Neuroscience Letters 562 (2014) 19–23
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Steady-state visual-evoked response to upright and inverted geometrical faces: A magnetoencephalography study Aki Tsuruhara a,∗ , Koji Inui a,b , Ryusuke Kakigi a,b a b
Department of Integrative Physiology, National Institute for Physiological Sciences, Okazaki 444-8585, Japan Department of Physiological Sciences, School of Life Sciences, The Graduate University for Advanced Studies, Hayama, Kanagawa 240-0193, Japan
h i g h l i g h t s • • • • •
The face inversion effect on steady-state visual-evoked magnetic fields was studied. Upright and inverted faces elicited clear SSVEF responses. Face inversion delayed the latency of SSVEFs in the right temporal area. SSVEF amplitudes were not different between upright and inverted faces. Different systems may lie in the delayed and enhanced responses by face inversion.
a r t i c l e
i n f o
Article history: Received 26 August 2013 Received in revised form 19 December 2013 Accepted 2 January 2014 Keywords: Magnetoencephalography Steady-state visual-evoked response Face Inversion effect
a b s t r a c t The face is one of the most important visual stimuli in human life, and inverted faces are known to elicit different brain responses than upright faces. This study analyzed steady-state visual-evoked magnetic fields (SSVEFs) in eleven healthy participants when they viewed upright and inverted geometrical faces presented at 6 Hz. Steady-state visual-evoked responses are useful measurements and have the advantages of robustness and a high signal-to-noise ratio. Spectrum analysis revealed clear responses to both upright and inverted faces at the fundamental stimulation frequency (6 Hz) and harmonics, i.e. SSVEFs. No significant difference was observed in the SSVEF amplitude at 6 Hz between upright and inverted faces, which was different from the transient visual-evoked response, N170. On the other hand, SSVEFs were delayed with the inverted face in the right temporal area, which was similar to N170 and the results of previous steady-state visual-evoked potentials studies. These results suggest that different mechanisms underlie the larger amplitude and delayed latency observed with face inversion, though further studies are needed to fully elucidate these mechanisms. Our study revealed that SSVEFs, which have practical advantages for measurements, could provide novel findings in human face processing. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Face perception has been considered as one of the most important factors of daily life in humans, and previous studies have suggested that processing human faces induces specialized brain responses. Studies using event-related potentials (ERPs) showed that a human face elicited a transient brain response (N170), a posterior negativity peaking approximately 170 ms after face presentation [1–4]. N170 and its corresponding response on magnetoencephalography (MEG), M170, are known to be affected by face inversion. An inverted face elicits delayed [1,5] or delayed and enhanced [6–8] N170, and delayed M170 at least in the right
∗ Corresponding author. Tel.: +81 564 55 7752; fax: +81 564 52 7913. E-mail addresses: [email protected] (A. Tsuruhara), [email protected] (K. Inui), [email protected] (R. Kakigi). 0304-3940/$ – see front matter © 2014 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neulet.2014.01.001
occipito-temporal area [5,9,10]. This effect has been suggested to be face-specific [6], and is referred to as the face inversion effect. However, face inversion does not appear to induce the same effect in the steady-state visual-evoked response, i.e. the response to a repeatedly presented visual stimulus at a specific frequency [11]. Previous studies reported that steady-state visual-evoked potentials (SSVEPs) were affected by face inversion. Rossion et al. [12] presented upright and inverted face stimuli at 4 Hz and found that both upright and inverted faces elicited SSVEPs in the right occipito-temporal areas at a fundamental frequency of 4 Hz, and that the phase of the SSVEP was delayed for inverted faces, similar to the inversion effect for N170. In contrast, face orientation had no effect on the amplitude of the fundamental frequency response when identical faces were presented repeatedly. Similar results were also found in their other study in which the presentation rate was 3.5 Hz [13]. Gruss et al. [14] demonstrated that the larger SSVEP
20
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
Fig. 1. The Face stimulus and stimulation sequence. The luminance of the stimuli (the three elements in the white background circle) was sinusoidally changed at a rate of 6 Hz so that the lower contrast face stimulus in the midline, in between the background and the full face stimulus, appeared at an intermediary stage of stimulation. The background and fixation circles were kept unchanged during each 8.67-s stimulation sequence.
