| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Graduate School of Letters, Kyoto University, Kyoto, Japan; 2Department of Psychology, Royal Holloway, University of London, London, United Kingdom; and 3Cognitive Science Laboratory, The University of Trento, Trento, Italy
Submitted 16 April 2006; accepted in final form 27 November 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
), such as boundaries defined by contrast differences or local speed differences, as well as motion of first-order signals (boundaries defined by luminance gradients; for a review, see Smith 1994b). Pure second-order motion is a laboratory phenomenon, requiring careful removal of first-order signals to yield an unnatural stimulus, yet we can perceive such motion very readily. In real-life situations, where both types of cue are typically present and are correlated, second-order signals can help to characterize a moving target in a noisy environment with higher accuracy than if first-order signals alone were used (Smith and Scott-Samuel 2001).
Computational studies have depicted three different views of how second-order motion is detected. First, a separate motion pathway was proposed that has nonlinear preprocessing before standard motion computation (Chubb and Sperling 1988). Second, a single pathway of nonlinear detectors was previously shown to be capable of processing both types of motion (Benton and Johnston 2001; Johnston et al. 1992). Such a model can extract velocity information directly from the space–time images (Marr and Ullman 1981). A third possibility is that a feature-tracking mechanism handles second-order motion (as well as first-order motion in some circumstances), whereas first-order motion is normally detected by dedicated low-level motion sensors (Del Viva and Morrone 1998; Derrington et al. 2004). All models can explain most of the data and, despite nearly two decades of debate, the controversy is unresolved. A parallel debate has more recently developed concerning first- and second-order spatial vision. The two debates are separate: even if the existence of a mechanism for static second-order pattern discrimination is established (e.g., Landy and Oruc 2002), mechanisms for second-order motion detection must still be sought.
Human psychophysics (e.g., Derrington and Badcock 1985; Ledgeway and Smith 1994; Nishida et al. 1997) has tended to indicate separate mechanisms for first-order and second-order motion, but this is not always the case (Hock and Gilroy 2005; Taub et al. 1997). Neuropsychological studies on brain damaged patients have favored separate-pathway models by showing selective deficits for second-order motion (Vaina and Cowey 1996) or first-order motion (Vaina et al. 2000), or differential degradation of detection threshold and speed perception for the two types of stimuli (Greenlee and Smith 1997). Earlier electrophysiology studies showed that the same cells respond to both types of motion in monkey V1/MT (O'Keefe and Movshon 1998) and cat area 17/18 (Mareschal and Baker 1998), which apparently is consistent with single-path models, although their responses are somewhat different, lending some support to dual-path models (see DISCUSSION).
Human functional magnetic resonance imaging (fMRI) studies are expected to provide more information on this issue not only because of their ability to see multiple cortical areas at the same time but also because they allow more direct comparison to results from human psychophysics; however, the results published so far are not conclusive. The first such fMRI study (Smith et al. 1998) suggested separate mechanisms and tentatively identified V3/VP as the earliest cortical site at which second-order motion is encoded. More recently, whereas one fMRI study implicated separate cortical networks (Dumoulin et al. 2003), others showed that both first-order and second-order motion are represented as early as V1 (Nishida et al. 2003; Seiffert et al. 2003), suggesting either a common mechanism or separate mechanisms that both originate in V1.
One reason for the inconsistency among fMRI studies is the relatively coarse spatial resolution of fMRI measurement. Previously, two processes occurring in the same voxel could not be distinguished and so similar activations from two different stimuli would not necessarily indicate a single underlying mechanism. However, fMRI adaptation now allows dissociation of different functional properties of neural populations within voxels (Grill-Spector and Malach 2001). Repeated presentation of similar stimuli causes a reduction of the blood oxygen level–dependent (BOLD) response in cortical regions consisting of neurons that are insensitive to the changes between them, whereas little or no reduction occurs if the two stimuli activate different sets of neurons. We can therefore infer the stimulus selectivity of neurons with a resolution beyond that of the fMRI voxels. Nishida et al. (2003) used adaptation between stimulus blocks and found direction-selective adaptation to both first-order and second-order motion, although they did not test cross-adaptation between the two types and so did not address the question of separate, parallel mechanisms.
fMRI adaptation can be observed both in rapid event-related sequences and in longer blocks (Grill-Spector et al. 2006). A short stimulus of a few seconds can yield selective reduction of responses for basic visual features such as orientation (Boynton and Finney 2003; Fang et al. 2005) or motion direction (Huettel et al. 2004). This approach allows more flexible testing of stimulus specificity within a limited scan time. Larsson et al. (2006) used this technique to assess the mechanisms underlying first-order and second-order orientation discrimination and suggested two separate mechanisms by showing different distributions of adaptation across visual areas and a lack of clear cross adaptation. Note that this result does not necessarily imply separate mechanisms underlying motion detection, as discussed above and in their introduction.
