As a functional neuroanatomist, I applaud Kurzban et al.'s efforts to integrate parallel literatures, despite the jargon barriers (oh, yes, phenomenology = feelings!). Of course, it's difficult even within our own fields. I'm certain that psychologists who adhere to “self-perception theory” (see Laird Reference Laird2007) would have difficulty with the statement that “feelings motivate behavior,” as do I.

I mapped an ascending neural pathway for affective feelings from the body and identified the cortical image of these elemental feelings in an interoceptive representation of the physiological condition of the body, which anchors the insular cortex of primates (Craig Reference Craig2002). A convergence of evidence across disparate fields led me to suggest that the progression of interoceptive processing from posterior to anterior in the insula culminates in a moment-to-moment global representation of homeostatic salience in the anterior insula, which we experience as feelings. The model incorporates each finite momentary representation in a cinemascopic structure across present time built upon learned predictions. The comparators needed to use this structure for the optimization of moment-to-moment emotional behavior also substantialize reflective subjectivity (as a “spandrel”), which is always one moment behind and cannot “see” itself. It may or may not be veridical, and it can become functionally distorted (Craig Reference Craig2009; Reference Craig2010).

The key concept is that feelings are an amodal valuation of energy utilization constructed on the homeostatic sensori-motor template generated by the elemental interoceptive feelings from the body. This construct, which we experience as feelings, can represent any pattern of neural activity in the brain; for example, a neural activation pattern in a sensory system, in the cognitive network, or in one of the motor planning networks. Being able to feel every percept, thought, motivation, or potential behavior provides a powerful “common currency” for evaluation and comparison of energy costs and benefits.

In my view, the homeostatic principle of optimal energy utilization in an organism in which the brain consumes 25% of the entire energy budget compelled the evolutionary development of a common valuation system for energy cost-benefit analysis. That system had to be capable of representing not only the homeostatic significance of any behavior, but also the homeostatic significance of the potential behaviors inherent in any network pattern of neural activity (Craig Reference Craig2011). I believe that feelings are the coinage of that valuation system, the interoceptive indicators of actual or potential homeostatic energy significance. Very recent evidence confirms that the bilateral insular cortex encodes action-specific valuation (Fitzgerald et al. Reference Fitzgerald, Friston and Dolan2012). I believe the precise control of emotional behavior is effected by the opponent processes of energy nourishment (positive affect, approach motivation) and expenditure (negative affect, avoidance motivation) that are engendered in the bivalent, bicameral brain of all vertebrates (Craig Reference Craig2009; MacNeilage et al. Reference MacNeilage, Rogers and Vallortigara2009).

It is important to recall that emotional behaviors evolved as energy-efficient means of producing goal-directed actions that fulfill homeostatic and social needs, as Darwin and others taught. In other words, the emotional feelings of happiness, anger, and the like, can be viewed as the sensory complement of fundamental neural patterns that are genetically ingrained in our (limbic) emotional motor system. Like the pleasantness or unpleasantness of a cool stimulus on the hand, which is inverted in hyperthermia and hypothermia, an emotional feeling is the perceptual correlate of a behavioral motivation. Sometimes we are aware of the feeling that accompanies each motivated behavior, and sometimes we are not (Wegner Reference Wegner2002). All behavior is emotional (Dunn et al. Reference Dunn, Evans, Makarova, White and Clark2012; Montague Reference Montague2006a).

In this model, the limbic (emotional) motor cortex (the anterior cingulate and medial prefrontal cortex) works together with the limbic sensory cortex (anterior insula); that's why they are almost always co-active, and why they are the most commonly activated regions of the brain (Nelson et al. Reference Nelson, Dosenbach, Cohen, Wheeler, Schlaggar and Petersen2010; Sepulcre et al. Reference Sepulcre, Sabuncu, Yeo, Liu and Johnson2012). The anterior insula is involved in thoughts, emotions, and bodily feelings, and it is the foundation for emotional awareness (Oosterwijk et al. Reference Oosterwijk, Lindquist, Anderson, Dautoff, Moriguchi and Barrett2012; Zaki et al. Reference Zaki, Davis and Ochsner2012). It is also the bottleneck in the attentional blink, the basis for inspection time, and a basic resource that underlies fluid intelligence (Britz et al. Reference Britz, Van De Ville and Michel2010; Cole et al. Reference Cole, Yarkoni, Repovs, Anticevic and Braver2012; Craig Reference Craig2009; Menon & Uddin Reference Menon and Uddin2010).

The homeostatic (Group III/IV) sensory input from muscles to the insular cortex is a primary generator of feelings of “fatigue” and “effort” (Amann et al. Reference Amann, Blain, Proctor, Sebranek, Pegelow and Dempsey2010; Hilty et al. Reference Hilty, Jancke, Luechinger, Boutellier and Lutz2011). Individuals who are more aware of their homeostatic condition self-regulate their energy consumption more efficiently, and highly trained athletes and warriors use interoceptive sensory activity to produce optimal performance (Herbert et al. Reference Herbert, Ulbrich and Schandry2007; Paulus et al. Reference Paulus, Simmons, Fitzpatrick, Potterat, Van Orden, Bauman and Swain2010; Reference Paulus, Flagan, Simmons, Gillis, Kotturi, Thom, Johnson, Van Orden, Davenport and Swain2012).

And lastly, the anterior insula serves as the evidence accumulator for decision-making, the source of error negativity, and in my opinion, the basis for mental energy and effort (Gluth et al. Reference Gluth, Rieskamp and Buchel2012; Jansma et al. Reference Jansma, Ramsey, de Zwart, van Gelderen and Duyn2007; Wessel et al. Reference Wessel, Danielmeier, Morton and Ullsperger2012). Unfortunately, its activation is not registered in certain studies, probably because of its structural variability and the use of subtractive contrasts.

I enjoyed reading the authors' views of the literature they are familiar with. I hope they will enjoy reading another literature, and I hope they find these comments constructive. Our respective views are conceptually distinct, yet quite similar teleologically. I would very much enjoy an opportunity to discuss these ideas further with such knowledgeable scholars.