pre-stimulus epoch, was subtracted and then trials were averaged according to the stimulus categories. The components of interest were P1 and N1 measured at the posterior electrodes: O1, O2, PO9, PO10, P3, P4, P7, P8, TP9 and TP10, where the largest and clearest VEPs were observed. Broadly distributed frontal activity was seen but likely reflected the other pole of the posterior activity (George et al. 1996). Analyses of variance with repeated measures were conducted on latency and amplitude data by depth stimuli (two levels: 3D and 2D), attention condition (two levels: 3D and 2D attended), electrodes (five levels) and hemisphere (two levels). When appropriate (e.g. with repeated measures, to reduce non-sphericity of the variance/covariance matrix) GreenhouseGeisser adjustments were used. Post hoc tests using the methods of contrasts investigated the electrode effects. Source modeling was performed using LORETA, lowresolution tomography (www.unizh.ch/keyinst/NewLORETA; (Pascual-Marqui et al. 1994). LORETA computes the smoothest 3D intracerebral distribution of current density (in lA/mm2) by minimizing the norm of the Laplacian of the current vectors without assuming a specific a priori location or number of sources. LORETA computations are restricted to 2,394 voxels in the cortical gray matter and hippocampus, based on the Talairach atlas (Talairach and Tournoux 1988), giving a 7 mm voxel resolution. In the version used here, the head model was three-shell spherical, using reported EEG electrode coordinates (Towle et al. 1993), based on several individual measurements from which the mean best-fitting sphere was determined. Then a surface-fitting algorithm was performed to transfer the best-fitting sphere to the mean MRI brain. The source locations were given as xyz coordinates, with an estimate of the corresponding brain region and Brodmann area obtained after the transformation of MNI coordinates into Talairach space. It has been shown that dipole localization on grand averages was a valid approach and could be more stable than on individual averages because of the better signalto-noise ratio (Giard et al. 1994; Pantev et al. 1995; Clark and Hillyard 1996). We therefore opted for a 2-step process: first, to work with a better signal-to-noise ratio, we entered our grand mean data into LORETA, which gave us a solution at both the P1 and N1 latencies. Second, to be able to assess the statistical significance of the modeling, we ran LORETA on the individual subject’s data and then performed a paired t-test voxel by voxel, on the set of 2,394 voxels, and for each time frame, to compare individual LORETA models in both the 2D and the 3D conditions. We obtained a statistical parametric map and used P<0.0001 as a correction for multiple testing. Results The stimuli evoked a biphasic potential (P1–N1) recorded over the posterior scalp. The positive deflection P1 was not seen for inferior electrodes PO9, PO10, TP9 and TP10, but appeared clearly on P3, P4, P7, P8 and O1, O2. The general distribution, as seen on grand average maps, was fairly symmetrical (Fig. 2), but with an asymmetry to the right at P1 latency. Stimuli with 3D perspective evoked a much larger activity than did 2D stimuli, whereas the task did not influence the recorded activity: the presentation of 2D stimuli evoked a similar response whether attended or not, as did 3D stimulation. At the N1 latency a simultaneous large frontal positivity was recorded. Peak analyses The P1 and N1 latencies did not differ between 2D and 3D stimuli: P1 at 90.1±1.4 ms and N1 at 148.8±4.6 ms. Neither the attention task (target vs. non-target) nor the P1 N1 attended P1 N1 non attended P1 N1 P1 N1 -7.9 -7 -6.2 -5.3 -4.5 -3.6 -2.7 -1.9 -1 0 0.7 1.6 2.4 3.3 4.2 5 5.9 (µV) Fig. 2 Grand average scalp voltage maps (n=12) at P1 (90 ms) and N1 (150 ms) mean latencies, to the 2D and 3D stimuli (left and right panels, respectively), in the two attended/non-attended conditions (upper and lower panels, respectively). The 2D and 3D labels were stimuli in the corresponding category 460 presence of perspective influenced latencies significantly (for P1 P=0.582 and P=1; for N1 P=0.101 and P=0.508, for the ‘‘attention’’ and ‘‘depth’’ factors, respectively). Figure 3 presents the grand averages of waveforms obtained for the four conditions, 2D/3D, attended/non-attended, at