Here we report the results of a rapid fMRI adaptation experiment on first-order and second-order motion. Each type of model predicts a different pattern of results. Separate adaptation, up to a point where the signals are integrated, would be found if there are dual pathways. Similar intra- and cross-adaptation would be evidence for a single mechanism. Partial cross-adaptation is expected if second-order motion detection relies on feature tracking because the feature-tracking mechanism can also be adapted by first-order motion, whereas low-level motion sensors would be adapted only by first-order motion. Our results showed direction-selective adaptation effects in MT+—the "motion area"—for both first-order and second-order stimuli but no cross-adaptation between them, consistent with the dual-processing model.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Two of the authors (HA, MW) and three naïve volunteers (JM, PM, WE) participated (aged 20–37, two males and three females). All had normal vision. The experiments were conducted in accordance with the Declaration of Helsinki and were approved by the ethical committee of Royal Holloway, University of London (RHUL). All the participants were screened using a standard procedure before going into the MR scanner. Written informed consent was obtained before each scanning session. One of the authors (HA) and another naïve volunteer (aged 27–38, both males) participated in the psychophysical experiment. Naïve participants were paid for their time according to standard rates and procedures.
Data acquisition
Magnetic resonance images were acquired by using a Siemens Trio scanner (3T) in RHUL, with an eight-channel phased-array head coil. T1-weighted anatomical images were acquired at the beginning of each session (MP-RAGE, Siemens, 1 x 1 x 1-mm voxels). Functional images were acquired with T2*-weighted gradient-recalled echo-planar imaging sequences. The voxel size was 3.0 x 3.0 x 3.0 mm. Fourteen oblique slices were acquired from the occipital lobe with TR = 1 s (TE = 31 ms, flip angle = 90°, field of view = 192 x 192 mm).
Visual stimulation
Visual stimuli were back-projected onto a screen by a liquid-crystal projector. These were viewed binocularly through a mirror above the head coil. The screen was visible as a circular aperture of 30° diameter. The luminance profile was linearized, with a mean luminance of about 2,000 cd/m2.
The stimulus formed an annulus that subtended from 1° (inner diameter) to 6° (outer diameter) visual angle, which was filled with random dots of 50% contrast (Fig. 1). Each dot consisted of a single screen pixel to minimize first-order artifacts (Smith and Ledgeway 1997). The outer area was filled with mean gray. A small disk (diameter 
0.4°) was presented in the center for fixation.
|
). For the first-order stimuli, this was created by adding luminance modulation (LM) to the static noise. For the second-order stimuli, the contrast of the noise was modulated by the sinusoid [contrast modulation (CM)]. In addition to standard gamma correction, the modulation depth of pixels that were darker than the average was adjusted to remove first-order artifacts in the contrast-modulated stimuli (Ashida et al. 2001). The effective modulation contrast was roughly matched between the two motion types (LM: 20% for adaptation and 10% for test, CM: 100% for adaptation and 50% for test). The noise was present throughout the run, with only the modulation appearing and disappearing during events. The speed of motion was always 3.33 Hz (33.3 rpm). Procedure
A trial began with a moving grating as an adaptation stimulus, S1 (2 s), followed by an interstimulus interval with no modulation (2 s) and then the test stimulus, S2 (1 s). There was then a 3-s interval with no modulation before the next trial.
There were five test conditions: LM-same, LM-opposite, CM-same, CM-opposite, and control. Same/different refers to direction of motion and the control contained S1 only. These five conditions were repeated five times in a single run with the order balanced using m-sequences (Buracas and Boynton 2002). This ensured that any contamination of the averaged time courses by the response to the preceding trial equally affected all event types. Events occurred every 8 s. A run consisted of 26 events, with the first one excluded from the analyses. There were blank periods of 6 s in the beginning and 11 s at the end of each run.