amplitude with face inversion was absent at low presentation rates (5 and 10 Hz), but was present at higher rates (15 and 20 Hz). This study analyzed steady-state visual-evoked magnetic fields (SSVEFs) when upright and inverted geometrical faces were presented. Although previous studies examined the face inversion effect on SSVEP as above, its effects on SSVEFs remain unknown. Measuring steady-state responses was shown to have practical advantages such as a high signal-to-noise ratio and shorter experimental time with frequency analyses [e.g. 12]. SSVEFs, as well as SSVEPs, are known to be useful for exploring visual processing [15,16]. The results of the present study may provide clear evidence that the face inversion effect is similar between SSVEF and SSVEP, and that face processing can be examined by measuring SSVEF. 2. Methods 2.1. Participants Eleven healthy adult volunteers (four females and seven males), aged between 25 and 49 years (mean 35.2; SD 6.4), participated in the present study. All participants had normal or corrected-tonormal visual acuity. This study was approved in advance by the Ethics Committee of authors’ affiliation, and all participants gave informed consent. 2.2. Stimuli Stimuli consisted of a face-like geometrical figure (Face), as shown in Fig. 1, and its inversion (iFace). The face inversion effect on N170 was observed with a simple geometrical face-like figure [8]. Thus, a geometrical figure was presented in this study, to avoid any effect of race, gender, attractiveness, or any other factors that could be found in real photographs. All participants reported that they perceived the upright stimulus as a face, and considered the inverted stimulus to be the inversion of the upright stimulus. The figure was composed of a large white circle (30.8 cm in diameter, the same hereafter), three black circles (5.4 cm), and a small red fixation circle (0.4 cm). Visual stimuli were presented by a personal computer (PC, IBM) and video projector (Mirage 2000; CHRISTIE DIGITAL SYSTEM Inc, Kitchener, Canada) placed outside the magnetically-shielded room, and projected on a screen placed 200 cm from the participant. The refresh rate of the projector was 60 Hz. The projected area was subtended 44.2 cm × 32.9 cm. The luminance of the black circles and
the black background was 0.14 cd/m2 , of the large white circle was 2.82 cd/m2 , and of the fixation circle was 0.47 cd/m2 . The stimulation was given with repeating sessions including two stimulus sequences. In a sequence, one of the two stimuli (Face/iFace) appeared and disappeared 52 times on the screen at a rate of 6 Hz (one face every 166.7 ms). One sequence lasted 8.67 s and followed by another sequence with another type of the stimulus, after a 10-s rest. Epochs with MEG signals larger than 2.7 pT/cm were rejected, and a session, i.e. two sequences, was repeated until four sessions were completed without data rejection for each participant. A 30-s or longer rest was inserted every two sessions. The order of the two sequences (Face/iFace) was randomized in each session and across participants. The stimulation frequency was determined as 6 Hz for the following reasons. First of all, we wanted to avoid the contamination of noise at the alpha range (8–12 Hz) at the fundamental frequency. The projector refresh rate (60 Hz) was also considered because the stimulus was presented as a sinusoidal (rather than abrupt, as in a square wave function; see Fig. 1) function, consistent with in previous SSVEP studies [12,13,17]. In addition, a recent study demonstrated that a frequency rate of 6 Hz gave the largest SSVEP response to faces [17]. The experiment was conducted in a quiet, magneticallyshielded, and darkened room. Participants passively viewed the display and were asked to maintain their gaze at the fixation circle throughout the experiment. 2.3. MEG recordings Magnetic signals were recorded using a 306-channel wholehead type MEG system (Vector-view, ELEKTA Neuromag, Helsinki, Finland), which comprised 102 identical triple sensor elements. Each sensor element consisted of two orthogonal planar gradiometers and one magnetometer coupled to a multi-superconducting quantum interference device (SQUID), which provided three independent measurements of the magnetic fields. Signals were recorded with a band-pass filter of 0.1–200 Hz and digitized at 1004 Hz. To record the timing of the stimulus presentation, a trigger was sent from the parallel port of the stimulation computer to the MEG recording. 2.4. Data analysis We analyzed MEG signals recorded from 204 planar-type gradiometers. These planar gradiometers were powerful enough to detect the largest signal just over local cerebral sources. To examine SSVEFs, we conducted spectrum analysis on MEG signals and computed the amplitude spectrum as follows. First, a Fast Fourier Transform (FFT) algorithm was applied to an 8159ms (8192 sampled data) window in each stimulation sequence, for the MEG waveform of each sensor for sequence. To avoid contamination from the transient responses triggered by the onset of the stimulation sequence, the initial 333-ms (2 cycles) waveforms were not included in the spectrum analysis. The frequency resolution was 0.12 Hz (dividing the sampling rate, 1004 Hz, by the sampled data, 8192). The amplitude and phase of each frequency was then averaged across sequences for Face and iFace. The signal-to-noise ratio (SNR) of the amplitude at each sensor was then computed. The SNR was the ratio of the amplitude at the frequency of interest to the average amplitude of 20 neighboring bins [12,13,18]. The SNR at the fundamental frequency (6 Hz) was extracted for further analyses. To compare the effect of face orientation on SSVEFs across the visual areas, we selected representative sensors in each visual area (bilateral temporal and occipital areas) as follows. First, sensors
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
21
Table 1 SNRs and phase components of SSVEP at the fundamental frequency (6 Hz) in the representative sensors. SNR
Temporal Occipital
Mean
Left Right Left Right
Phase Temporal Occipital
Fig. 2. Grouping of the MEG sensors.
were grouped (Fig. 2) according to the reference manual of the software for the MEG system (Graph, ELEKTA Neuromag). We then selected the representative sensor showing the maximum SNR of the amplitude at 6 Hz in each visual area for Face and iFace, respectively. To examine the effect of stimulus orientation on the amplitude at the stimulation frequency (6 Hz), the SNR at 6 Hz of the representative sensor was compared between Face and iFace. To explore the temporal difference induced in SSVEFs by Face and iFace, the phase at the fundamental frequency (6 Hz) of the representative sensors were also compared between the two types of stimuli. Three-way ANOVAs (orientation: Face/iFace, area: occipital/temporal, hemisphere: right/left) were conducted on the SNR of the amplitude and phase at 6 Hz, with the level of significance at 0.05. 3. Results Spectrum analysis revealed clear peaks in the fundamental frequency (6 Hz) and its harmonics under both Face and iFace conditions (Fig. 3). The representative sensor in each visual area
Left Right Left Right
Face
iFace
3.0 ± 0.3 3.6 ± 0.5 5.3 ± 0.6 5.4 ± 0.5
3.0 ± 0.3 3.1 ± 0.2 4.8 ± 0.3 5.1 ± 0.4
Mean Face
iFace
−1.7 ± 29.6 41.0 ± 20.4 −7.4 ± 32.1 −46.7 ± 25.5
39.7 ± 21.2 −58.7 ± 31.5 13.4 ± 22.0 −24.7 ± 20.9
Note: Data were expressed as mean ± SE from eleven participants. A smaller phase value indicates a delay.
