Much of science is reductionist, whereby methodology and theory are expressed in ever-smaller and fine-grained entities in order to investigate and characterize a phenomenon or process of interest (Koch Reference Koch2012). In the target article Morsella et al. apply this method and introduce a provocative conceptual framework to advance our understanding of consciousness. Specifically, they argue that in order to study this most elusive of neural processes, one should look at it in its most basic form. Indeed, they focus on a hypothetical, simplified human-like mammal that, although conscious, does not possess high-level abilities associated with human information-processing – for example, indirect cognitive control. In addition, they study the olfactory system, as opposed to the more complex visual system which has been the focus of previous consciousness research (Crick & Koch Reference Crick and Koch2003). I commend the authors for taking on such a daunting problem, their thorough analysis of the relevant literature, and their novel approach; however, the basic framework Morsella et al. present may be limited it terms of its explanatory power. Indeed, it appears unable to account for a wide array of literature which shows that, in the visual domain (humans' dominant sense), higher-level cognitive processes, such as top-down attention and re-entrant processing in the brain – whereby frontal-parietal regions project back to sensory areas – play a vital role in consciousness.
A recent trend in the empirical literature has been to argue that conscious and top-down attention are distinct and do not share a causal relationship. Put differently, attention can be devoted to unconscious stimuli, and consciousness can be achieved without a stimulus being the subject of attention (Koch & Tsuchiya Reference Koch and Tsuchiya2007). However, as elegantly argued by Cohen et al. (Reference Cohen, Cavanagh, Chun and Nakayama2012), there is little evidence for this. Specifically, attention has to be able to be deployed to unconscious stimuli, otherwise how would these items reach consciousness? Having said this, attention is not all or none, and consequently, just because an item is attended does not guarantee that it will enter awareness (thus, attention is necessary but not sufficient). In addition, Cohen et al. provide a thorough review of the literature and demonstrate that previous findings reporting properties of the environment (e.g., gist of a scene; Mack & Rock Reference Mack and Rock1998) entering awareness without attention are flawed and do not hold up to dual-task attentional manipulations (Cohen et al. Reference Cohen, Alvarez and Nakayama2011). In short, there is strong evidence that top-down attention and consciousness are inherently linked. Given that top-down attention is capacity limited/effortful, this suggests that consciousness is not entirely passive and that it relies on higher-level goal-related information processing (not possessed by the creature in the cave).
Evidence from neuroscience also suggests that visual consciousness draws on high-level resources and that the neural basis for consciousness goes beyond subcortical regions. Arguably the most heavily used phenomenon for studying visual consciousness is the attentional blink (AB): This refers to subjects' typically impaired performance in reporting the second of two targets in a rapid serial visual presentation stream if it appears within 200–500 msec of the first target relative to when it appears outside this window (Raymond et al. Reference Raymond, Shapiro and Arnell1992). The AB is thought to reflect a failure of perceptual awareness, where the missed second target does not enter consciousness. Considerable functional imaging work has investigated the neural substrates of this phenomenon and localized it to frontal-parietal regions. For example, Marois et al. (Reference Marois, Yi and Chun2004) found, using a dual-target Rapid Serial Visual Presentation (RSVP) task with scene stimuli for target 2, that when this item was missed, and it failed to enter awareness, it still activated the parahippocampal place, whereas the lateral prefrontal cortex did not respond to missed second targets. Similarly, research on object substitution masking (OSM; Enns & Di Lollo Reference Enns and Di Lollo1997), where a sparse (e.g., four-dot), temporally trailing mask obscures the visibility of a briefly presented target, also implicates top-down processing. Indeed, Hirose et al. (Reference Hirose, Kihara, Mima, Ueki, Fukuyama and Osaka2007) observed that disruption of visual area MT (middle temporal/V5) via repetitive transcranial magnetic stimulation (rTMS) influenced the magnitude of OSM. This is consistent with the re-entrant processing account of OSM offered by Di Lollo et al. (Reference Di Lollo, Enns and Rensink2000) who argue that the phenomenon represents iterative hypothesis testing with anterior (e.g., frontal) regions sampling high-resolution sensory information from posterior visual areas via feedback. Finally, recent work by Zhang et al. (Reference Zhang, Jamison, Engel, He and He2011) has shown that even the inter-ocular fluctuations in consciousness that are observed during binocular rivalry (BR), which were previously thought to be automatic and spontaneous, are dependent on top-down attention. These authors employed electroencephalogram (EEG) frequency tagging via steady-state visual evoked potentials to show that stimulus representations stopped rivaling when attention was diverted away from the rivalry display. Collectively, these results using the prominent visual consciousness phenomena of the AB, OSM, and BR, all show that top-down attention processing plays a vital role in consciousness. In addition, when considering that top-down attention has been shown to tap frontal-parietal cortical circuits (see Corbetta et al. Reference Corbetta, Patel and Shulman2008), this suggests that consciousness is, at least partially, both high level and active rather than passive.
So where does this leave us with the passive frame theory offered by Morsella et al.? It is clear that theirs is an internally consistent framework that offers a unique perspective and impressive account of the most basic form of consciousness. However, the challenge of characterizing higher levels of consciousness (although AB, OSM, and BR are thought to be perceptual phenomena) in a species that leads a richer and more challenging life and that has more sophisticated and complex cognitive hardware, remains to be addressed. It may be the case that passive frame theory could be extended to accommodate a broader range of findings and phenomena in humans; however, this would require significant revision. It must also be noted that such a revision is necessary if this theory is to provide a comprehensive framework that facilitates the interpretation of consciousness data. Having said this, Morsella et al. are not alone here, and indeed the difficulty of marrying reductionism and generalizability is one that all of us working in this field face.