As a precaution in case adaptation effects carried over from one event to the next, the motion type (LM/CM) and direction [clockwise/counterclockwise (CW/CCW)] of S1 were fixed within a given scan run. Eight runs were conducted for each type of S1 across two sessions, four in CW and four in CCW directions, in four alternating blocks of two runs of the same direction (S1 directions were collapsed in the data analyses). There were thus 40 events of each condition per participant. A short rest was given between runs.
A central disk served as a fixation mark, the color of which changed randomly at a rate of 1 Hz. To promote fixation and maintain a constant attentional state, participants were asked to count the number of blue disks and report it after each run.
Regions of interest
Moving LM stimuli of the same type (Fig. 1A) were used to identify the MT+ region of interest (ROI) in separate block-design runs (motion vs. blank, 15-s blocks). During the motion blocks, the modulation contrast was 20% and the rate of motion was 3.33 Hz (33.3 rpm). The motion direction was reversed every 2.5 s. During the blank blocks, the fixation disk was presented with the unmodulated noise ring. General linear model (GLM) contrast was computed between motion and blank blocks for each participant. The ROI corresponding to MT+ was identified on the basis of both functional activity and its known approximate anatomical location (Dumoulin et al. 2000).
For four of the five participants, areas MT, MST, and V3A were identified for auxiliary analyses. MST was identified on the basis of ipsilateral activation to a circular patch of expanding random dots, presented in one hemifield only (Dukelow et al. 2001; Huk et al. 2002). This method is described elsewhere (Smith et al. 2006). MT was defined by eliminating overlapping MST voxels from MT+. A rotating wedge stimulus (24° angle), used to identify retinotopic areas (Sereno et al. 1995), consisted of a flickering (8 Hz) red-green checkerboard pattern that was scaled with eccentricity 
12°. It rotated clockwise at the rate of 64 s/cycle, and eight cycles were presented. V3A was identified manually on the cortical flatmaps based on the field sign of the polar-phase maps and the characteristic presence of an upper quadrant representation. Whereas MT+, MT, and MST were defined on individual anatomical spaces, V3A was defined in Talairach space that was required for cortical flattening.
Data analysis
The images were analyzed using Brain Voyager software (Brain Innovation, Maastricht, The Netherlands). The functional images were corrected for slice timing and head movement and temporally high-pass filtered at 0.014 Hz. They were then coregistered to the standard anatomical image for each participant.
Event-related averages of time courses were computed within each ROI for each hemisphere. Responses during the period 0–4 s from the onset of S1 were averaged for each run and were used as baselines to compute percentage signal change. The time courses were averaged across runs within each participant and then across participants, for each type of S1–S2 pair, with the absolute motion directions of S1 collapsed.
An adaptation index for each adapt-test condition was computed on the basis of the averages, across participants, of percentage signal changes collapsed between t = 6 s and t = 12 s, corresponding to the period when most of the response relates to S2. The average of the baseline condition was subtracted from that of each condition to give Aopp and Asame for each stimulus pair (LM–LM, LM–CM, CM–LM, CM–CM). Then the adaptation index AI was given as (Aopp – Asame)/(Aopp + Asame) for each pair. The 95% confidence intervals were computed by using a nonparametric bootstrap procedure (Efron and Tibshirani 1993). Adaptation indices were computed 10,000 times from random sampling of hemispheres with replacement, with the size of the data set being the same as the original data. The 2.5 and 97.5% quantiles were taken as the lower and upper bounds of the confidence intervals.
Psychophysics
In fMRI, as in psychophysics, adaptation may also result in improved performance. An example is facilitation in the form of priming (Maccotta and Buckner 2004). We thus conducted an auxiliary psychophysical experiment to consider the possible perceptual manifestation of the effect we found, by measuring the changes of detection threshold in each S1–S2 condition.
The stimulus configuration was the same as in the fMRI experiment, but to create low-contrast modulation (especially for LM stimuli), we used a VSG 2/5 framebuffer (Cambridge Research Systems) that has 14-bit grayscale resolution and a 17-in. CRT monitor (NANAO F57T, 1,024 x 768 pixels, 60-Hz refresh, mean luminance of 40 cd/m2), which was viewed from a distance of 44 cm with the aid of a head-and-chin rest. After the S1 stimulus of 2 s and a blank (with background noise) of 2 s, a test stimulus (S2) was presented for 1 s. The participants judged whether they saw the second moving grating and responded by using a keypad.