(bilateral occipital and temporal areas) was extracted as described in Section 2. Fig. 4 shows the representative MEG waveforms of a participant that were averaged and filtered by a band-pass filter at the fundamental frequency (6 Hz). Although the peak right after the zero point in Fig. 4 appears to indicate a shorter latency than that of N170, the peak may not indicate a response in such a short latency. The stimulus was presented every 167 ms, and overlapping MEG epochs from 167 ms before to 333 ms after the onset of stimulation were averaged. Therefore, the peak right after the zero point may reflect the response to the previous stimulus. Another possibility is that the high contrast black-and-white stimulus used in this study may have caused a shorter latency. The grand-averaged SNR at 6 Hz in the representative sensors are shown in Fig. 5 and Table 1. A three-way ANOVA (orientation: Face/iFace, area: occipital/temporal, hemisphere: right/left) showed that the main effect of area was significant (F(1,10) = 45.84, p < 0.001). Neither the main effect of orientation nor interactions including orientation were significant. These results indicated that SSVEFs in the occipital areas were significantly larger than those in
Fig. 3. Spectrum analysis of a representative participant. (a) The amplitude spectrum of all gradiometers for Face and iFace. Top-view topography at the fundamental frequency (6 Hz), and its second and third harmonics are shown. (b) The signal-to-noise ratio (SNR) of all gradiometers for Face and iFace, and top-view topography at the fundamental frequency (6 Hz), and its second and third harmonics. This participant exhibited large alpha-band activities, but this was not clear in SNRs.
22
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
Fig. 4. MEG waveforms of a representative participant. The enlarged waveforms were the representative sensors’ that were selected for analyses. The figures attached to each enlarged waveform indicate that the sensor is the representative one for Face, iFace, or both. Overlapping MEG epochs from 167 ms before to 333 ms after the onset of stimulation were averaged for each stimulus. The zero time point of each epoch corresponded to the beginning of the appearance of a stimulus. We used MEG epochs starting 3 cycles after the beginning of the stimulation sequence for that averaging. At least 180 MEG epochs were averaged for each stimulus, and were then filtered with a band-pass filter of 6 Hz (1 Hz width).
the temporal areas regardless of differences in the hemisphere and stimulus orientation. Fig. 5 and Table 1 also show the grand-averaged phase components of SSVEF at 6 Hz in the representative sensors. A three-way ANOVA (orientation: Face/iFace, area: occipital/temporal, hemisphere: right/left) revealed that a one-way interaction between orientation and hemisphere (F(1, 10) = 6.53, p = 0.029) and the twoway interaction across orientation, area, and hemisphere (F(1, 10) = 11.20, p = 0.007) were significant. No other main effect or interaction was significance. Post hoc analyses showed that a simple-simple main effect of orientation was only significant in the right temporal area (F(1, 10) = 12.90, p = 0.005). The average phase difference between Face and iFace in the area was 99.7◦ , which
corresponds to 46 ms at 6 Hz. This simple-simple main effect in the right temporal area may have caused the significant one-way interaction between orientation and hemisphere and the opposite effect of orientation between the two hemispheres, while the main effect of orientation was not significant in either hemisphere. 4. Discussion Both upright and inverted geometrical faces presented at 6 Hz elicited a clear response oscillating at that specific frequency and its harmonics, i.e. SSVEFs. The topography of the fundamental and harmonic responses was basically distributed over the occipital and temporal areas.
Fig. 5. Vector plots of the SNRs and phase components of SSVEP at the fundamental frequency (6 Hz) in the representative sensors. Data were averaged across participants. The radius of the small semicircle indicates SNR = 1, and of the large one does SNR = 5. Phase delays (counter-clockwise lags) were observed under iFace condition, especially in the right temporal area.
