Elsevier

Neuropsychologia

Volume 51, Issue 13, November 2013, Pages 2863-2875
Neuropsychologia

The 6 Hz fundamental stimulation frequency rate for individual face discrimination in the right occipito-temporal cortex

https://doi.org/10.1016/j.neuropsychologia.2013.08.018Get rights and content

Highlights

  • Different faces elicit larger EEG responses than same faces between 4 and 8.33 Hz.

  • We define the frequency-tuning function for individual face discirmination.

  • The peak of the individual face discrimination response is at 5.88 Hz (170 ms cycle).

  • Visual areas of the right occipito-temporal cortex individualize a face in 170 ms.

  • Future paradigms relying on repetition suppression should be optimized accordingly.

Abstract

What is the stimulus presentation rate at which the human brain can discriminate each exemplar of a familiar visual category? We presented faces at 14 frequency rates (1.0–16.66 Hz) to human observers while recording high-density electroencephalogram (EEG). Different face exemplars elicited a larger steady-state visual evoked (ssVEP) response than when the same face was repeated, but only for stimulation frequencies between 4 and 8.33 Hz, with a maximal difference at 5.88 Hz (170 ms cycle). The effect was confined to the exact stimulation frequency and localized over the right occipito-temporal cortex. At high frequency rates (>10 Hz), the response to different and identical exemplars did not differ, suggesting that the fine-grained analysis needed for individual face discrimination cannot be completed before the next face interrupts, or competes, with the processed face. At low rates (<3 Hz), repetition suppression could not be identified at the stimulation frequency, suggesting that the neural response to an individual face is temporally dispersed and distributed over different processes. These observations indicate that at a temporal rate of 170 ms (6 faces/s) the face perception network is able to fully discriminate between each individual face presented, providing information about the temporal bottleneck of individual face discrimination in humans. These results also have important practical implications for optimizing paradigms that rely on repetition suppression, and open an avenue for investigating complex visual processes at an optimal range of stimulation frequency rates.

Introduction

The visual system is able to extract diagnostic information not only at a coarse level to categorize a visual stimulus as a face (“face detection”, e.g., (Crouzet, Kirchner, & Thorpe, 2010)) but also at a finer-grained level in order to discriminate it from other individual faces (“individual face discrimination”, or more simply “face discrimination”). Behaviorally, human observers are able to discriminate different individual faces accurately in a few hundreds of milliseconds (e.g., (Jacques, d′Arripe, & Rossion, 2007)). Face-selective cells of the same neuronal population in the monkey infero-temporal (IT) cortex (Gross, Rocha-Miranda, & Bender, 1972) discharge at different rates to different individual faces, suggesting a mechanism to discriminate faces based on sparse population coding (Abbott et al., 1996, Leopold et al., 2006, Young and Yamane, 1992). Yet, in humans, neuroimaging studies have identified several areas of the ventral occipito-temporal cortex that are sensitive to differences between individual faces, showing a larger neural response when different faces are presented successively compared to the repetition of the exact same face (“repetition suppression”, or “neural adaptation”, e.g., (Davies-Thompson, Gouws, & Andrews, 2009; Gilaie-Dotan, Gelbard-Sagiv, & Malach, 2010; Grill-Spector & Malach, 2001)).

An unexplored issue concerns the sensitivity of the human visual system to stimulus presentation rate. That is, how many faces can be discriminated in 1 s of time? This issue is different than defining the speed of conduction of information about faces, information that is usually inferred from either the latency of face-related event-related potentials (ERPs, e.g., (Bentin, Allison, Puce, Perez, & McCarthy, 1996; Jeffreys, 1996b; Rossion & Jacques, 2011; VanRullen, 2011)) or the latency of discharges of face-selective neurons in the monkey infero-temporal cortex (e.g. (Kiani, Esteky, & Tanaka, 2005; Sugase, Yamane, Ueno, & Kawano, 1999)). It is also a different issue than the presentation duration that is needed to recognize a single individual face and activate its neural representation (Tanskanen, Nasanen, Ojanpaa, & Hari, 2007). Rather, it concerns the time that is necessary to process a face at the individual level before the next one can be handled. This is an important issue to resolve for at least four reasons. First, generally speaking, the temporal interval over which the system blurs information together, known as its ‘temporal resolution’, is important for understanding human visual perception and may be a fundamental attribute of brain function (Brown, 1965, Hawken et al., 1996, Holcombe, 2009, Keysers et al., 2001, Krukowski and Miller, 2001, Tovee et al., 1994). Second, more specifically, humans live in a highly dynamic visual world in which they can be exposed to many different faces simultaneously, or within a short timeframe. Being able to discriminate these faces rapidly may be critical for adequate social interactions. Third, clarifying the maximal rate at which individual faces can be discriminated would have important implications for understanding the neuro-functional basis of face perception, putting constraints on the nature of the information that can be extracted during a certain amount of time and potentially transferred to higher levels of processing. Fourth, determining the frequency (rate) tuning function of the discrimination of individual faces would have practical implications for optimizing studies that rely on neural repetition suppression in high-level vision (Grill-Spector, Henson, & Martin, 2006).