The type and direction of S1 were fixed in a single run, as in the main experiment, and one type of S2 was tested in a single run. Modulation contrast was adjusted by a 1-up–1-down double-random staircase procedure, in which two staircases for both motion directions were intermixed within a single run. Two runs were conducted for each S1–S2 pair. By collapsing the motion direction of S1, four measures of thresholds were obtained for each condition.
Adaptation indices were computed in the same way as described above for the fMRI data, on the basis of averaged sensitivities (i.e., reciprocals of thresholds) across participants as Asame and Aopp. The 95% confidence intervals were computed by using a nonparametric bootstrap procedure, as described earlier for the fMRI data, in which the runs were randomly sampled with replacement.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Figure 2 shows averaged time courses of the BOLD signals in MT+, with separate panels for adaptation to LM and CM. Averaged responses between t = 6 s and t = 12 s (after the offset of the S2 stimuli) are shown to the right. As is evident from these plots, direction-selective adaptation occurred in MT+ when the stimulus type (LM/CM) was the same for S1 and S2. That is, responses were reduced, reflecting common neural processing, when S2 moved in the same direction as S1. This is consistent with neural adaptation of direction-sensitive neurons within each of two separate detection mechanisms. When S1 was LM and S2 was CM (Fig. 2, left, red and orange lines) and when S1 was CM and S2 was LM (Fig. 2, right, green and turquoise lines), there was no direction-selective adaptation. The adaptation indices in Fig. 3 confirm that reliable adaptation occurred only when the stimulus type was the same for S1 and S2. Cross-adaptation between LM and CM motion was not found after adapting to either LM or CM motion.
|
|
). However, we believe that such an effect, if any, was not crucial for the following reasons. First, the demanding task at fixation should have minimized attention to the stimuli. Second, the modulation contrast of S2 was halved from that of S1 and the two stimuli never looked the same. It is unlikely that the participant's attention was captured only when the stimulus type changed. Third, attention alone cannot explain the whole pattern of results. Attention to direction change could lead to larger responses for opposite motion regardless of the stimulus type, but this was not the case. Attention to a change in stimulus type could lead to larger responses for cross-adaptation conditions regardless of direction, but neither was this consistently the case. Simultaneous attention to both variables predicts that the largest response would occur for the condition with stimulus change and direction change, but these responses were no larger than two of the remaining three conditions. It is possible that attention to stimulus change could obscure the effect of direction change if there were saturation of the response, but the use of a low S2 contrast makes saturation unlikely. There was a slight hint of adaptation in the LM probe stimuli after adapting to CM stimuli. Although partial cross-adaptation from CM to LM is one possible interpretation, it must be remembered that complete elimination of first-order artifacts is not easy even after careful calibration. It is therefore safer (and parsimonious) not to take this small effect as genuine unless confirmed by more thorough investigation. Minor contamination could be crucial at the threshold level but is not necessarily so in our suprathreshold experiment. Note that a first-order artifact would also be expected to result in apparent cross-adaptation in the opposite direction (from LM to CM); this was not the case.
In four of the five participants, two subcomponents of MT+, MST and MT, and area V3A were separately analyzed (Figs. 4, 5, and 6, respectively). The results showed very similar trends in all these areas.