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
Our results revealed no significant difference in the amplitude of SSVEFs between the upright and inverted faces, which is inconsistent with the inversion effect of N170 reported in some previous VEP studies [6–8]. A geometrical face-like figure was used as a stimulus in this study, and the lack of reality may have led to the absence of an enlarged response to the inverted stimulus. However, Tomalski and Jonson [8] presented a simple geometrical face-like figure that was similar to ours, and reported an enlarged N170 with inversion. In addition, previous SSVEP studies that presented face photographs did not find a larger amplitude with face inversion [12,13]. Therefore, it is more likely that factors other than the reality of the face stimulus were related to the absence of the face inversion effect on amplitude. One possibility is the stimulus presentation rate. Gruss et al. [14] recently reported that the larger SSVEP amplitude with face inversion was only observed when the presentation rate of the face stimulus was high (15 Hz∼), which is consistent with the present results as well as previous SSVEP studies [7,8]. Another possibility is that the effect of face inversion on amplitude may not necessarily be significant [1], especially with MEG measurement [5,9,10]. In contrast to the absence of an effect on amplitude, the results of the present study showed a phase delay in the right temporal area with face inversion. Delayed response to face inversion were also observed in previous studies regardless of the measured response (N170 [1,5–8], M170 [5,9,10] or SSVEPs [13]), the reality of the face stimulus (a photograph [e.g. 1] or a face-like figure [8]), or the presentation rate to measure steady-state responses. Although further studies are needed to clarify the mechanism relating the larger amplitude and delayed latency with face inversion, our results suggest that different mechanisms underlie the two indices of the face inversion effect. As reported in previous SSVEP and SSVEF studies [e.g. 12], measuring steady-state responses has advantages, such as a relatively short measuring time and highly distinctive response, which allows a focus to be placed on just the amplitude at the chosen frequency, with no ambiguity in selecting the component of interest. In addition, our results suggest that an independent mechanism may affect the delayed latency and increased amplitude with face inversion because the former was found in both steady-state and transient response, whereas the latter was not necessarily. This finding indicates that measuring steady-state responses could contribute to novel findings on face processing in the human brain with many practical advantages. Acknowledgments This study was supported by Grants-in-Aid for Scientific Research on Innovative Areas “Face perception and recognition”
23
from MEXT KAKENHI (20119007, 23119722), and a Grant-in-Aid for Young Scientists (B) (23730709) from JSPS. References [1] S. Bentin, T. Allison, A. Puce, E. Perez, G. McCarthy, Electrophysiological studies of face perception in humans, J. Cogn. Neurosci. 8 (1996) 551–565. [2] M. Eimer, Effects of face inversion on the structural encoding and recognition of faces. Evidence from event-related brain potentials, Brain Res. Cogn. Brain Res. 10 (2000) 145–158. [3] N. George, J. Evans, N. Fiori, J. Davidoff, B. Renault, Brain events related to normal and moderately scrambled faces, Brain Res. Cogn. Brain Res. 4 (1996) 65–76. [4] R.J. Itier, M.J. Taylor, N170 or N1? Spatiotemporal differences between object and face processing using ERPs, Cereb. Cortex. 14 (2004) 132–142. [5] K. Linkenkaer-Hansen, J.M. Palva, M. Sams, J.K. Hietanen, H.J. Aronen, R.J. Ilmoniemi, Face-selective processing in human extrastriate cortex around 120 ms after stimulus onset revealed by magneto- and electroencephalography, Neurosci. Lett. 253 (1998) 147–150. [6] B. Rossion, J.F. Delvenne, D. Debatisse, V. Goffaux, R. Bruyer, M. Crommelinck, J.M. Guérit, Spatio-temporal localization of the face inversion effect: an eventrelated potentials study, Biol. Psychol. 50 (1999) 173–189. [7] B. Rossion, I. Gauthier, M.J. Tarr, P. Despland, R. Bruyer, S. Linotte, M. Crommelinck, The N170 occipito-temporal component is delayed and enhanced to inverted faces but not to inverted objects: an electrophysiological account of face-specic processes in the human brain, Neuroreport 11 (2000) 69–74. [8] P. Tomalski, M.H. Johnson, Cortical sensitivity to contrast polarity and orientation of faces is modulated by temporal-nasal hemifield asymmetry, Brain Imaging Behav. 6 (2012) 88–101. [9] S. Watanabe, R. Kakigi, A. Puce, The spatiotemporal dynamics of the face inversion effect: A magneto- and electro-encephalographic study, Neuroscience 116 (2003) 879–895. [10] J. Liu, M. Higuchi, a Marantz, N. Kanwisher, The selectivity of the occipitotemporal M170 for faces, Neuroreport 11 (2000) 337–341. [11] D. Regan, Some characteristics of average steady-state and transient responses evoked by modulated light, Electroencephalogr. Clin. Neurophysiol. 20 (1966) 238–248. [12] B. Rossion, E.A. Prieto, A. Boremanse, D. Kuefner, G. Van Belle, A steadystate visual evoked potential approach to individual face perception: Effect of inversion, contrast-reversal and temporal dynamics, NeuroImage 63 (2012) 1585–1600. [13] B. Rossion, A. Boremanse, Robust sensitivity to facial identity in the right human occipito-temporal cortex as revealed by steady-state visual-evoked potentials, J. Vis. 11 (2011) 1–21. [14] L.F. Gruss, M.J. Wieser, S.R. Schweinberger, A. Keil, Face-evoked steady-state visual potentials: effects of presentation rate and face inversion, Front. Hum. Neurosci. 6 (2012) 316 1–10. [15] Y. Chen, A.K. Seth, J.A. Gally, G.M. Edelman, The power of human brain magnetoencephalographic signals can be modulated up or down by changes in an attentive visual task., Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 3501–3506. [16] B. Cottereau, J. Lorenceau, A. Gramfort, M. Clerc, B. Thirion, S. Baillet, Phase delays within visual cortex shape the response to steady-state visual stimulation, NeuroImage 54 (2011) 1919–1929. [17] E. Alonso-Prieto, G. Van Belle, J. Liu-Shuang, A.M., Norcia, B. Rossion, The 6 Hz fundamental stimulation frequency rate for individual face discrimination in the right occipito-temporal cortex, Neuropsychologia (2013) 1–13. [18] R. Srinivasan, D.P. Russell, G.M. Edelman, G. Tononi, Increased synchronization of neuromagnetic responses during conscious perception, J. Neurosci. 19 (1999) 5435–5448.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Steady-state visual-evoked response to upright and inverted geometrical faces: A magnetoencephalography study Aki Tsuruhara a,∗ , Koji Inui a,b , Ryusuke Kakigi a,b a b
Department of Integrative Physiology, National Institute for Physiological Sciences, Okazaki 444-8585, Japan Department of Physiological Sciences, School of Life Sciences, The Graduate University for Advanced Studies, Hayama, Kanagawa 240-0193, Japan
h i g h l i g h t s • • • • •
The face inversion effect on steady-state visual-evoked magnetic fields was studied. Upright and inverted faces elicited clear SSVEF responses. Face inversion delayed the latency of SSVEFs in the right temporal area. SSVEF amplitudes were not different between upright and inverted faces. Different systems may lie in the delayed and enhanced responses by face inversion.
