Coombs and Trestman’s approach determined that “cognition” evolved in three very different phyla. They admit the term, as used, is “very fuzzy,” denoting “pivotal traits” such as problem-solving, theory-of-mind, planning, and tool use. We note that cognition in this sense also applies to machines. A self-driving automobile distinguishes other cars and obstacles from pedestrians, for example, and computes adaptive steering accordingly. However, it is not conscious. Here we consider a closely related endpoint of Darwinian evolution, barely mentioned by the authors: conscious awareness, conscious “experience,” or simply “consciousness”. We ask: “When did consciousness, the “me inside” arise in Darwinian evolution?”

We posit that consciousness may not be a late-emerging feature of large-brained vertebrates and perhaps octopuses. Rather, it evolved early as an adaptive response to life-threatening events signaled by specific sensory stimuli. Primitive nocifensive behaviors are present in protozoa and single-celled eukaryotes, which avoid/escape from regions of excess salt, pH, and/or temperature using stimulus-sensitive receptors linked to cilia and flagella. Plants have negative tropisms, and vertebrates have nocifensive reflexes. However, such responses do not necessarily denote pain. Spinal reflexes can be triggered in human paraplegics, but evoke no pain. Adaptive responses to injury, and avoiding injury, may be intelligent, “cognitive” in the sense used by the authors. But they do not reliably indicate pain experience. Pain is experienced by a conscious brain.

We are unaware of any mechanistic theory of how consciousness is implemented in the brain beyond the proposal that it “emerges” when certain conditions are met. Nor is there a definitive way to detect consciousness in another organism, or an intelligent silicon-based device. Nevertheless we, like you the reader, experience conscious perception continuously, except during anesthesia or dreamless sleep. The “me inside” sees, hears, smells, and understands. Note that the integral of all sensory experience and thought is not what makes us conscious. When closing my eyes, the “me inside” remains 100% present. Even covering my eyes, ears, and nose, all together, does not diminish my conscious presence, even slightly. Try it!

We do not dispute that high-order sensory perception depends on complex computations executed in the neocortex. However, we propose that phenomenal consciousness, the “me inside,” does not. Rather, the experience of pain is rooted subcortically, plausibly in the “primitive” brainstem. A corollary is that pain’s counterpart, consciousness, also originated subcortically. Laminated cortex evolved later, not to generate consciousness, but to provide the computational power needed for already conscious animals to prosper (Baron & Devor, Reference Baron and Devor2022; Kaiser, Reference Kaiser2007).

Many clues point to pain being computed subcortically. One is direct brain stimulation in awake humans. Across modalities, cortical stimulation evokes sensory experiences: vision in (V1), audition (A1), touch (S1), etc. The one exception is pain. There is no recognized P1 (primary pain cortex). Wilder Penfield wrote: “…only very rarely does a patient use the word pain to describe the result of cortical stimulation….,” continuing: “It is obvious, therefore, that the pathway of pain conduction reaches the thalamus and consciousness without essential conduction to the cortex” (Libet, Reference Libet and Iggo1973; Penfield, Reference Penfield1947). Some buried cortices are inaccessible, but seizures access them all. Yet, pain, unlike smells or sounds, is rarely reported as an aura in epileptic patients (Kuloğlu Pazarci et al., Reference Kuloğlu Pazarci, Bebek, Baykan, Gürses and Gökyiğit2016; Nair et al., Reference Nair, Najm, Bulacio, Lüders and Luders2001; Young & Blume, Reference Young and Blume1983). The insular cortex may be an exception. But insular stimulation using depth electrodes yielded pain in only ∼1.4% of trials (Afif et al., Reference Afif, Hoffmann, Minotti, Benabid and Kahane2008; Isnard et al., Reference Isnard, Magnin, Jung, Mauguière and Garcia-Larrea2011). In marked contrast, stimulating subcortical locations along ascending pain pathways, e.g., parabrachial complex and somatosensory thalamus, readily evokes pain (Davis et al., Reference Davis, Kiss, Luo, Tasker, Lozano and Dostrovsky1998; Nashold et al., Reference Nashold, Wilson and Slaughter1969). To be sure, neurons in S1, S2, insula, and elsewhere in Melzack’s “pain matrix” are activated by painful stimuli. But this probably reflects generalized arousal, not pain (Iannetti & Mouraux, Reference Iannetti and Mouraux2010; Mouraux et al., Reference Mouraux, Diukova, Lee, Wise and Iannetti2011; Vierck et al., Reference Vierck, Whitsel, Favorov, Brown and Tommerdahl2013).