|
|
|
), which we expected to ease saturation, but shorter intervals may be optimal for V1. Finally, our preliminary observations suggested that the short TR of 1 s could have reduced the signal-to-noise ratio. Because Nishida et al. (2003) previously reported that direction selectivity to second-order motion is found as early as V1, it is likely that individual adaptation to LM and CM motion occurs in V1. There was a hint of such individual adaptation in V1 in some participants, but it was not reliable. Psychophysics
Figure 7A shows the results for one of the authors (HA) and a paid naïve volunteer (NH). For both observers, an increase of threshold in the "same" direction is seen for LM–LM and CM–CM but not in the LM–CM and CM–LM conditions. Figure 7B shows the adaptation indices. This plot resembles the adaptation index plots in Figs. 3–6, showing clear intraorder adaptation but no cross-adaptation. Adaptation in the CM–CM condition is weaker than that in the fMRI data, whereas adaptation in the LM–LM condition is comparable, but this is largely explained by the naïve observer who did not participate in the fMRI experiment. These results are in line with those obtained in the more thorough investigation of Nishida et al. (1997). In sum, psychophysical results are similar to the fMRI results and it is highly likely that the fMRI adaptation we observed is correlated with perceptual adaptation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
We found order-specific adaptation of BOLD responses to first-order (LM) and second-order (CM) motion. This pattern of results parallels the psychophysical results of adaptation. It is also consistent with order-specific disorders of motion perception found in some patients (Greenlee and Smith 1997; Vaina and Cowey 1996; Vaina et al. 2000). Taken together, our findings strongly suggest the existence of separate neural mechanisms that are responsible for encoding the two types of motion, even if these mechanisms are intermixed in the same brain areas (Nishida et al. 2003; Seiffert et al. 2003). Our result therefore favors dual-pathway models that assume a separate second-order motion pathway rather than single-mechanism models (Benton et al. 2001; Johnston et al. 1992). Such dual-path models often assume nonlinear preprocessing followed by standard motion detection (Chubb and Sperling 1988; Lu and Sperling 1995; Nishida and Ashida 2000; Wilson et al. 1992), although our result does not directly support any specific scheme of motion sensors. It also remains an open question whether the initial-stage mechanism of the second-order motion pathway is shared by the second-order pattern analysis mechanisms shown by Larsson et al. (2006).
Second-order motion can be detected by a higher-order feature-tracking mechanism (Del Viva and Morrone 1998; Derrington et al. 2004; Scott-Samuel and Georgeson 1999; Seiffert and Cavanagh 1998; Smith 1994a). Smith (1994a) presented evidence for two methods of detecting second-order motion: a dedicated mechanism based on second-order motion energy and a feature-tracking method probably shared by first-order stimuli. In contrast, by examining many psychophysical results in the literature, Derrington et al. (2004) concluded that there is no reason to assume dedicated second-order sensors but that a feature-tracking mechanism suffices. Portions of our results may be explained if second-order motion is detected by a higher-order feature-tracking mechanism, whereas first-order motion is detected by lower-level motion sensors. However, a feature-tracking mechanism requires greater attentional resources (Ashida et al. 2001; Derrington et al. 2004) and should have been highly suppressed by the fixation task. Adaptation to CM stimuli therefore should have been much smaller than actually observed. Another point is that feature tracking does not operate exclusively for CM stimuli but for both LM and CM stimuli (Del Viva and Morrone 1998; Smith and Ledgeway 2001). Therefore if CM motion is detected entirely by feature tracking whereas LM motion is mainly detected by lower-level filters, partial cross-adaptation from LM to CM motion is expected even if it does not occur in the opposite direction; however, we did not find cross-adaptation in either direction. Third, if second-order motion is detected by feature tracking, it is necessary to conclude that this process occurs in V1 because direction-specific adaptation for second-order motion occurs in V1 (Nishida et al. 2003). This seems unlikely, given the high-level nature of the process. Our finding indicates that dedicated second-order detectors do exist at least for relatively high modulation contrast. However, we also do not exclude the possibility that a feature-tracking mechanism works cooperatively (Smith and Ledgeway 2001) or exclusively at lower contrasts.
Relation to single-unit physiology
O'Keefe and Movshon (1998) reported that a subset of cells in macaque MT respond to both first-order motion and second-order motion. They also found similar cells in V1, although the sample was much smaller. They did not confirm cue-invariant processing (i.e., invariant responses to first-order and second-order motion) in most MT cells, but suggested that some cells may be cue invariant. They also found that preferred spatial frequency was lower for second-order motion than for first-order motion, whereas direction tuning was similar for both types of stimuli. Their finding is generally consistent with our finding of independent adaptation in human MT+, although the exact site of fMRI adaptation is unknown, as we discuss in the next subsection.
Mareschal and Baker (1998, 1999) examined cells in cat area 18 and argued in support of the dual-pathway model by showing significant differences in the spatiotemporal properties of first-order and second-order stimuli to which the cells respond. Ledgeway et al. (2005) also examined cells in cat area 18 and found different contrast response functions for first-order and second-order motion stimuli. Their conclusions too are generally consistent with ours.