the electrodes where the effect was the strongest. The larger waveforms were seen at the occipital and parieto-occipital electrodes. Waveforms obtained in the target and non-target conditions were perfectly superimposable until the end of the N1 component (Fig. 3). Then, 200 ms after the stimulus onset, an influence of the attentional task could be observed, especially on the left hemisphere electrodes, but as it occurred after the early biphasic event sensitive to perspective, this was not further studied here. Figure 4 presents bar graphs (mean±SEM) of P1 and N1 amplitudes for the electrodes of interest. The P1 amplitude was similar for 2D and 3D stimuli, regardless of the task, but it was significantly larger over the right hemisphere (‘‘hemisphere’’ factor P<0.05) for the parietal electrodes (P8 vs. P7: +92%, P4 vs. P3: +74%; ‘‘electrode·hemisphere’’ interaction: P<0.05). In contrast, N1 was clearly larger (‘‘depth’’ factor, P<0.001) for the 3D than for the 2D stimuli ( 4.9±0.8 vs. 3.4±0.6 lV) and the increase due to perspective cues was maximal over the parietal, mostly the right parietal (see bar graphs on Fig. 4: +30% on P3 Fig. 3 Grand average (n=12) waveforms recorded over the six electrodes of interest in the four conditions: 2D (green) and 3D (purple) for attended (dark) and non-attended (pale) stimuli and +150% on P4), rather than ventral sites (+10 and +37% at TP and PO sites, respectively, not shown, and +50% at O1, O2 and P7, P8 sites) (‘‘electrode·depth’’ interaction: P<0.003; ‘‘depth·hemisphere’’ interaction, P<0.04). Source modeling As it was evident that the attended/non-attended status of the stimulus did not yield differences, we collapsed the data from the two conditions to improve the signal-tonoise ratio for the source analyses. Grand average modeling At both P1 and N1 latencies, the proposed model of the grand average data presented the same pattern, with the same source localization, for both 2D and 3D stimuli. The model included five main regions (Table 1): two lateral, symmetrical sources at the level of occipitotemporal junctions (left BA39 and right BA22), two medial: one occipital (right/left BA18) and one occipitoparietal (right/left BA7), and one in the region of the right superior parietal lobule (right BA7). A further source was located in the region of premotor area (right/ left BA6). P3 P4 P1 P8 N1 P7 O2 O1 + - 1 µV 100 ms 2D attended 2D non-attended 3D attended 3D non-attended 461 then, between 162 and 170 ms, below in the temporal lobe (x=46, y= 60 and z=22). This difference between 2D and 3D models occurred during the N1 time range and was congruent with the larger N1 increase we observed over the right parietal and parieto-occipital electrodes. No statistical difference was found either in the left hemisphere or in the medial parietal region, between 2D and 3D stimuli. P4 P3 P8 P7 Discussion O2 O1 1 γV 2-D attended 3-D attended 2-D non-attended 3-D non-attended Fig. 4 Bar graphs representing the mean amplitude±SEM of P1 (the four left-most bars at each of the six electrode locations) and N1 (four bars on the right for each electrode) amplitudes for the four conditions at each electrode of interest Subject by subject modeling The sources modeled on grand averages were found in most of the individual modeling (Table 1). We analyzed the individual results of the LORETA inversion using a paired t-test design, at each time frame and for each voxel. This statistical approach assesses the ‘‘depth effect’’ without any a priori assumption on the timing and locations of the sources as it is performed at each time frame and for each voxel (see Materials and methods). For P<0.0001, one cluster of significantly different voxels emerged at N1 latency in the right hemisphere, around the region of the angular gyrus (see Fig. 5), appearing first (156–162 ms post-stimulus) slightly above, in the parietal lobe (x=46, y= 67 and z=36), Adding 3D perspective cues to visual stimuli significantly modified the second component of the cortical response: the posterior N1 peak was always larger for the 3D than the 2D stimuli but the processing of perspective cues did not delay the cortical response to the visual stimuli. Attention to the stimulus category modified the waveforms only at a later, post-N1 stage. The two approaches used to analyze the cortical activity