You have
Access
Much of science is reductionist, whereby methodology and theory are expressed in ever-smaller and fine-grained entities in order to investigate and characterize a phenomenon or process of interest (Koch Reference Koch2012). In the target article Morsella et al. apply this method and introduce a provocative conceptual framework to advance our understanding of consciousness. Specifically, they argue that in order to study this most elusive of neural processes, one should look at it in its most basic form. Indeed, they focus on a hypothetical, simplified human-like mammal that, although conscious, does not possess high-level abilities associated with human information-processing – for example, indirect cognitive control. In addition, they study the olfactory system, as opposed to the more complex visual system which has been the focus of previous consciousness research (Crick & Koch Reference Crick and Koch2003). I commend the authors for taking on such a daunting problem, their thorough analysis of the relevant literature, and their novel approach; however, the basic framework Morsella et al. present may be limited it terms of its explanatory power. Indeed, it appears unable to account for a wide array of literature which shows that, in the visual domain (humans' dominant sense), higher-level cognitive processes, such as top-down attention and re-entrant processing in the brain – whereby frontal-parietal regions project back to sensory areas – play a vital role in consciousness.
A recent trend in the empirical literature has been to argue that conscious and top-down attention are distinct and do not share a causal relationship. Put differently, attention can be devoted to unconscious stimuli, and consciousness can be achieved without a stimulus being the subject of attention (Koch & Tsuchiya Reference Koch and Tsuchiya2007). However, as elegantly argued by Cohen et al. (Reference Cohen, Cavanagh, Chun and Nakayama2012), there is little evidence for this. Specifically, attention has to be able to be deployed to unconscious stimuli, otherwise how would these items reach consciousness? Having said this, attention is not all or none, and consequently, just because an item is attended does not guarantee that it will enter awareness (thus, attention is necessary but not sufficient). In addition, Cohen et al. provide a thorough review of the literature and demonstrate that previous findings reporting properties of the environment (e.g., gist of a scene; Mack & Rock Reference Mack and Rock1998) entering awareness without attention are flawed and do not hold up to dual-task attentional manipulations (Cohen et al. Reference Cohen, Alvarez and Nakayama2011). In short, there is strong evidence that top-down attention and consciousness are inherently linked. Given that top-down attention is capacity limited/effortful, this suggests that consciousness is not entirely passive and that it relies on higher-level goal-related information processing (not possessed by the creature in the cave).
Evidence from neuroscience also suggests that visual consciousness draws on high-level resources and that the neural basis for consciousness goes beyond subcortical regions. Arguably the most heavily used phenomenon for studying visual consciousness is the attentional blink (AB): This refers to subjects' typically impaired performance in reporting the second of two targets in a rapid serial visual presentation stream if it appears within 200–500 msec of the first target relative to when it appears outside this window (Raymond et al. Reference Raymond, Shapiro and Arnell1992). The AB is thought to reflect a failure of perceptual awareness, where the missed second target does not enter consciousness. Considerable functional imaging work has investigated the neural substrates of this phenomenon and localized it to frontal-parietal regions. For example, Marois et al. (Reference Marois, Yi and Chun2004) found, using a dual-target Rapid Serial Visual Presentation (RSVP) task with scene stimuli for target 2, that when this item was missed, and it failed to enter awareness, it still activated the parahippocampal place, whereas the lateral prefrontal cortex did not respond to missed second targets. Similarly, research on object substitution masking (OSM; Enns & Di Lollo Reference Enns and Di Lollo1997), where a sparse (e.g., four-dot), temporally trailing mask obscures the visibility of a briefly presented target, also implicates top-down processing. Indeed, Hirose et al. (Reference Hirose, Kihara, Mima, Ueki, Fukuyama and Osaka2007) observed that disruption of visual area MT (middle temporal/V5) via repetitive transcranial magnetic stimulation (rTMS) influenced the magnitude of OSM. This is consistent with the re-entrant processing account of OSM offered by Di Lollo et al. (Reference Di Lollo, Enns and Rensink2000) who argue that the phenomenon represents iterative hypothesis testing with anterior (e.g., frontal) regions sampling high-resolution sensory information from posterior visual areas via feedback. Finally, recent work by Zhang et al. (Reference Zhang, Jamison, Engel, He and He2011) has shown that even the inter-ocular fluctuations in consciousness that are observed during binocular rivalry (BR), which were previously thought to be automatic and spontaneous, are dependent on top-down attention. These authors employed electroencephalogram (EEG) frequency tagging via steady-state visual evoked potentials to show that stimulus representations stopped rivaling when attention was diverted away from the rivalry display. Collectively, these results using the prominent visual consciousness phenomena of the AB, OSM, and BR, all show that top-down attention processing plays a vital role in consciousness. In addition, when considering that top-down attention has been shown to tap frontal-parietal cortical circuits (see Corbetta et al. Reference Corbetta, Patel and Shulman2008), this suggests that consciousness is, at least partially, both high level and active rather than passive.
So where does this leave us with the passive frame theory offered by Morsella et al.? It is clear that theirs is an internally consistent framework that offers a unique perspective and impressive account of the most basic form of consciousness. However, the challenge of characterizing higher levels of consciousness (although AB, OSM, and BR are thought to be perceptual phenomena) in a species that leads a richer and more challenging life and that has more sophisticated and complex cognitive hardware, remains to be addressed. It may be the case that passive frame theory could be extended to accommodate a broader range of findings and phenomena in humans; however, this would require significant revision. It must also be noted that such a revision is necessary if this theory is to provide a comprehensive framework that facilitates the interpretation of consciousness data. Having said this, Morsella et al. are not alone here, and indeed the difficulty of marrying reductionism and generalizability is one that all of us working in this field face.