The fact that these single-cell studies report that cells responsive to second-order stimuli also respond to first-order stimuli is a potential problem for the dual-pathway interpretation. There appear to be no reports of cells that respond to second-order motion exclusively, whereas our results suggest the existence of such exclusively second-order cells. We can explain this apparent discrepancy by three points. First, cells that code second-order motion may be sensitive to the carrier structure in a direction-selective way, because spatiotemporal properties are different for first-order and second-order stimuli (Mareschal and Baker 1998). Second, as discussed by Larsson et al. (2006), second-order–only cells may have been overlooked by the procedure of isolating cells using first-order stimuli (Baker and Mareschal 2001). Finally, Sawamura et al. (2006) suggested, based on single-unit recording of macaque inferior temporal cortex, that adaptation might be more selective to stimulus properties than the response of cells. In other words, adaptation measured either by single-unit recording or by BOLD fMRI might underestimate the range of stimuli to which the neurons are responsive, suggesting that stimulus-specific fMRI adaptation might be observed even when the underlying mechanisms involve neurons that respond to both types of stimuli, albeit with different response profiles. If so, our results may not imply completely separate neural populations but partially intermixed ones. This would explain why some neurons respond to both types of motion stimuli even if there are distinct neural channels. Note, however, that there has been no evidence that the finding of Sawamura et al. is directly applicable to our case of adaptation in MT+ because the repetition suppression in their study, measured in the inferotemporal cortex with fast repetitions of many kinds of pictures, is likely correlated with priming (facilitation) rather than conventional adaptation (Maccotta and Buckner 2004).
Possible site of adaptation
Our result of separate direction-selective adaptation to first- and second-order motion in MT+ appears to be consistent with the incomplete cue invariance found by O'Keefe and Movshon (1998) as described above. fMRI adaptation found in area MT+, however, does not necessarily imply that adaptation occurs within MT+ because adapted input signals from the previous processing levels could lead to a similar result (Grill-Spector et al. 2006; Larsson et al. 2006). Actually, a single-unit study of macaques suggested that adaptation to visual motion in MT probably reflects gain control in earlier sites because adaptation was specific to the spatial location despite the larger receptive fields of MT cells (Kohn and Movshon 2003). Our result of similar trends in MT+ and V3A is consistent with this idea. Absence of cross-adaptation in MT+ therefore does not necessarily contradict a popular idea that separate detection is followed by integration of the two streams at the level of MT/MST (Nishida and Ashida 2000; Wilson et al. 1992). We could not identify the origin of adaptation because of unstable results in the early retinotopic areas including V1 with our experimental design. Further investigation would require optimized stimulation and scanning parameters for such areas.
In conclusion, initial separate detection of first-order and second-order motion is suggested by our finding of direction-selective BOLD adaptation to visual motion that is specific to first-order and second-order motion. Two questions remain unanswered: Where does direction-selective adaptation originate? and how/where are motion signals integrated across stimulus types? These questions highlight methodological challenges of fMRI adaptation in terms of asking about the hierarchical structure of the system, given that upstream adaptation affects later areas and adaptability might vary across areas.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: H. Ashida, Graduate School of Letters, Kyoto University, Kyoto 6068501, Japan (E-mail: ashida{at}bun.kyoto-u.ac.jp)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Baker CL Jr, Mareschal I. Processing of second-order stimuli in the visual cortex. Prog Brain Res 134: 171–191, 2001.[ISI][Medline]
Benton CP, Johnston A. A new approach to analysing texture-defined motion. Proc R Soc Lond B Biol Sci 268: 2435–2443, 2001.[Medline]
Benton CP, Johnston A, McOwan PW, Victor JD. Computational modeling of non-Fourier motion: further evidence for a single luminance-based mechanism. J Opt Soc Am A 18: 2204–2208, 2001.
Boynton GM, Finney EM. Orientation-specific adaptation in human visual cortex. J Neurosci 23: 8781–8787, 2003.
Buracas GT, Boynton GM. Efficient design of event-related fMRI experiments using M-sequences. Neuroimage 16: 801–813, 2002.[CrossRef][ISI][Medline]
Cavanagh P, Mather G. Motion: the long and short of it. Spat Vis 4: 103–129, 1989.[Medline]
Chubb C, Sperling G. Drift-balanced random stimuli: a general basis for studying non-Fourier motion perception. J Opt Soc Am A 5: 1986–2007, 1988.[ISI][Medline]
Del Viva MM, Morrone MC. Motion analysis by feature tracking. Vision Res 38: 3633–3653, 1998.[CrossRef][ISI][Medline]
Derrington AM, Allen HA, Delicato LS. Visual mechanisms of motion analysis and motion perception. Annu Rev Psychol 55: 181–205, 2004.[CrossRef][ISI][Medline]
Derrington AM, Badcock DR. Separate detectors for simple and complex grating patterns? Vision Res 25: 1869–1878, 1985.[CrossRef][ISI][Medline]
Dukelow SP, DeSouza JF, Culham JC, van den Berg AV, Menon RS, Vilis T. Distinguishing subregions of the human MT+ complex using visual fields and pursuit eye movements. J Neurophysiol 86: 1991–2000, 2001.