led to the same conclusion: processing 3D information from perspective cues involves the same cortical areas as matched 2D stimuli, with an increase in activity of the regions contributing to the N1 component, and a preferentially greater increase in right parieto-occipital areas. Perspective cues comparable to those used here may initiate vergence eye movements (Enright 1987) and an N1 component related to vergence eye movements has been recently reported (Tzelepi et al. 2004). However, the possibility of the involvement of oculomotor cues in our results appears unlikely. In the Tzelepi et al. (2004) study, the amplitude of vergence movement that evoked a comparable N1 amplitude was 13. Monocular exploration of a drawing with perspective cues seen at comparable viewing distance to the present study yielded much smaller (0.5–1.5) vergence movements. Furthermore, this size of vergence movements disappear in binocular vision (Enright 1987), and we used binocular viewing for the present study. As expected, the early stages of cortical processing of the visual stimuli occurred in posterior areas, as shown Table 1 Grand average and individual modeling x y z Number of subjects P1 (86 ms) N1 (156 ms) 2D 3D 2D 3D 2D 3D 2D 3D 2D 3D 2D 3D 12 12 11 11 10 10 7 9 11 12 10 10 11 10 12 12 9 10 10 11 9 11 7 6 L BA 39 45 67 15 R BA 22 53 60 15 R/L BA 18 3 88 6 R/L BA 7 3 74 43 32 74 43 3 11 64 R BA 7 R/L BA 6 Co-ordinates (mm) in the digitized atlas of Talairach and Tournoux space (1988) of the LORETA peaks of activation processed from grand averages for both the 2D and 3D conditions and number of individual subjects showing activity in the corresponding area, at P1 and N1 latency. Approximation of the localization in Brodmann areas is given 462 Fig. 5 Results of the t-test on the individual LORETA modeling. Localization of the voxels that were significantly more active (P<0.0001) for 3D versus 2D stimuli, at 158 ms post-stimulus latency by the sharply focal posterior VEPs compared to the broad waves recorded frontally. The fact that LORETA modeling also localized the sources in the posterior cortex reinforces the hypothesis of 3D processing occurring in visual regions and that there was not a contributing frontal source. The synchronous dispersed frontal activity was likely the positive ends of the N1 dipoles. Source modeling, although requiring caution in its interpretation when based on 32-channel recordings, suggests the following cortical network is engaged in perspective processing. The first positive component, P1, reflects early cortical activation by the stimuli. The voltage topography and the source modeling at this latency suggest occipital, possibly BA 18 or 17, bilateral occipito-temporal and medial to right parieto-occipital localization of this activity. This is compatible with physiological data based on studies with monkeys showing V1, V3 and MST as regions where neurons have the shortest latencies (Nowak and Bullier 1997; Schmolesky et al. 1998; Bullier 2001). Foxe and Simpson (2002) suggested in a study presenting parafoveal stimuli, that multiple visual generators are active in that latency range. Using highdensity mapping, they provided evidence that despite a timing advantage for dorsal over ventral regions, temporal stream areas are active as early as 82 ms. This is consistent with the bilateral source LORETA proposed around area 37 at P1 latency. This first group of active areas, the ‘‘fast brain’’, or the first wave of information through the visual pathway, was not differentially active for 2D and 3D stimuli since no difference, either in P1 amplitude or in amount of source activity, was observed between the two conditions. It was for the second, negative component, that perspective cues produced an effect. The significantly larger amplitude of N1 to 3D than to 2D stimuli is attributed to enhanced activity in occipital, right and left occipito-temporal regions, and particularly in the right parietal region, around BA7. The symmetrical occipitotemporal sources found at N1 and P1 latencies are possibly related to the V3A and/or MT/MST activation found in fMRI studies of human (Backus et al. 2001; Iwami et al. 2002) and monkey stereopsis (Uka and DeAngelis 2003). These studies investigated depth perception based on binocular cues, but 3D vision based on monocular cues, such