a r t i c l e
i n f o
Article history: Received 26 August 2013 Received in revised form 19 December 2013 Accepted 2 January 2014 Keywords: Magnetoencephalography Steady-state visual-evoked response Face Inversion effect
a b s t r a c t The face is one of the most important visual stimuli in human life, and inverted faces are known to elicit different brain responses than upright faces. This study analyzed steady-state visual-evoked magnetic fields (SSVEFs) in eleven healthy participants when they viewed upright and inverted geometrical faces presented at 6 Hz. Steady-state visual-evoked responses are useful measurements and have the advantages of robustness and a high signal-to-noise ratio. Spectrum analysis revealed clear responses to both upright and inverted faces at the fundamental stimulation frequency (6 Hz) and harmonics, i.e. SSVEFs. No significant difference was observed in the SSVEF amplitude at 6 Hz between upright and inverted faces, which was different from the transient visual-evoked response, N170. On the other hand, SSVEFs were delayed with the inverted face in the right temporal area, which was similar to N170 and the results of previous steady-state visual-evoked potentials studies. These results suggest that different mechanisms underlie the larger amplitude and delayed latency observed with face inversion, though further studies are needed to fully elucidate these mechanisms. Our study revealed that SSVEFs, which have practical advantages for measurements, could provide novel findings in human face processing. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Face perception has been considered as one of the most important factors of daily life in humans, and previous studies have suggested that processing human faces induces specialized brain responses. Studies using event-related potentials (ERPs) showed that a human face elicited a transient brain response (N170), a posterior negativity peaking approximately 170 ms after face presentation [1–4]. N170 and its corresponding response on magnetoencephalography (MEG), M170, are known to be affected by face inversion. An inverted face elicits delayed [1,5] or delayed and enhanced [6–8] N170, and delayed M170 at least in the right
∗ Corresponding author. Tel.: +81 564 55 7752; fax: +81 564 52 7913. E-mail addresses: [email protected] (A. Tsuruhara), [email protected] (K. Inui), [email protected] (R. Kakigi). 0304-3940/$ – see front matter © 2014 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.neulet.2014.01.001
occipito-temporal area [5,9,10]. This effect has been suggested to be face-specific [6], and is referred to as the face inversion effect. However, face inversion does not appear to induce the same effect in the steady-state visual-evoked response, i.e. the response to a repeatedly presented visual stimulus at a specific frequency [11]. Previous studies reported that steady-state visual-evoked potentials (SSVEPs) were affected by face inversion. Rossion et al. [12] presented upright and inverted face stimuli at 4 Hz and found that both upright and inverted faces elicited SSVEPs in the right occipito-temporal areas at a fundamental frequency of 4 Hz, and that the phase of the SSVEP was delayed for inverted faces, similar to the inversion effect for N170. In contrast, face orientation had no effect on the amplitude of the fundamental frequency response when identical faces were presented repeatedly. Similar results were also found in their other study in which the presentation rate was 3.5 Hz [13]. Gruss et al. [14] demonstrated that the larger SSVEP
20
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
Fig. 1. The Face stimulus and stimulation sequence. The luminance of the stimuli (the three elements in the white background circle) was sinusoidally changed at a rate of 6 Hz so that the lower contrast face stimulus in the midline, in between the background and the full face stimulus, appeared at an intermediary stage of stimulation. The background and fixation circles were kept unchanged during each 8.67-s stimulation sequence.