Dumoulin SO, Baker CL Jr, Hess RF, Evans AC. Cortical specialization for processing first- and second-order motion. Cereb Cortex 13: 1375–1385, 2003.
Dumoulin SO, Bittar RG, Kabani NJ, Baker CL Jr, Le Goualher G, Bruce Pike G, Evans AC. A new anatomical landmark for reliable identification of human area V5/MT: a quantitative analysis of sulcal patterning. Cereb Cortex 10: 454–463, 2000.
Efron B, Tibshirani RJ. An Introduction to the Bootstrap. New York: Chapman & Hall, 1993.
Fang F, Murray SO, Kersten D, He S. Orientation-tuned fMRI adaptation in human visual cortex. J Neurophysiol 94: 4188–4195, 2005.
Greenlee MW, Smith AT. Detection and discrimination of first- and second-order motion in patients with unilateral brain damage. J Neurosci 17: 804–818, 1997.
Grill-Spector K, Henson R, Martin A. Repetition and the brain: neural models of stimulus-specific effects. Trends Cogn Sci 10: 14–23, 2006.[CrossRef][ISI][Medline]
Grill-Spector K, Malach R. fMR-adaptation: a tool for studying the functional properties of human cortical neurons. Acta Psychol (Amst) 107: 293–321, 2001.[CrossRef][Medline]
Hock HS, Gilroy LA. A common mechanism for the perception of first-order and second-order apparent motion. Vision Res 45: 661–675, 2005.[CrossRef][ISI][Medline]
Huettel SA, Obembe OO, Song AW, Woldorff MG. The BOLD fMRI refractory effect is specific to stimulus attributes: evidence from a visual motion paradigm. Neuroimage 23: 402–408, 2004.[CrossRef][ISI][Medline]
Huk AC, Dougherty RF, Heeger DJ. Retinotopy and functional subdivision of human areas MT and MST. J Neurosci 22: 7195–7205, 2002.
Huk AC, Ress D, Heeger DJ. Neuronal basis of the motion aftereffect reconsidered. Neuron 32: 161–172, 2001.[CrossRef][ISI][Medline]
Johnston A, McOwan PW, Buxton H. A computational model of the analysis of some first-order and second-order motion patterns by simple and complex cells. Proc R Soc Lond B Biol Sci 250: 297–306, 1992.[Medline]
Kohn A, Movshon JA. Neuronal adaptation to visual motion in area MT of the macaque. Neuron 39: 681–691, 2003.[CrossRef][ISI][Medline]
Landy MS, Oruc I. Properties of second-order spatial frequency channels. Vision Res 42: 2311–2329, 2002.[CrossRef][ISI][Medline]
Larsson J, Landy MS, Heeger DJ. Orientation-selective adaptation to first- and second-order patterns in human visual cortex. J Neurophysiol 95: 862–881, 2006.
Ledgeway T, Smith AT. Evidence for separate motion-detecting mechanisms for first- and second-order motion in human vision. Vision Res 34: 2727–2740, 1994.[CrossRef][ISI][Medline]
Ledgeway T, Zhan C, Johnson AP, Song Y, Baker CL Jr. The direction-selective contrast response of area 18 neurons is different for first- and second-order motion. Vis Neurosci 22: 87–99, 2005.[CrossRef][ISI][Medline]
Lu ZL, Sperling G. The functional architecture of human visual motion perception. Vision Res 35: 2697–2722, 1995.[CrossRef][ISI][Medline]
Macotta L, Buckner RL. Evidence for neural effects of repetition that directly correlate with behavioral priming. J Cogn Neurosci 16: 1625–1632, 2004.
Mareschal I, Baker CL Jr. Temporal and spatial response to second-order stimuli in cat area 18. J Neurophysiol 80: 2811–2823, 1998.
