Edited by:
Reviewed by:
*Correspondence:
†
This article was submitted to Consciousness Research, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
When viewing the Rubin face–vase illusion, our conscious perception spontaneously alternates between the face and the vase; this illusion has been widely used to explore bistable perception. Previous functional magnetic resonance imaging (fMRI) studies have studied the neural mechanisms underlying bistable perception through univariate and multivariate pattern analyses; however, no studies have investigated the issue of category selectivity. Here, we used fMRI to investigate the neural mechanisms underlying the Rubin face–vase illusion by introducing univariate amplitude and multivariate pattern analyses. The results from the amplitude analysis suggested that the activity in the fusiform face area was likely related to the subjective face perception. Furthermore, the pattern analysis results showed that the early visual cortex (EVC) and the face-selective cortex could discriminate the activity patterns of the face and vase perceptions. However, further analysis of the activity patterns showed that only the face-selective cortex contains the face information. These findings indicated that although the EVC and face-selective cortex activities could discriminate the visual information, only the activity and activity pattern in the face-selective areas contained the category information of face perception in the Rubin face–vase illusion.
The human brain addresses the three-dimensional world by constructing visual representations of two-dimensional retinal information (
In bistable perception, perceptual states alter between one object and another object originating from complex brain activities (
The activation of the visual cortex was found closely related to perception of the Rubin illusion (
Activities in high-level areas such as the right superior parietal lobule (rSPL) preceded the perceptual alteration of the Rubin face–vase illusion (
Modulation of the visual cortex from high-level areas also influenced perceptual alternations during the Rubin illusion perception (
Studies of the Rubin illusion have greatly advanced our understanding of the complex neural mechanisms underlying bistable perception through univariate (
In the current study, we investigated human cortical activation related to the Rubin face–vase illusion. Using multivariate pattern analysis, more information presented in the multivariate pattern of responses to visual stimuli could be studied compared to univariate analysis; therefore, we could observe more activated areas than possible using univariate analysis. Thus, we used both amplitude and multivariate pattern analyses in this study. We hypothesized that regions of interest (ROIs) (face-selective regions) could discriminate between the different perceptual states of bistable perception and would contain category information about the content of the Rubin face–vase illusion.
Twenty-two (18 women, 18–25 years of age) healthy graduate students from Southwest University were paid to participate in this experiment. All subjects were right-handed and had normal or corrected-to-normal vision. The participants were screened to confirm healthy development using a self-report questionnaire before undergoing scanning. Participants with a history of psychiatric or neurological disorders, those who had received mental health treatment or those who had taken psychiatric medications were excluded. All participants were graduates or undergraduates at Southwest University. Informed written consent was obtained from each participant. The Brain Imaging Center Institutional Review Board of Southwest China University approved this study. The consent and experiment procedures were performed in accordance with the World Medical Association Code of Ethics (the Declaration of Helsinki).
In this study, all stimuli were back-projected through a DLP video projector (Tokyo, Japan) (refresh rate: 60 Hz; spatial resolution: 1024 × 768) onto a translucent screen placed inside the fMRI scanner bore. The subjects viewed the stimuli through a mirror located above their eyes. The viewing distance was 83 cm. The tasks were completed on 2 separate days: on day 1, the retinotopic visual areas and the localizer (ROIs) were scanned, and on day 2, the ambiguous (Rubin face–vase illusion task) and unambiguous conditions were scanned.
To define the visual cortex boundaries, retinotopic mapping scanning was used, according to a standard phase-encoded method developed by
The fMRI tests also contained a block-design run to localize the ROIs, including the face-selective areas. The subjects viewed images of faces, non-face objects, and texture patterns (scrambled faces), which subtended 6.2 × 6.2° and were presented at the center of the screen. The images appeared at a rate of 2 Hz in blocks of 12 s, interleaved with 12-s blank blocks. Each image was presented for 300 ms, followed by a 200-ms blank interval. Each block type was repeated five times in the run, which lasted 360 s. The subjects performed a one-back task during which they pressed a key to indicate the same images. This scanning enabled us to define all ROIs in the visual cortex from V1 to higher level areas.
In the scanner, the participants were presented the Rubin face–vase illusion. They were instructed to fixate on a small red circle dot in the middle of the Rubin face–vase stimulus and to indicate any perceptual alternations between the face and vase by continuously pressing one of the two buttons as soon as the new perception was perceived. There were three phases in each “Rubin face–vase illusion” run. First, a red circle dot was represented centrally on the screen for 6 s to prepare. Second, the Rubin face–vase illusion was viewed for 276 s, and the subjects had to keep pressing one of the two buttons to indicate their perceptions. Third, for another 6 s, a red circle dot was placed in the middle of the screen until the end of a run. The ambiguous condition comprised six runs in total, and the subjects were instructed to keep their heads still throughout the entire scanning of the runs (
In the unambiguous condition, the stimuli were grayscale photos of faces and vases (6.2 × 6.2°). Examples are shown in
For each participant, high-resolution T1-weighted structural images were acquired. A 3-T Siemens Trio MRI scanner (Siemens Medical, Erlangen, Germany) was used to collect all images. fMRI data were collected using the scanner with a 12-channel phase-array coil. Blood oxygenation level dependent (BOLD) signals were measured with an EPI sequence (TR: 2000 ms; TE: 30 ms; FOV: 192 mm × 192 mm; matrix: 64 × 64; flip angle: 90; slice thickness: 3.0 mm; gap: 0 mm; number of slices: 33; slice orientation: axial). The bottom slice was positioned at the bottom of the temporal lobes. A three-dimensional MPRAGE structural data set (resolution: 1 mm × 1 mm × 1 mm; TR: 2000 ms; TE: 3.02 ms; FOV: 256 mm × 224 mm; flip angle: 8; number of slices: 176; slice orientation: sagittal) was collected before the task-related functional runs.