amplitude with face inversion was absent at low presentation rates (5 and 10 Hz), but was present at higher rates (15 and 20 Hz). This study analyzed steady-state visual-evoked magnetic fields (SSVEFs) when upright and inverted geometrical faces were presented. Although previous studies examined the face inversion effect on SSVEP as above, its effects on SSVEFs remain unknown. Measuring steady-state responses was shown to have practical advantages such as a high signal-to-noise ratio and shorter experimental time with frequency analyses [e.g. 12]. SSVEFs, as well as SSVEPs, are known to be useful for exploring visual processing [15,16]. The results of the present study may provide clear evidence that the face inversion effect is similar between SSVEF and SSVEP, and that face processing can be examined by measuring SSVEF. 2. Methods 2.1. Participants Eleven healthy adult volunteers (four females and seven males), aged between 25 and 49 years (mean 35.2; SD 6.4), participated in the present study. All participants had normal or corrected-tonormal visual acuity. This study was approved in advance by the Ethics Committee of authors’ affiliation, and all participants gave informed consent. 2.2. Stimuli Stimuli consisted of a face-like geometrical figure (Face), as shown in Fig. 1, and its inversion (iFace). The face inversion effect on N170 was observed with a simple geometrical face-like figure [8]. Thus, a geometrical figure was presented in this study, to avoid any effect of race, gender, attractiveness, or any other factors that could be found in real photographs. All participants reported that they perceived the upright stimulus as a face, and considered the inverted stimulus to be the inversion of the upright stimulus. The figure was composed of a large white circle (30.8 cm in diameter, the same hereafter), three black circles (5.4 cm), and a small red fixation circle (0.4 cm). Visual stimuli were presented by a personal computer (PC, IBM) and video projector (Mirage 2000; CHRISTIE DIGITAL SYSTEM Inc, Kitchener, Canada) placed outside the magnetically-shielded room, and projected on a screen placed 200 cm from the participant. The refresh rate of the projector was 60 Hz. The projected area was subtended 44.2 cm × 32.9 cm. The luminance of the black circles and
the black background was 0.14 cd/m2 , of the large white circle was 2.82 cd/m2 , and of the fixation circle was 0.47 cd/m2 . The stimulation was given with repeating sessions including two stimulus sequences. In a sequence, one of the two stimuli (Face/iFace) appeared and disappeared 52 times on the screen at a rate of 6 Hz (one face every 166.7 ms). One sequence lasted 8.67 s and followed by another sequence with another type of the stimulus, after a 10-s rest. Epochs with MEG signals larger than 2.7 pT/cm were rejected, and a session, i.e. two sequences, was repeated until four sessions were completed without data rejection for each participant. A 30-s or longer rest was inserted every two sessions. The order of the two sequences (Face/iFace) was randomized in each session and across participants. The stimulation frequency was determined as 6 Hz for the following reasons. First of all, we wanted to avoid the contamination of noise at the alpha range (8–12 Hz) at the fundamental frequency. The projector refresh rate (60 Hz) was also considered because the stimulus was presented as a sinusoidal (rather than abrupt, as in a square wave function; see Fig. 1) function, consistent with in previous SSVEP studies [12,13,17]. In addition, a recent study demonstrated that a frequency rate of 6 Hz gave the largest SSVEP response to faces [17]. The experiment was conducted in a quiet, magneticallyshielded, and darkened room. Participants passively viewed the display and were asked to maintain their gaze at the fixation circle throughout the experiment. 2.3. MEG recordings Magnetic signals were recorded using a 306-channel wholehead type MEG system (Vector-view, ELEKTA Neuromag, Helsinki, Finland), which comprised 102 identical triple sensor elements. Each sensor element consisted of two orthogonal planar gradiometers and one magnetometer coupled to a multi-superconducting quantum interference device (SQUID), which provided three independent measurements of the magnetic fields. Signals were recorded with a band-pass filter of 0.1–200 Hz and digitized at 1004 Hz. To record the timing of the stimulus presentation, a trigger was sent from the parallel port of the stimulation computer to the MEG recording. 2.4. Data analysis We analyzed MEG signals recorded from 204 planar-type gradiometers. These planar gradiometers were powerful enough to detect the largest signal just over local cerebral sources. To examine SSVEFs, we conducted spectrum analysis on MEG signals and computed the amplitude spectrum as follows. First, a Fast Fourier Transform (FFT) algorithm was applied to an 8159ms (8192 sampled data) window in each stimulation sequence, for the MEG waveform of each sensor for sequence. To avoid contamination from the transient responses triggered by the onset of the stimulation sequence, the initial 333-ms (2 cycles) waveforms were not included in the spectrum analysis. The frequency resolution was 0.12 Hz (dividing the sampling rate, 1004 Hz, by the sampled data, 8192). The amplitude and phase of each frequency was then averaged across sequences for Face and iFace. The signal-to-noise ratio (SNR) of the amplitude at each sensor was then computed. The SNR was the ratio of the amplitude at the frequency of interest to the average amplitude of 20 neighboring bins [12,13,18]. The SNR at the fundamental frequency (6 Hz) was extracted for further analyses. To compare the effect of face orientation on SSVEFs across the visual areas, we selected representative sensors in each visual area (bilateral temporal and occipital areas) as follows. First, sensors
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
21
Table 1 SNRs and phase components of SSVEP at the fundamental frequency (6 Hz) in the representative sensors. SNR
Temporal Occipital
Mean
Left Right Left Right
Phase Temporal Occipital
Fig. 2. Grouping of the MEG sensors.
were grouped (Fig. 2) according to the reference manual of the software for the MEG system (Graph, ELEKTA Neuromag). We then selected the representative sensor showing the maximum SNR of the amplitude at 6 Hz in each visual area for Face and iFace, respectively. To examine the effect of stimulus orientation on the amplitude at the stimulation frequency (6 Hz), the SNR at 6 Hz of the representative sensor was compared between Face and iFace. To explore the temporal difference induced in SSVEFs by Face and iFace, the phase at the fundamental frequency (6 Hz) of the representative sensors were also compared between the two types of stimuli. Three-way ANOVAs (orientation: Face/iFace, area: occipital/temporal, hemisphere: right/left) were conducted on the SNR of the amplitude and phase at 6 Hz, with the level of significance at 0.05. 3. Results Spectrum analysis revealed clear peaks in the fundamental frequency (6 Hz) and its harmonics under both Face and iFace conditions (Fig. 3). The representative sensor in each visual area
Left Right Left Right
Face
iFace
3.0 ± 0.3 3.6 ± 0.5 5.3 ± 0.6 5.4 ± 0.5
3.0 ± 0.3 3.1 ± 0.2 4.8 ± 0.3 5.1 ± 0.4
Mean Face
iFace
−1.7 ± 29.6 41.0 ± 20.4 −7.4 ± 32.1 −46.7 ± 25.5
39.7 ± 21.2 −58.7 ± 31.5 13.4 ± 22.0 −24.7 ± 20.9
Note: Data were expressed as mean ± SE from eleven participants. A smaller phase value indicates a delay.
