Category-selective attention modulates unconscious processes in the middle occipital gyrus

https://doi.org/10.1016/j.concog.2013.02.007Get rights and content

Abstract

Many studies have revealed the top-down modulation (spatial attention, attentional load, etc.) on unconscious processing. However, there is little research about how category-selective attention could modulate the unconscious processing. In the present study, using functional magnetic resonance imaging (fMRI), the results showed that category-selective attention modulated unconscious face/tool processing in the middle occipital gyrus (MOG). Interestingly, MOG effects were of opposed direction for face and tool processes. During unconscious face processing, activation in MOG decreased under the face-selective attention compared with tool-selective attention. This result was in line with the predictive coding theory. During unconscious tool processing, however, activation in MOG increased under the tool-selective attention compared with face-selective attention. The different effects might be ascribed to an interaction between top-down category-selective processes and bottom-up processes in the partial awareness level as proposed by Kouider, De Gardelle, Sackur, and Dupoux (2010). Specifically, we suppose an “excessive activation” hypothesis.

Introduction

Understanding how top-down processes modulate bottom-up processes is a major challenge to reveal the mechanisms of adaptive perception processes in the brain. Based on much evidence, perception processes depend to a large degree on expectations derived from previous experience, and on generalized knowledge (Engel, Fries, & Singer, 2001). Gilbert and Sigman (2007) have advanced the concept of “brain states” to expand the conventional definition of top-down influences which are usually equated with attention (Gilbert & Sigman, 2007). They consider that the role of top-down influences is to set the cortex in a specific working state according to behavioral requirements.

Various previous studies demonstrated that even early visual processes can be modulated by top-down processes across a wide range of experimental paradigms (Rauss, Schwartz, & Pourtois, 2011). For example, Mechelli et al., 2003, Mechelli et al., 2004 used the method of Dynamic Causal Modelling (DCM) and found that supraliminal category-selective (faces, houses and chairs) activations in extrastriate cortex are mediated both by bottom-up connections from early visual areas and by top-down connections from prefrontal cortex (Mechelli et al., 2004, Mechelli et al., 2003). In addition, some studies showed that attentional load could modify activity in regions of the visual cortex (Kelley and Lavie, 2011, Rauss et al., 2009). Other studies revealed that conscious visual perception is modulated by spatial attention (Gandhi, Heeger, & Boynton, 1999), feature-selective attention (Müller et al., 2006), and category-selective attention (Gazzaley, Cooney, McEvoy, Knight, & D’esposito, 2005). Moreover, in animal experiments, results showed that top-down processes could modulate the early cortical activity of sensory processes (Ito and Gilbert, 1999, Li et al., 2004). Even in the absence of visual stimuli, activity in the visual cortex could be modulated by attention (Kastner, Pinsk, De Weerd, Desimone, & Ungerleider, 1999).

Besides the evidences of top-down effects on supraliminal processes, it is an important aspect to understand whether top-down processes can modulate unconscious processes to understand the mechanisms of perception, because we often process the majority of sensory information without consciousness. In addition, because unconscious stimuli can exert effects on motivation (Custers & Aarts, 2010), experience of control (Linser & Goschke, 2007), semantic processing (Dehaene et al., 2001), emotional processing (Luo et al., 2004, Whalen et al., 1998), cognitive control (Lau and Passingham, 2007, van Gaal et al., 2010), and object recognition (Stoerig & Cowey, 1997), top-down effects on unconscious processes might modulate these unconscious influences. In fact, some studies have revealed that unconscious (prime) processes are modulated by conscious top-down factors. For example, spatial attention can influence sensorimotor processes that are entirely separated from conscious perception (Sumner, Tsai, Yu, & Nachev, 2006) and unconscious letter prime effects (Marzouki, Grainger, & Theeuwes, 2007). There is also evidence showing that the effect of unconscious masked priming depends on temporal attention (Kiefer and Brendel, 2006, Naccache et al., 2002). Using fMRI and a continuous flash suppression paradigm, Bahrami, Lavie, and Rees (2007) observed reduced V1 activity under high attentional load even for invisible stimuli compared with that under low attentional load (Bahrami et al., 2007). Moreover, recent studies measuring event-related brain potentials demonstrated selective top-down influence of task sets on subliminal semantic processing and motor response selection (Kiefer and Martens, 2010, Martens et al., 2011).

However, there is little research about how category-selective attention could modulate unconscious visual processes. Category-selective attention is based on discrimination and categorization of features of all the objects around us. Studies showed that feature-based attention is different from other attentional processes such as spatial attention (Kanai et al., 2006, McMains et al., 2007). Combined with the evidence that early visual processes can be modulated by top-down factors, in the present study, we aimed at testing the hypothesis that the activity in the early visual cortex elicited by unconsciously perceived categories of pictures (i.e. faces and tools) might be selectively modulated by top-down processes of different types of category cues (“face” and “tool”). Specifically, according to the predictive coding theory (Rao & Ballard, 1999), we postulated that activity in the early visual cortex would decrease under congruent conditions (unconscious pictures with same type of category cues) compared with incongruent conditions (unconscious pictures with different type of category cues).

