The article by Morsella et al. is an interesting and thought-provoking review. Two primary issues kept coming to mind during my reading, both of which seem critical to address. The first is that the review avoids mentioning anything about the relationship between attention and consciousness. The second concerns whether the olfactory system in all of its simplicity is suitable as a model system to study consciousness for action (particularly in humans).
Cited in the target article but not elaborated, Keller (Reference Keller2011) emphasizes the relation between attention and consciousness when considering olfactory consciousness. In a similar vein, work in my lab has considered the link between consciousness and attention; however, the focus is not on olfaction, but on conflict situations between the limbs (bimanual actions) in an allocation of attention to action (AAA) framework (Franz Reference Franz2012). As in the target article by Morsella et al., the AAA framework was inspired by the earlier writings of Sperry (Reference Sperry1952) and James (Reference James1890/1918), and a major tenet is ideomotor theory (Hommel Reference Hommel and Bruya2010). Thus, the similarities across approaches are quite remarkable.
Higher levels of the 3-level AAA framework incorporate examples of bimanual conflict in actions such as drawing a circle and a line, or circle and square (Franz Reference Franz1997; Reference Franz and Johnston-Frey2003; Reference Franz2012; Franz & Ramachandran Reference Franz and Ramachandran1998; Franz et al. Reference Franz, Zelaznik and McCabe1991; Reference Franz, Eliassen, Ivry and Gazzaniga1996), or reaching with one hand to a far target and one hand to a near target (Franz & McCormick Reference Franz and McCormick2010; Kelso et al. Reference Kelso, Southard and Goodman1979). The “empirical proof” of integration (of lower and higher levels of processing) comes from additional examples in which the two hands produce a unified action rather than conflicting actions as a result of the presence of visual perceptual features or specific verbal instructional commands (Franz & McCormick Reference Franz and McCormick2010; Franz et al. Reference Franz, Waldie and Smith2000; Reference Franz, Zelaznik, Swinnen and Walter2001); this propensity to unify information at the highest level possible is proposed as a fundamental property of the “economical brain” (Franz Reference Franz2010; Reference Franz2012; Franz & McCormick Reference Franz and McCormick2010). Higher levels are further discussed in terms of automatic or conscious processes of activation and inhibition (level 2: AAA), and action selection (level 3: AAA), with all levels implicating a mix of conscious and unconscious processing.
Consider a specific point of overlap found in the target article, which states that “there must be a (unconscious) mechanism by which one action plan can influence behavior more than other activated action plans” (sect. 3.3, para. 3). AAA argues the same, building on an action selection model (Mink Reference Mink1996) in which activation of an action plan occurs from among other prepotent action plans that must be inhibited (akin to action options). This process implicates basal-ganglia (striatal) and thalamic-cortical (primarily frontal) circuits as being key (Franz Reference Franz and Bezard2006; Reference Franz2012; Mink Reference Mink1996) and can occur automatically through activation functions.
The target article focuses primarily on what most resembles the lowest level of AAA. In level 1 of AAA (see Figure 1 in Franz Reference Franz2012), primarily exogenous inputs capture awareness of an organism transiently (and automatically), leading eventually to trial-and-error learning which results in formation of a perceptual-action code akin to ideomotor theory (Hommel Reference Hommel and Bruya2010); recent experimental support comes from studies using evoked response potentials (Bednark et al. Reference Bednark, Reynolds, Stafford, Redgrave and Franz2013; Bednark & Franz Reference Bednark and Franz2014).
Three differences are clear across frameworks: (1) the focus on attention and its relation to consciousness (Franz Reference Franz2012; Keller Reference Keller2011) not present in Morsella et al.; (2) the focus on “in the service of action” (Franz Reference Franz2012; Morsella et al.) not present in Keller (Reference Keller2011); and (3) the focus on olfaction (Keller Reference Keller2011; Morsella et al.) not present in Franz (Reference Franz2012).
Furthermore, the work of Grossberg, although briefly cited, seems far more relevant than suggested in the target article. Grossberg (Reference Grossberg1999) addresses the question of “how an intelligent being can autonomously adapt to a changing world” (p. 4, emphasis in original). Grossberg (Reference Grossberg1999) suggests that top-down expectation selectively amplifies some features of bottom-up input while suppressing others, and bottom-up signals reactivate those same top-down expectations (p. 10). This reactivation leads to a resonant state which is necessary for the brain to achieve consciousness (a conscious percept). Without it, there is no conscious percept. Furthermore, Grossberg (Reference Grossberg1999) suggests that a form of automatic attention can operate at very low levels of processing (p. 32). Perhaps this also suggests that the distinction between “passive” and “active” becomes rather blurry within the context of a process model?