(bilateral occipital and temporal areas) was extracted as described in Section 2. Fig. 4 shows the representative MEG waveforms of a participant that were averaged and filtered by a band-pass filter at the fundamental frequency (6 Hz). Although the peak right after the zero point in Fig. 4 appears to indicate a shorter latency than that of N170, the peak may not indicate a response in such a short latency. The stimulus was presented every 167 ms, and overlapping MEG epochs from 167 ms before to 333 ms after the onset of stimulation were averaged. Therefore, the peak right after the zero point may reflect the response to the previous stimulus. Another possibility is that the high contrast black-and-white stimulus used in this study may have caused a shorter latency. The grand-averaged SNR at 6 Hz in the representative sensors are shown in Fig. 5 and Table 1. A three-way ANOVA (orientation: Face/iFace, area: occipital/temporal, hemisphere: right/left) showed that the main effect of area was significant (F(1,10) = 45.84, p < 0.001). Neither the main effect of orientation nor interactions including orientation were significant. These results indicated that SSVEFs in the occipital areas were significantly larger than those in
Fig. 3. Spectrum analysis of a representative participant. (a) The amplitude spectrum of all gradiometers for Face and iFace. Top-view topography at the fundamental frequency (6 Hz), and its second and third harmonics are shown. (b) The signal-to-noise ratio (SNR) of all gradiometers for Face and iFace, and top-view topography at the fundamental frequency (6 Hz), and its second and third harmonics. This participant exhibited large alpha-band activities, but this was not clear in SNRs.
22
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
Fig. 4. MEG waveforms of a representative participant. The enlarged waveforms were the representative sensors’ that were selected for analyses. The figures attached to each enlarged waveform indicate that the sensor is the representative one for Face, iFace, or both. Overlapping MEG epochs from 167 ms before to 333 ms after the onset of stimulation were averaged for each stimulus. The zero time point of each epoch corresponded to the beginning of the appearance of a stimulus. We used MEG epochs starting 3 cycles after the beginning of the stimulation sequence for that averaging. At least 180 MEG epochs were averaged for each stimulus, and were then filtered with a band-pass filter of 6 Hz (1 Hz width).
the temporal areas regardless of differences in the hemisphere and stimulus orientation. Fig. 5 and Table 1 also show the grand-averaged phase components of SSVEF at 6 Hz in the representative sensors. A three-way ANOVA (orientation: Face/iFace, area: occipital/temporal, hemisphere: right/left) revealed that a one-way interaction between orientation and hemisphere (F(1, 10) = 6.53, p = 0.029) and the twoway interaction across orientation, area, and hemisphere (F(1, 10) = 11.20, p = 0.007) were significant. No other main effect or interaction was significance. Post hoc analyses showed that a simple-simple main effect of orientation was only significant in the right temporal area (F(1, 10) = 12.90, p = 0.005). The average phase difference between Face and iFace in the area was 99.7◦ , which
corresponds to 46 ms at 6 Hz. This simple-simple main effect in the right temporal area may have caused the significant one-way interaction between orientation and hemisphere and the opposite effect of orientation between the two hemispheres, while the main effect of orientation was not significant in either hemisphere. 4. Discussion Both upright and inverted geometrical faces presented at 6 Hz elicited a clear response oscillating at that specific frequency and its harmonics, i.e. SSVEFs. The topography of the fundamental and harmonic responses was basically distributed over the occipital and temporal areas.
Fig. 5. Vector plots of the SNRs and phase components of SSVEP at the fundamental frequency (6 Hz) in the representative sensors. Data were averaged across participants. The radius of the small semicircle indicates SNR = 1, and of the large one does SNR = 5. Phase delays (counter-clockwise lags) were observed under iFace condition, especially in the right temporal area.