The anatomical volume for each subject was first aligned according to the AC-PC locations and then transformed into Talairach coordinates. Functional volumes in all ambiguous, unambiguous, localizer, and retinotopic visual areas scanning were preprocessed using BrainVoyager QX, including three-dimensional motion correction, linear trend removal, and high-pass (0.015 Hz) filtering. Head motion within any fMRI test was less than 3 mm for all subjects. The functional volumes were then aligned to the anatomical volume. The first 6 s of BOLD signals was discarded to minimize any transient magnetic saturation effects.
A general linear model (GLM) procedure was applied to the retinotopic visual area scanning and localizer runs, and seven ROIs were defined from the early to late visual cortices. Face-selective regions were defined as regions that responded more strongly to faces than to non-face objects (
For the unambiguous condition, data were first extracted from each of the ROIs. Then, the activity amplitude in each ROI was used as the estimated beta value through a GLM procedure.
For the ambiguous condition, the event-related averaging method was used. Event-related BOLD signals were calculated separately for each ROI for each subject and condition, following the method used by
The multivariate pattern analysis introduced in the analysis of the ambiguous and unambiguous conditions was a standard correlation analysis of spatial activity pattern, as used by
There was substantial variability in the mean alternation frequency for each subject. The frequency histogram showed the number of participants in the different reported perceptual alternations (
We measured the BOLD signals in response to the face and vase perception in six ambiguous fMRI runs and three unambiguous fMRI runs. As expected, in the unambiguous condition, the differences (face–vase) between the face and vase activity amplitudes were significant in the bilateral OFA (lOFA:
We then investigated the correlation between neural activities and face perceptual duration in all seven ROIs (ambiguous condition). We found the activity amplitudes to face perception of the lOFA (
According to the results, the higher activity amplitude of the face-selective areas might imply that a longer face perception duration leads to lower alternation frequency. However, this correlation might be inconclusive. The correlation might be an artifact of the experimental design because the time of perceptual duration
We first calculated the correlations between the spatial patterns of activities corresponding to ambiguous face and vase perception. Then, the discrimination index was defined as the difference between the correlation coefficients calculated from the same category (face vs. face or vase vs. vase) and from different categories (face vs. vase). A significantly positive index demonstrated that a specific brain region could discriminate the different perceptual states of a subject. As shown in
In the current study, we systematically investigated the neural mechanisms underlying the Rubin face–vase illusion. During unambiguous face perception, the face-selective areas, including the bilateral OFA, the FFA, and the STS, were significantly activated. However, the results of the ambiguous condition indicated that only the bilateral FFA responded significantly to face perception when viewing the Rubin face–vase illusion.
The amplitude analysis results revealed that only the activities of the bilateral FFA were significantly higher during face perception than during vase perception in the ambiguous condition, which was consistent with previous findings (
The OFA and STS results might be explained because they selectively responded to some specific attributes of the face that could be absent in the bistable stimulus. For example, the STS was believed to encode the changeable aspects of the face, such as facial expressions and gaze (
Further analysis of the activity patterns showed that not only the face-selective areas (the bilateral FFA, bilateral STS, and right OFA) but also the EVC could discriminate face perception from vase perception in the ambiguous condition. However, only the face perception activity patterns of the face-selective regions (i.e., the right OFA, bilateral FFA, and bilateral STS) showed a greater similarity to the activity patterns induced by the real face than those induced by the real vase. Although the EVC could discriminate different perceptual states of bistable perception, it did not contain any information about the content of the Rubin face–vase illusion.
The results from the discrimination index showed that we could discriminate the two perceptual states based on the activity patterns from both the EVC and the face-selective cortex. This pattern correlation method is similar to the classification method used by
As we expected, the face-selective regions not only showed discrimination of face and vase perception but also contained face information when the subjects perceived the face. Together, the results of the activity amplitude analysis and activity pattern analysis indicate that both the EVC and high-level face-selective cortex were regulated by conscious awareness. However, only the face-selective cortex was related to the content of the face in the subjective experience. These results strongly support the role of face-selective areas in the conscious processing of faces.
TB, NS, and LH responsible for the experiment design, data collecting, and data analyzing and JQ, YZ, and XW responsible for the manuscript preparation.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This research was supported by the National Natural Science Foundation of China (31271087, 31470981, 31571137, 31500885, 31400960), National Outstanding Young People Plan, the Program for the Top Young Talents by Chongqing, the Fundamental Research Funds for the Central Universities (SWU1509383, SWU1509451), Natural Science Foundation of Chongqing (cstc2015jcyjA10106), Fok Ying Tung Education Foundation (151023), General Financial Grant from the China Postdoctoral Science Foundation (2015M572423, 2015M580767), Special Funds from the Chongqing Postdoctoral Science Foundation (Xm2015037), and Key Research for Humanities and Social Sciences of Ministry of Education (14JJD880009).