Furthermore, Kouider et al. assume that the representations at each processing level could be accessed independently from each other and advocated the concept of “partial awareness” (Kouider et al., 2010). In our experiment, the masked tools underlay partial awareness at a low processing level. The participants could sense the changes in contour and orientation of tool pictures but they were unable to recognize them. Therefore, we additionally analyzed the effect of category-selective top-down modulation on unrecognized masked tools with partial awareness.

Section snippets

Participants

As paid volunteers, 21 adults (10 women, 11 men) aged 18–26 years (mean age 22.7 years) from Southwest University in China participated in this experiment. All participants gave written informed consent, were right-handed, had no history of current or past neurological or psychiatric illness, and had normal or corrected-to-normal vision. This study has been approved by the IRB at Southwest China University. Informed consent was obtained from each participant after the nature of the study was

Visibility test

The participants reported not recognizing the masked faces/tools. But they could experience the changes of contour and orientation under the condition of masked tools but not under the condition of masked faces. In the conservative measure of stimulus visibility, 18 of 21 participants scored at chance level (binomial test, p > .05), suggesting that these individuals were unable to perceive masked faces/tools. The other three participants were excluded from analyses. Also at the group level,

Discussion

In the present fMRI study, we investigated how top-down category-selective attention modulates unconscious processes. The results showed that category-selective top-down processes modulated unconscious face/tool processing in the MOG (middle occipital gyrus), but effects for face stimuli differed from effects for tool stimuli. Processing unconsciously perceived faces resulted in decreased activations in MOG under top-down face-selective attention compared with top-down tool-selective attention.

Acknowledgments

This research was supported by the National Natural Science Foundation of China (30970892) and the Fundamental Research Funds for the Central Universities (NSKD11002).

References (43)

  • Q. Luo et al.

    Emotional valence of words modulates the subliminal repetition priming effect in the left fusiform gyrus: An event-related fMRI study

    Neuroimage

    (2004)
  • Y. Marzouki et al.

    Exogenous spatial cueing modulates subliminal masked priming

    Acta Psychologica

    (2007)
  • K. Rauss et al.

    Top-down effects on early visual processing in humans: A predictive coding framework

    Neuroscience and Biobehavioral Reviews

    (2011)
  • A. Alink et al.

    Stimulus predictability reduces responses in primary visual cortex

    Journal of Neuroscience

    (2010)
  • V. Cardin et al.

    Top-down modulations in the visual form pathway revealed with dynamic causal modeling

    Cerebral Cortex

    (2011)
  • M.M. Chun et al.

    A taxonomy of external and internal attention

    Annual review of psychology

    (2011)
  • R. Custers et al.

    The unconscious will: How the pursuit of goals operates outside of conscious awareness

    Science

    (2010)
  • E. Dahlin et al.

    Transfer of learning after updating training mediated by the striatum

    Science

    (2008)
  • S. Dehaene et al.

    Cerebral mechanisms of word masking and unconscious repetition priming

    Nature Neuroscience

    (2001)
  • A.K. Engel et al.

    Dynamic predictions: Oscillations and synchrony in top-down processing

    Nature Reviews Neuroscience

    (2001)
  • F. Fang et al.

    Perceptual grouping and inverse fMRI activity patterns in human visual cortex

    Journal of Vision

    (2008)
  • Cited by (50)

    • Common and Distinct Neural Patterns of Attention-Deficit/Hyperactivity Disorder and Borderline Personality Disorder: A Multimodal Functional and Structural Meta-analysis

      2023, Biological Psychiatry: Cognitive Neuroscience and Neuroimaging
      Citation Excerpt :

      Given the predominant linkage with the action domain, the SMG contributes to action reprogramming for rapid adaptation to external alterations and to maintaining the memory flow of serial orders (87,88). Apart from relating to visuospatial functioning, the middle occipital gyrus processes the category-selective attention via facial recognition (89). Notably, large-scale mega-analyses revealed that reduced volume of subcortical regions was not found in adults, but rather was found only in children with ADHD compared with typically developing control subjects, suggesting a neuropathological model of altered subcortical trajectories (90,91).

    • Common and distinct neurofunctional representations of core and social disgust in the brain: Coordinate-based and network meta-analyses

      2022, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      The corresponding meta-analytic functional decoding of the regions revealed primarily terms associated with salience and social communication, including attention as well as face and communication. In line with the functional decoding, the identified regions have been previously associated with emotional face processing (Fusar-Poli et al., 2009b), with the middle occipital cortex involved in early face processing stages (Tu et al., 2013) and the fusiform gyrus being involved in visual attention and higher level face perception processing (Weiner and Zilles, 2016). In addition, the superior temporal regions have been strongly involved in social cognitive processes including theory of mind and mentalizing (Frith, 2007; Zilbovicius et al., 2006).

    View all citing articles on Scopus
    View full text