The steady state visual evoked potential (ssVEP) technique recorded by means of electroencephalogram (EEG) or magnetoencephalogram (MEG) is well suited to address the above mentioned issue because it benefits from the observation that in humans (and monkeys, see e.g., (Nakayama & Mackeben, 1982)), a periodic stimulation elicits a periodic EEG/MEG response at the exact frequency of stimulation and its harmonics, the ssVEPs (Regan, 1966, Regan, 1989). Thus, it can be used to characterize the visual system’s sensitivity to stimulus presentation rate noninvasively, at a macroscopic brain level, with high signal-to-noise ratio and high frequency resolution (Regan, 1966, Regan, 1989). Recently, the ssVEP approach was extended to study face detection (Ales, Farzin, Rossion, & Norcia, 2012) and individual face discrimination (Rossion and Boremanse, 2011, Rossion et al., 2012), using single stimulation frequencies. In line with these latter studies, the temporal frequency tuning function for individual face discrimination was assessed here by comparing the electrophysiological response when individual faces changed at every cycle to the response when the same face was presented at every cycle.

Specifically, we measured neural temporal frequency tuning for face discrimination by showing human observers pictures of faces presented at different frequency rates (1 to 16.66 Hz, i.e., a face every 1000  to every ~60 ms; Fig. 1), while recording high-density scalp EEG. This range of frequencies was selected by considering several factors. First, a minimal stimulation frequency of 1 Hz is sufficiently low to clearly observe the transient ERP responses following face stimulation, given that the return to baseline in typical visual ERP studies usually takes about half a second (Nunez & Srinivasan, 2006). Second, the maximal responses to low-level visual stimulation in EEG/MEG studies are observed either at about 15 Hz (Hermann, 2001, Pastor et al., 2003) or below (e.g., 8–10 Hz in (Fawcett et al., 2004, Regan, 1966, Regan, 1989, Singh et al., 2003, Srinivasan et al., 2006, Van Der Tweel and Lunel, 1965).2 Given that the rate of discrimination should be lower for high-level complex visual stimuli such as faces than for low-level visual stimuli, a range between 1 and 16.66 Hz should cover most if not all of the responses of interest. Finally, intermediate frequency rates were selected in order to be able to define a meaningful frequency tuning function for individual face discrimination. For this reason, we used a large number of frequency rates (12) which were relatively equally spaced, considering the technical constraints (refresh rate of the monitor). We also included one condition encompassing an exact cycle of 170 ms (i.e., 5.88 Hz) because a cycle of 170 ms corresponds to the peak of the face-sensitive N170 component (Bentin et al., 1996, Rossion and Jacques, 2008). Most importantly, the earliest repetition effects for individual faces are observed at the peak of the N170 (e.g., Itier & Taylor, 2002; Jacques et al., 2007, Rossion and Jacques, 2011) and these effects are often prolonged until 250–300 ms (e.g., Caharel et al., 2009, Schweinberger et al., 2002). Even though the latency of ERPs does not reflect the temporal rate of processing, these observations suggest that a range of stimulation rates between 3.5/4 and 6 Hz might be associated with the largest difference between repeated and different individual faces.

Section snippets

Participants

Four healthy adult participants (ages 28, 30, 35, 36 years), all right-handed males with normal or corrected-to-normal vision, took part in the study for payment. They were tested four times on different days, over a period of four weeks (16 EEG recordings in total). Eight new right-handed participants (two males) were tested in a single EEG recording session for a complementary experiment. Written informed consent was obtained from all participants prior to the experiments, which were approved

Results

At every stimulation frequency and in both conditions, there were large EEG responses confined to single frequency bins (0.02 Hz) located at the fundamental frequency of stimulation (1F) and its harmonics (2F, 3F,…) (Fig. 2). These responses indicate that the brain synchronized precisely with the rate of visual stimulation, leading to clear ssVEPs (Regan, 1966, Regan, 1989). In both conditions, SNR at the fundamental frequency (first harmonic) was by far the largest, with the exception of the

Discussion

Adaptation to individual faces – as measured by suppression of neural responses to face repetition – is tuned to a relatively narrow range of temporal frequencies (>3.03–<9.09 Hz) centered on approximately 6 Hz. At all suitable frequency ranges, this effect is observed over the right occipito-temporal cortex, a distinct signature of face-specific perceptual processes (e.g., (Bentin et al., 1996; Sergent, Ohta, & MacDonald, 1992; Rossion & Jacques, 2008).