as perspective, could share the same processing areas. In addition to these bilateral occipito-temporal junction areas, the different analysis methods used in the present study converge on the finding that the parieto-occipital region is where differences between 2D and 3D stimuli are the most marked: the greatest difference in N1 amplitude between 2D and 3D was found over the right parietal electrodes, where, at this latency, LORETA analyses located a cluster of voxels statistically different between 2D and 3D. This emphasizes that the parietal region, particularly in the right hemisphere, plays a major role in the detection of 3D shape, as Taira et al. (2001) found in an fMRI study based on shading. Other fMRI studies, with RDS or with solid stereograms, suggested that stereopsis processes were dorsally located in the parieto-occipital cortex, particularly on the right hemisphere (Nishida et al. 2001; Iwami et al. 2002). In light of the increased amplitude to perspective cues that we found over the right compared to the left parietal electrode, that region likely processes monocular as well as binocular depth cues, as suggested above for the occipito-temporal sources. Tsutsui et al. (2002) reported that among the many neurons of the caudal part of the lateral bank of the intraparietal sulcus selective to 3D surface orientation, most are sensitive to retinal disparity. The same result was recently obtained in neurons of the lower bank of STS (Liu et al. 2004). In summary, the difference between the processing of 2D and 3D stimuli appeared at the N1 component of the VEP response, and not at the earlier P1 stage. Source analyses allowed us to localize the brain areas implicated. A hypothesis is that the first structures activated, lying mainly but not only on the dorsal, magnocellular pathway, send feedback loops to early processing areas to reinforce their activity for 3D stimulus processing. The increased activity at N1 in occipital, occipito-temporal 463 and right parietal regions would reflect this neural network for refining the ongoing 3D sensory construction, based on monocular as well as binocular 3D cues. Acknowledgements The authors are grateful to Renaud Lestringant for his constructive comments and highly valuable contribution to the statistical analysis of LORETA models, to Laurent Le Guyader for helping with LORETA inversion, and to the 12 subjects who participated in the study. This work was supported by grants from CNRS (Centre National de la Recherche Scientifique) and CNES (Centre National d’Etudes Spatiales). References Backus BT, Fleet DJ, Parker AJ, Heeger DJ (2001) Human cortical activity correlates with stereoscopic depth perception. J Neurophysiol 86:2054–2068 Bullier J (2001) Integrated model of visual processing. Brain Res Rev 36:96–107 Clark VP, Hillyard SA (1996) Spatial selective attention affects early extrastriate but not striate components of the visual evoked potential. J Cogn Neurosci 8:387–402 Cumming BG, Parker AJ (2000) Local disparity not perceived depth is signaled by binocular neurons in cortical area V1 of the macaque. J Neurosci 20:4758–4767 Enright JT (1987) Art and the oculomotor system: perspective illustrations evoke vergence changes. Perception 16:731–746 Ferraina S, Pare M, Wurtz RH (2000) Disparity sensitivity of frontal eye field neurons. J Neurophysiol 83:625–629 Foxe JJ, Simpson GV (2002) Flow of activation from V1 to frontal cortex in humans. A framework for defining ‘‘early’’ visual processing. Exp Brain Res 142:139–150 George N, Evans J, Fiori N, Davidoff J, Renault B (1996) Brain events related to normal and moderately scrambled faces. Brain Res Cogn Brain Res 4:65–76 Giard MH, Perrin F, Echallier JF, Thevenet M, Froment JC, Pernier J (1994) Dissociation of temporal and frontal components in the human auditory N1 wave: a scalp current density and dipole model analysis. Electroencephalogr Clin Neurophysiol 92:238–252 Gonzalez F, Perez R (1998) Modulation of cell responses to horizontal disparities by ocular vergence in the visual cortex of the awake Macaca mulatta monkey. Neurosci Lett 245:101–104 Inui T, Tanaka S, Okada T, Nishizawa S, Katayama M, Konishi J (2000) Neural substrates for depth perception of the Necker cube; a functional magnetic resonance imaging study in human subjects. Neurosci Lett 282:145–148 Iwami T, Nishida