Mareschal I, Baker CL Jr. Cortical processing of second-order motion. Vis Neurosci 16: 527–540, 1999.[CrossRef][ISI][Medline]
Marr D, Ullman S. Directional selectivity and its use in early visual processing. Proc R Soc Lond B Biol Sci 211: 151–180, 1981.[Medline]
Nishida S, Ashida H. A hierarchical structure of motion system revealed by interocular transfer of flicker motion aftereffects. Vision Res 40: 265–278, 2000.[CrossRef][ISI][Medline]
Nishida S, Ledgeway T, Edwards M. Dual multiple-scale processing for motion in the human visual system. Vision Res 37: 2685–2698, 1997.[CrossRef][ISI][Medline]
Nishida S, Sasaki Y, Murakami I, Watanabe T, Tootell RB. Neuroimaging of direction-selective mechanisms for second-order motion. J Neurophysiol 90: 3242–3254, 2003.
O'Keefe LP, Movshon JA. Processing of first- and second-order motion signals by neurons in area MT of the macaque monkey. Vis Neurosci 15: 305–317, 1998.[CrossRef][ISI][Medline]
Sawamura H, Orban GA, Vogels R. Selectivity of neuronal adaptation does not match response selectivity: a single-cell study of the FMRI adaptation paradigm. Neuron 49: 307–318, 2006.[CrossRef][ISI][Medline]
Scott-Samuel NE, Georgeson MA. Feature matching and segmentation in motion perception. Proc R Soc Lond B Biol Sci 266: 2289–2294, 1999.[Medline]
Seiffert AE, Cavanagh P. Position displacement, not velocity, is the cue to motion detection of second-order stimuli. Vision Res 38: 3569–3582, 1998.[CrossRef][ISI][Medline]
Seiffert AE, Somers DC, Dale AM, Tootell RB. Functional MRI studies of human visual motion perception: texture, luminance, attention and after-effects. Cereb Cortex 13: 340–349, 2003.
Sereno MI, Dale AM, Reppas JB, Kwong KK, Belliveau JW, Brady TJ, Rosen BR, Tootell RB. Borders of multiple visual areas in humans revealed by functional magnetic resonance imaging. Science 268: 889–893, 1995.
Smith AT. Correspondence-based and energy-based detection of second-order motion in human vision. J Opt Soc Am A 11: 1940–1948, 1994a.
Smith AT. The detection of second-order motion. In: Visual Detection of Motion, edited by Smith AT and Snowden RJ. London: Academic Press, 1994b, p. 145–176.
Smith AT, Greenlee MW, Singh KD, Kraemer FM, Hennig J. The processing of first- and second-order motion in human visual cortex assessed by functional magnetic resonance imaging (fMRI). J Neurosci 18: 3816–3830, 1998.
Smith AT, Ledgeway T. Separate detection of moving luminance and contrast modulations: fact or artifact? Vision Res 37: 45–62, 1997.[CrossRef][ISI][Medline]
Smith AT, Ledgeway T. Motion detection in human vision: a unifying approach based on energy and features. Proc R Soc Lond B Biol Sci 268: 1889–1899, 2001.[Medline]
Smith AT, Scott-Samuel NE. First-order and second-order signals combine to improve perceptual accuracy. J Opt Soc Am A 18: 2267–2272, 2001.
Smith AT, Wall MB, Williams AL, Singh KD. Sensitivity to optic flow in human cortical areas MT and MST. Eur J Neurosci 23: 561–569, 2006.[CrossRef][ISI][Medline]
Taub E, Victor JD, Conte MM. Nonlinear preprocessing in short-range motion. Vision Res 37: 1459–1477, 1997.[CrossRef][ISI][Medline]
Vaina LM, Cowey A. Impairment of the perception of second order motion but not first order motion in a patient with unilateral focal brain damage. Proc R Soc Lond B Biol Sci 263: 1225–1232, 1996.[Medline]
Vaina LM, Soloviev S, Bienfang DC, Cowey A. A lesion of cortical area V2 selectively impairs the perception of the direction of first-order visual motion. Neuroreport 11: 1039–1044, 2000.[ISI][Medline]
Wilson HR, Ferrera VP, Yo C. A psychophysically motivated model for two-dimensional motion perception. Vis Neurosci 9: 79–97, 1992.[ISI][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