Lesions support this conclusion. Again, quoting Penfield “…no removal of cortex anywhere can prevent pain from being felt…”. If anything, cortical lesions augment pain [post-stroke pain; (Andersen et al., Reference Andersen, Vestergaard, Ingeman-Nielsen and Jensen1995)]. In rodents, decortication does not reduce nociception, cause loss-of-consciousness (LOC), or reduce MAC for systemic anesthesia and likewise for cortical agenesis in humans (Merker, Reference Merker2007; Panksepp et al., Reference Panksepp, Normansell, Cox and Siviy1994; Rampil et al., Reference Rampil, Mason and Singh1993). V1 lesions render patients perceptually blind, but no less conscious; thousands have undergone prefrontal lobotomy for psychiatric diagnoses without LOC, and large cortical strokes rarely induce LOC without added sub-tentorial involvement. The consensus region for lesion-induced coma is the mesopontine brainstem (Fischer et al., Reference Fischer, Boes, Demertzi, Evrard, Laureys, Edlow, Liu, Saper, Pascual-Leone, Fox and Geerling2016; Grady et al., Reference Grady, Boes and Geerling2022; Posner et al., Reference Posner, Saper, Schiff and Plum2007). In rodents, the mesopontine tegmental anesthesia area is a unique structure at which microinjection of anesthetics produces LOC, including dense analgesia [(Avigdor et al., Reference Avigdor, Minert, Baron and Devor2021; Baron & Devor, Reference Baron and Devor2023; Devor & Zalkind, Reference Devor and Zalkind2001; Minert et al., Reference Minert, Yatziv and Devor2017)].

These observations indicate profound differences in the cortical contribution to pain experience versus vision, audition, non-painful somato-sensation, etc. Our interpretation: For a prey animal to distinguish a cloud from a rhinoceros charging requires the computational power of a cortical supercomputer. In contrast, information about whether a stimulus on the snout is a gentle lick, or a predator’s bite, is available at the first central synapse in the trigeminal brainstem. The lick is signaled by cutaneous low-threshold touch afferents; the bite by nociceptors (Figure 1). We assume, without elaborating why, that conscious pain experience has profound survival value over-and-above simply differentiating a lick from a bite… perhaps in the realms of salience or learning. Otherwise, consciousness would not have evolved in the first place. Hence, we expect that evolution of a neural substrate for implementing consciousness, the “me inside,” would have emerged first in the brainstem, concerning injury (perhaps also food-related chemical signaling). Eons later, with a cortical supercomputer, it would make Darwinian sense for the survival valence of more subtle distinctions, in all sensory modalities, to be conveyed to an existing “me inside” via descending cortico-bulbar pathways. Species that instead chose to devote additional eons to generating a new, and potentially competing, “me inside” may not have survived to tell the story.

Figure 1.

Depending on the stimulus, the lick of a fawn, or the bite of a wolf, somatosensory activity is generated in low-threshold touch fibers, or in high-threshold nociceptors. The corresponding information content, and its survival valence, is available at the 1st central synapse, in this case in the trigeminal brainstem. We propose that the neural substrate of conscious experience emerged early, in the brainstem (orange patch), in association with noxious stimuli. In contrast, extracting information and survival valence from visual contrasts falling on the retina is far more complex. Is it a distant cloud, or a rhinoceros charging? Such distinctions require supercomputer computations, available only eons later with the development of laminated neocortex. We propose that in humans, the output of primary and higher-order visual processing is delivered to the primeval neural substrates of conscious experience, long-since present in the brainstem (Baron & Devor, Reference Baron and Devor2022).