According to AAA, increasingly embedded levels of processing are built upon a basic sensorimotor system as a result of evolved adaptations. This enables representation (possibly similar to conscious content), and re-representation in a hierarchical neural system, which has evolved through embodied actions, consistent with the very early views of John Hughlings Jackson (Reference Hughlings Jackson1884; cf. Franz & Gillett Reference Franz and Gillett2011; Gillett & Franz Reference Gillett and Franz2013). A statement in section 5 of the target article seems to capture the essence of this idea: “All influences on skeletomotor behavior, from the highest-level processes (e.g., language) to the lowest-level processes (e.g., pain withdrawal), must engage the skeletomotor output system” (para. 4).
If olfaction is being proposed as a model system for studying consciousness, and consciousness is for action, then what about the implied olfaction for action link? The skeletal action system is highly adaptive with growth of the organism (the limbs increase in size with age), and is controlled by numerous cognitive and perceptual processes together with influences of a dynamically changing environment (Franz Reference Franz2012; Grossberg Reference Grossberg1999). Limb systems respond with fast simple reaction times to visual and auditory stimuli (often between 200 and 300 milliseconds; Greenwald Reference Greenwald1972; Hughes & Franz Reference Hughes and Franz2007; Shen & Franz Reference Shen and Franz2005; Tassinari & Berlucchi Reference Tassinari and Berlucchi1993). Visual stimuli are experienced as uninterrupted, whereas olfactory experience is discontinuous due to sniffs occurring with gaps between them (1.6 seconds is cited as the gap; Laing Reference Laing1983; Mainland & Sobel Reference Mainland and Sobel2006; cited in Keller Reference Keller2011). Thus, olfactory experience is temporally unstructured, and olfaction is essentially a slower sense (Keller Reference Keller2011). Gottfried and Dolan (Reference Gottfried and Dolan2003) report a mean olfactory reaction time of 1.539 seconds, and further elaborate that “the nose smells what the eye sees” (p. 375), that is, odor perception is largely affected by interactions with vision. Is an olfaction to action link all that consistent with available evidence?
In sum, while I have no reason to disagree with the suggestion that olfaction is an ideal system to study consciousness, I am less convinced that olfaction is a suitable model system for an approach aimed at informing our understanding of “consciousness for action.”
You have
Access
The article by Morsella et al. is an interesting and thought-provoking review. Two primary issues kept coming to mind during my reading, both of which seem critical to address. The first is that the review avoids mentioning anything about the relationship between attention and consciousness. The second concerns whether the olfactory system in all of its simplicity is suitable as a model system to study consciousness for action (particularly in humans).
Cited in the target article but not elaborated, Keller (Reference Keller2011) emphasizes the relation between attention and consciousness when considering olfactory consciousness. In a similar vein, work in my lab has considered the link between consciousness and attention; however, the focus is not on olfaction, but on conflict situations between the limbs (bimanual actions) in an allocation of attention to action (AAA) framework (Franz Reference Franz2012). As in the target article by Morsella et al., the AAA framework was inspired by the earlier writings of Sperry (Reference Sperry1952) and James (Reference James1890/1918), and a major tenet is ideomotor theory (Hommel Reference Hommel and Bruya2010). Thus, the similarities across approaches are quite remarkable.
Higher levels of the 3-level AAA framework incorporate examples of bimanual conflict in actions such as drawing a circle and a line, or circle and square (Franz Reference Franz1997; Reference Franz and Johnston-Frey2003; Reference Franz2012; Franz & Ramachandran Reference Franz and Ramachandran1998; Franz et al. Reference Franz, Zelaznik and McCabe1991; Reference Franz, Eliassen, Ivry and Gazzaniga1996), or reaching with one hand to a far target and one hand to a near target (Franz & McCormick Reference Franz and McCormick2010; Kelso et al. Reference Kelso, Southard and Goodman1979). The “empirical proof” of integration (of lower and higher levels of processing) comes from additional examples in which the two hands produce a unified action rather than conflicting actions as a result of the presence of visual perceptual features or specific verbal instructional commands (Franz & McCormick Reference Franz and McCormick2010; Franz et al. Reference Franz, Waldie and Smith2000; Reference Franz, Zelaznik, Swinnen and Walter2001); this propensity to unify information at the highest level possible is proposed as a fundamental property of the “economical brain” (Franz Reference Franz2010; Reference Franz2012; Franz & McCormick Reference Franz and McCormick2010). Higher levels are further discussed in terms of automatic or conscious processes of activation and inhibition (level 2: AAA), and action selection (level 3: AAA), with all levels implicating a mix of conscious and unconscious processing.