Since the specific ~4–8 Hz range of the

Acknowledgments

This research was supported by an ERC grant (facessvep 284025) to BR. BR, JLS and GVB are supported by the Belgian National Fund for Scientific Research (FNRS).

References (80)

  • C. Keysers et al.

    Visual masking and RSVP reveal neural competition

    Trends in Cognitive Sciences

    (2002)
  • W. Klimesch

    EEG alpha and theta oscillations reflect cognitive and memory performance: A review and analysis

    Brain Research. Brain Research Reviews

    (1999)
  • W. Klimesch

    Alpha-band oscillations, attention, and controlled access to stored information

    Trends in Cognitive Sciences

    (2012)
  • E.V. Mnatsakanian et al.

    Familiar-face recognition and comparison: Source analysis of scalp-recorded event-related potentials

    Clinical Neurophysiology

    (2004)
  • A. Mouraux et al.

    Across-trial averaging of event-related EEG responses and beyond

    Magnetic Resonance Imaging

    (2008)
  • K. Nakayama et al.

    Steady state visual evoked potentials in the alert primate

    Vision Research

    (1982)
  • P.L. Nunez et al.

    A theoretical basis for standing and traveling brain waves measured with human EEG with implications for an integrated consciousness

    Clinical Neurophysiology

    (2006)
  • D. Regan

    Some characteristics of average steady-state and transient responses evoked by modulated light

    Electroencephalography and Clinical Neurophysiology

    (1966)
  • B. Rossion et al.

    Does physical interstimulus variance account for early electrophysiological face sensitive responses in the human brain? Ten lessons on the N170

    NeuroImage

    (2008)
  • B. Rossion et al.

    A steady-state visual evoked potential approach to individual face perception: Effect of inversion, contrast-reversal and temporal dynamics

    NeuroImage

    (2012)
  • R. Santarelli et al.

    Generation of human auditory steady-state responses (SSRs). II: Addition of responses to individual stimuli

    (1995)
  • C. Schiltz et al.

    Faces are represented holistically in the human occipito-temporal cortex

    NeuroImage

    (2006)
  • S.R. Schweinberger et al.

    Event-related brain potential evidence for a response of inferior temporal cortex to familiar face repetitions

    Brain Research. Cognitive Brain Research

    (2002)
  • R.B. Silberstein et al.

    Steady-state visually evoked potential topography during the Wisconsin card sorting test

    Electroencephalography and Clinical Neurophysiology

    (1995)
  • T. Tanskanen et al.

    Face recognition and cortical responses: Effect of stimulus duration

    NeuroImage

    (2007)
  • L.H. Van Der Tweel et al.

    Human visual responses to sinusoidally modulated light

    Electroencephalography and Clinical Neurophysiology

    (1965)
  • J.D. Victor et al.

    The human visual evoked potential: Analysis of components due to elementary and complex aspects of form

    Vision Research

    (1985)
  • L.F. Abbott et al.

    Representational capacity of face coding in monkeys

    Cerebral Cortex

    (1996)
  • J.M. Ales et al.

    An objective method for measuring face detection thresholds using the sweep steady-state visual evoked response

    Journal of Visualized

    (2012)
  • S. Bentin et al.

    Electrophysiological studies of face perception in humans

    Journal of Cognitive Neuroscience

    (1996)
  • J.L. Brown

    Flicker an intermittent stimulation

  • N.A. Busch et al.

    Spontaneous EEG oscillations reveal periodic sampling of visual attention

    Proceedings of the National Academy of Sciences of the United States of America

    (2010)
  • G. Buzsaki

    Rhythms of the brain

    (2006)
  • G. Buzsaki et al.

    Oscillatory and intermittent synchrony in the hippocampus: Relevance to memory trace formation

  • A. Capilla et al.

    Steady-state visual evoked potentials can be explained by temporal superposition of transient event-related responses

    PLoS One

    (2011)
  • S.M. Crouzet et al.

    Fast saccades toward faces: Face detection in just 100 ms

    Journal of Visualized

    (2010)
  • J. Ding et al.

    Attentional modulation of SSVEP power depends on the network tagged by the flicker frequency

    Cerebral Cortex

    (2006)
  • S. Fletcher-Watson et al.

    Rapid detection of person information in a naturalistic scene

    Perception

    (2008)
  • W.A. Freiwald et al.

    A face feature space in the macaque temporal lobe

    Nature Neuroscience

    (2009)
  • R. Galambos et al.

    A 40-Hz auditory potential recorded from the human scalp

    Proceedings of the National Academy of Sciences of the United States of America

    (1981)
  • Cited by (86)

    View all citing articles on Scopus
    1

    Current address: Department of Ophthalmology & Visual Sciences, 2550 Willow Street, Vancouver, BC, Canada V5Z 3N9.

    View full text