A. Tsuruhara et al. / Neuroscience Letters 562 (2014) 19–23
Our results revealed no significant difference in the amplitude of SSVEFs between the upright and inverted faces, which is inconsistent with the inversion effect of N170 reported in some previous VEP studies [6–8]. A geometrical face-like figure was used as a stimulus in this study, and the lack of reality may have led to the absence of an enlarged response to the inverted stimulus. However, Tomalski and Jonson [8] presented a simple geometrical face-like figure that was similar to ours, and reported an enlarged N170 with inversion. In addition, previous SSVEP studies that presented face photographs did not find a larger amplitude with face inversion [12,13]. Therefore, it is more likely that factors other than the reality of the face stimulus were related to the absence of the face inversion effect on amplitude. One possibility is the stimulus presentation rate. Gruss et al. [14] recently reported that the larger SSVEP amplitude with face inversion was only observed when the presentation rate of the face stimulus was high (15 Hz∼), which is consistent with the present results as well as previous SSVEP studies [7,8]. Another possibility is that the effect of face inversion on amplitude may not necessarily be significant [1], especially with MEG measurement [5,9,10]. In contrast to the absence of an effect on amplitude, the results of the present study showed a phase delay in the right temporal area with face inversion. Delayed response to face inversion were also observed in previous studies regardless of the measured response (N170 [1,5–8], M170 [5,9,10] or SSVEPs [13]), the reality of the face stimulus (a photograph [e.g. 1] or a face-like figure [8]), or the presentation rate to measure steady-state responses. Although further studies are needed to clarify the mechanism relating the larger amplitude and delayed latency with face inversion, our results suggest that different mechanisms underlie the two indices of the face inversion effect. As reported in previous SSVEP and SSVEF studies [e.g. 12], measuring steady-state responses has advantages, such as a relatively short measuring time and highly distinctive response, which allows a focus to be placed on just the amplitude at the chosen frequency, with no ambiguity in selecting the component of interest. In addition, our results suggest that an independent mechanism may affect the delayed latency and increased amplitude with face inversion because the former was found in both steady-state and transient response, whereas the latter was not necessarily. This finding indicates that measuring steady-state responses could contribute to novel findings on face processing in the human brain with many practical advantages. Acknowledgments This study was supported by Grants-in-Aid for Scientific Research on Innovative Areas “Face perception and recognition”
23
from MEXT KAKENHI (20119007, 23119722), and a Grant-in-Aid for Young Scientists (B) (23730709) from JSPS. References [1] S. Bentin, T. Allison, A. Puce, E. Perez, G. McCarthy, Electrophysiological studies of face perception in humans, J. Cogn. Neurosci. 8 (1996) 551–565. [2] M. Eimer, Effects of face inversion on the structural encoding and recognition of faces. Evidence from event-related brain potentials, Brain Res. Cogn. Brain Res. 10 (2000) 145–158. [3] N. George, J. Evans, N. Fiori, J. Davidoff, B. Renault, Brain events related to normal and moderately scrambled faces, Brain Res. Cogn. Brain Res. 4 (1996) 65–76. [4] R.J. Itier, M.J. Taylor, N170 or N1? Spatiotemporal differences between object and face processing using ERPs, Cereb. Cortex. 14 (2004) 132–142. [5] K. Linkenkaer-Hansen, J.M. Palva, M. Sams, J.K. Hietanen, H.J. Aronen, R.J. Ilmoniemi, Face-selective processing in human extrastriate cortex around 120 ms after stimulus onset revealed by magneto- and electroencephalography, Neurosci. Lett. 253 (1998) 147–150. [6] B. Rossion, J.F. Delvenne, D. Debatisse, V. Goffaux, R. Bruyer, M. Crommelinck, J.M. Guérit, Spatio-temporal localization of the face inversion effect: an eventrelated potentials study, Biol. Psychol. 50 (1999) 173–189. [7] B. Rossion, I. Gauthier, M.J. Tarr, P. Despland, R. Bruyer, S. Linotte, M. Crommelinck, The N170 occipito-temporal component is delayed and enhanced to inverted faces but not to inverted objects: an electrophysiological account of face-specic processes in the human brain, Neuroreport 11 (2000) 69–74. [8] P. Tomalski, M.H. Johnson, Cortical sensitivity to contrast polarity and orientation of faces is modulated by temporal-nasal hemifield asymmetry, Brain Imaging Behav. 6 (2012) 88–101. [9] S. Watanabe, R. Kakigi, A. Puce, The spatiotemporal dynamics of the face inversion effect: A magneto- and electro-encephalographic study, Neuroscience 116 (2003) 879–895. [10] J. Liu, M. Higuchi, a Marantz, N. Kanwisher, The selectivity of the occipitotemporal M170 for faces, Neuroreport 11 (2000) 337–341. [11] D. Regan, Some characteristics of average steady-state and transient responses evoked by modulated light, Electroencephalogr. Clin. Neurophysiol. 20 (1966) 238–248. [12] B. Rossion, E.A. Prieto, A. Boremanse, D. Kuefner, G. Van Belle, A steadystate visual evoked potential approach to individual face perception: Effect of inversion, contrast-reversal and temporal dynamics, NeuroImage 63 (2012) 1585–1600. [13] B. Rossion, A. Boremanse, Robust sensitivity to facial identity in the right human occipito-temporal cortex as revealed by steady-state visual-evoked potentials, J. Vis. 11 (2011) 1–21. [14] L.F. Gruss, M.J. Wieser, S.R. Schweinberger, A. Keil, Face-evoked steady-state visual potentials: effects of presentation rate and face inversion, Front. Hum. Neurosci. 6 (2012) 316 1–10. [15] Y. Chen, A.K. Seth, J.A. Gally, G.M. Edelman, The power of human brain magnetoencephalographic signals can be modulated up or down by changes in an attentive visual task., Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 3501–3506. [16] B. Cottereau, J. Lorenceau, A. Gramfort, M. Clerc, B. Thirion, S. Baillet, Phase delays within visual cortex shape the response to steady-state visual stimulation, NeuroImage 54 (2011) 1919–1929. [17] E. Alonso-Prieto, G. Van Belle, J. Liu-Shuang, A.M., Norcia, B. Rossion, The 6 Hz fundamental stimulation frequency rate for individual face discrimination in the right occipito-temporal cortex, Neuropsychologia (2013) 1–13. [18] R. Srinivasan, D.P. Russell, G.M. Edelman, G. Tononi, Increased synchronization of neuromagnetic responses during conscious perception, J. Neurosci. 19 (1999) 5435–5448.