Y, Hayashi O, Kimura M, Sakai M, Kani K, Ito R, Shiino A, Suzuki M (2002) Common neural processing regions for dynamic and static stereopsis in human parietooccipital cortices. Neurosci Lett 327:29–32 Janssen P, Vogels R, Orban GA (1999) Macaque inferior temporal neurons are selective for disparity-defined three-dimensional shapes. Proc Natl Acad Sci USA 96:8217–8222 Janssen P, Vogels R, Orban GA (2000) Selectivity for 3D shape that reveals distinct areas within macaque inferior temporal cortex. Science 288:2054–2056 Jeffreys DA (1996) Simple methods of identifying the independently generated components of scalp-recorded responses evoked by stationary patterns. Exp Brain Res 111:100–112 Kourtzi Z, Kanwisher N (2000) Cortical regions involved in perceiving object shape. J Neurosci 20:3310–3318 Lehmann D, Julesz B (1978) Lateralized cortical potentials evoked in humans by dynamic random-dot stereograms. Vis Res 18:1265–1271 Liu Y, Vogels R, Orban GA (2004) Convergence of depth from texture and depth from disparity in macaque inferior temporal cortex. J Neurosci 24:3795–3800 Manning ML, Finlay DC, Dewis SA, Dunlop DB (1992) Detection duration thresholds and evoked potential measures of stereosensitivity. Doc Ophthalmol 79:161–175 Moore C, Engel SA (2001) Neural response to perception of volume in the lateral occipital complex. Neuron 29:277–286 Murray SO, Olshausen BA, Woods DL (2003) Processing shape, motion and three-dimensional shape-from-motion in the human cortex. Cereb Cortex 13:508–516 Nishida Y, Hayashi O, Iwami T, Kimura M, Kani K, Ito R, Shiino A, Suzuki M (2001) Stereopsis-processing regions in the human parieto-occipital cortex. Neuroreport 12:2259–2263 Nowak LG, Bullier J (1997) The timing of information transfer in the visual system. In: Rockland KS, Kaas JH, Peters A (eds) Extrastriate visual cortex in primates, vol 12. Plenum Press, NY, pp 205–241 Pantev C, Bertrand O, Eulitz C, Verkindt C, Hampson S, Schuierer G, Elbert T (1995) Specific tonotopic organizations of different areas of the human auditory cortex revealed by simultaneous magnetic and electric recordings. Electroencephalogr Clin Neurophysiol 94:26–40 Pascual-Marqui RD, Michel CM, Lehmann D (1994) Low resolution electromagnetic tomography: a new method for localizing electrical activity in the brain. Int J Psychophysiol 18:49–65 Schmolesky MT, Wang Y, Hanes DP, Thompson KG, Leutgeb S, Schall JD, Leventhal AG (1998) Signal timing across the macaque visual system. J Neurophysiol 79:3272–3278 Shikata E, Hamzei F, Glauche V, Knab R, Dettmers C, Weiller C, Buchel C (2001) Surface orientation discrimination activates caudal and anterior intraparietal sulcus in humans: an eventrelated fMRI study. J Neurophysiol 85:1309–1314 Skrandies W, Jedynak A (1999) Learning to see 3D: psychophysics and brain electrical activity. Neuroreport 10:249–253 Taira M, Nose I, Inoue K, Tsutsui K (2001) Cortical areas related to attention to 3D surface structures based on shading: an fMRI study. Neuroimage 14:959–966 Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human brain: 3-dimensional proportional system: an approach to cerebral imaging. Thieme Medical Publishing, NY Towle VL, Bolanos J, Suarez D, Tan K, Grzeszczuk R, Levin DN, Cakmur R, Frank SA, Spire J-P (1993) The spatial location of EEG electrodes: locating the best-fitting sphere relative to cortical anatomy. Electroencephalogr Clin Neurophysiol 86:1–6 Trotter Y (1995) Cortical representation of visual three-dimensional space. Perception 24:287–298 Tsutsui KI, Jiang M, Yara K, Sakata H, Taira M (2001) Integration of perspective and disparity cues in surface-orientationselective neurons of area CIP. J Neurophysiol 86:2856–2867 Tsutsui KI, Sakata H, Naganuma T, Taira M (2002) Neural correlates for perception of 3D surface orientation from texture gradient. Science 298:409–412 Tzelepi A, Lutz A, Kapoula Z (2004) EEG activity related to preparation and suppression of eye movements in threedimensional space. Exp Brain Res 155:439–449 Uka T, DeAngelis GC (2003) Contribution of middle temporal area to coarse depth discrimination: comparison of neuronal and psychophysical sensitivity. J Neurosci 23:3515–3530 Uka T, Tanaka H, Yoshiyama K, Kato M, Fujita I (2000) Disparity selectivity of neurons in monkey inferior temporal cortex. J Neurophysiol 84:120–132
© Copyright 2026 Paperzz