Consider a specific point of overlap found in the target article, which states that “there must be a (unconscious) mechanism by which one action plan can influence behavior more than other activated action plans” (sect. 3.3, para. 3). AAA argues the same, building on an action selection model (Mink Reference Mink1996) in which activation of an action plan occurs from among other prepotent action plans that must be inhibited (akin to action options). This process implicates basal-ganglia (striatal) and thalamic-cortical (primarily frontal) circuits as being key (Franz Reference Franz and Bezard2006; Reference Franz2012; Mink Reference Mink1996) and can occur automatically through activation functions.
The target article focuses primarily on what most resembles the lowest level of AAA. In level 1 of AAA (see Figure 1 in Franz Reference Franz2012), primarily exogenous inputs capture awareness of an organism transiently (and automatically), leading eventually to trial-and-error learning which results in formation of a perceptual-action code akin to ideomotor theory (Hommel Reference Hommel and Bruya2010); recent experimental support comes from studies using evoked response potentials (Bednark et al. Reference Bednark, Reynolds, Stafford, Redgrave and Franz2013; Bednark & Franz Reference Bednark and Franz2014).
Three differences are clear across frameworks: (1) the focus on attention and its relation to consciousness (Franz Reference Franz2012; Keller Reference Keller2011) not present in Morsella et al.; (2) the focus on “in the service of action” (Franz Reference Franz2012; Morsella et al.) not present in Keller (Reference Keller2011); and (3) the focus on olfaction (Keller Reference Keller2011; Morsella et al.) not present in Franz (Reference Franz2012).
Furthermore, the work of Grossberg, although briefly cited, seems far more relevant than suggested in the target article. Grossberg (Reference Grossberg1999) addresses the question of “how an intelligent being can autonomously adapt to a changing world” (p. 4, emphasis in original). Grossberg (Reference Grossberg1999) suggests that top-down expectation selectively amplifies some features of bottom-up input while suppressing others, and bottom-up signals reactivate those same top-down expectations (p. 10). This reactivation leads to a resonant state which is necessary for the brain to achieve consciousness (a conscious percept). Without it, there is no conscious percept. Furthermore, Grossberg (Reference Grossberg1999) suggests that a form of automatic attention can operate at very low levels of processing (p. 32). Perhaps this also suggests that the distinction between “passive” and “active” becomes rather blurry within the context of a process model?
According to AAA, increasingly embedded levels of processing are built upon a basic sensorimotor system as a result of evolved adaptations. This enables representation (possibly similar to conscious content), and re-representation in a hierarchical neural system, which has evolved through embodied actions, consistent with the very early views of John Hughlings Jackson (Reference Hughlings Jackson1884; cf. Franz & Gillett Reference Franz and Gillett2011; Gillett & Franz Reference Gillett and Franz2013). A statement in section 5 of the target article seems to capture the essence of this idea: “All influences on skeletomotor behavior, from the highest-level processes (e.g., language) to the lowest-level processes (e.g., pain withdrawal), must engage the skeletomotor output system” (para. 4).
If olfaction is being proposed as a model system for studying consciousness, and consciousness is for action, then what about the implied olfaction for action link? The skeletal action system is highly adaptive with growth of the organism (the limbs increase in size with age), and is controlled by numerous cognitive and perceptual processes together with influences of a dynamically changing environment (Franz Reference Franz2012; Grossberg Reference Grossberg1999). Limb systems respond with fast simple reaction times to visual and auditory stimuli (often between 200 and 300 milliseconds; Greenwald Reference Greenwald1972; Hughes & Franz Reference Hughes and Franz2007; Shen & Franz Reference Shen and Franz2005; Tassinari & Berlucchi Reference Tassinari and Berlucchi1993). Visual stimuli are experienced as uninterrupted, whereas olfactory experience is discontinuous due to sniffs occurring with gaps between them (1.6 seconds is cited as the gap; Laing Reference Laing1983; Mainland & Sobel Reference Mainland and Sobel2006; cited in Keller Reference Keller2011). Thus, olfactory experience is temporally unstructured, and olfaction is essentially a slower sense (Keller Reference Keller2011). Gottfried and Dolan (Reference Gottfried and Dolan2003) report a mean olfactory reaction time of 1.539 seconds, and further elaborate that “the nose smells what the eye sees” (p. 375), that is, odor perception is largely affected by interactions with vision. Is an olfaction to action link all that consistent with available evidence?
In sum, while I have no reason to disagree with the suggestion that olfaction is an ideal system to study consciousness, I am less convinced that olfaction is a suitable model system for an approach aimed at informing our understanding of “consciousness for action.”