Mark Solms, a neuropsychologist and psychoanalyst, argues that consciousness does not originate in the cortex but in the brainstem, specifically in raw feeling or affect. He contends that the scientific study of consciousness made a foundational mistake by starting with visual perception — a cortical function that can operate unconsciously — rather than with feeling, which is intrinsically conscious and generated by upper brainstem structures. His dual-aspect monist framework treats mind and body as two observational perspectives on a single physical system: a self-organizing, homeostatic mechanism that persists by minimizing entropy through a Markov blanket that separates it from the environment. Consciousness arises when a prioritized need (affect) becomes felt, enabling voluntary action guided by valence (good/bad for the organism). This view is grounded in converging evidence from lesion studies (hydranencephalic children without cortex remain conscious; tiny brainstem lesions abolish consciousness), neuroimaging, neurotransmitter pharmacology, and dream research showing the dopaminergic seeking system generates dreaming. Solms integrates neuroscience with psychoanalysis (neuropsychoanalysis) to account for both objective mechanisms and subjective experience, and is now engineering artificial agents that implement this architecture to test whether feeling can be synthetically created.
Core Theory: Consciousness as Feeling, Not Perception
Consciousness is “what it is like to be you,” adopting Thomas Nagel’s definition: an organism is conscious iff there is something it is like for that organism.
This definition is simple, all-encompassing, and applies beyond humans; it captures the qualitative feel of being as the essence of consciousness.
The scientific program initiated by Francis Crick in the mid-1990s erred by taking visual perception as the model system for consciousness.
Visual perception is a cortical function that routinely operates unconsciously (e.g., face recognition, reading comprehension can occur without awareness), so it is not an intrinsically conscious function.
Feeling (affect) is intrinsically conscious: you cannot have a feeling without feeling it. The mechanism of feeling must explain why it is felt.
The upper brainstem — particularly the periaqueductal gray (PAG) and reticular activating system — generates raw feeling; damaging these structures extinguishes consciousness entirely.
Cortical consciousness depends on brainstem consciousness; the cortex elaborates content but does not generate the “lights on” of awareness.
The Brainstem Evidence
Hydranencephaly: children born without a cortex are not in a coma; they express a wide range of emotions (protest, giggle, rage, fear, startle) and bodily needs — they are feeling beings.
Conversely, a tiny lesion in the upper brainstem (PAG, reticular activating system) in a person with an intact cortex causes total loss of consciousness.
Electrical stimulation of these brainstem nuclei in conscious humans produces dramatic changes in raw feeling.
Positron emission tomography (PET) during intense raw feeling shows activation epicenters in the upper brainstem.
Psychopharmacology: neurotransmitters targeted by psychiatry (serotonin, dopamine, noradrenaline) are sourced in the reticular activating system; drugs act on the feeling-generating substrate.
Single-cell recordings, microdialysis (measuring dopamine release), and functional imaging all converge on the dopaminergic seeking/wanting system as the generator of dreaming and, by extension, the core conscious mechanism.
Affect, Feeling, Sentience, Consciousness: Definitions and Distinctions
Affect: the valenced registration of how far or near the organism is to its homeostatic set points (good = near, bad = far). Not all affects are conscious; many homeostatic regulations (e.g., blood pressure) operate without affect.
Feeling: the prioritized affect that comes to consciousness. Only one affective need is felt at a time; others are regulated in the background. Feeling enables voluntary action by signaling “this is going better/worse for me” so the organism can change course.
Sentience: the extension of affect onto the sensorium — “I feel like this about that.” The sensorium contextualizes affect, telling the organism what is causing the feeling or could ameliorate it. Only salient aspects of the sensorium (those maximizing opportunity to reduce the prioritized need) become conscious; most perception is unconscious.
Consciousness: the state of having a prioritized, felt affect extended onto a salient sensorium. It is the “bottleneck” where the musculoskeletal system can only execute one voluntary project at a time, so the prioritized need dominates.
Self-Organizing Systems, Homeostasis, and the Origin of Subjectivity
Consciousness arises only in self-organizing systems that act to persist (minimize entropy) against the second law of thermodynamics.
Persistence is a value: it is “good” for the system to continue, “bad” to dissipate. This value is subjective — it applies only to and for the system, not the universe.
Self-organizing systems must be sequestered from the environment by a Markov blanket (a boundary that separates internal states from external states), giving them a point of view.
The system must infer the causes of its sensory states via a generative model of the world; it acts to sample sensory data compatible with its continued existence.
Felt uncertainty — confidence or lack thereof in the generative model — is a fundamental physical mechanism for how subjective experience arises. “Palpating your uncertainty” is part of the statistical physics of consciousness.
Homeostasis is extended into affect: the organism registers deviation from set points not just reflexively but as a demand on the mind to perform work, enabling flexible responses to novel situations.
Multiple Needs, Categorical Variables, and the Ground of Qualia
Living systems have multiple, incommensurable needs (thirst, hunger, sleep, safety, etc.) that cannot be reduced to a common currency; you cannot “sleep off” thirst.
Each need is a categorical variable with its own qualitative valence — eight-out-of-ten thirst is not the same as eight-out-of-ten sleepiness. This categorical, qualitative distinction is the “ground zero of qualia.”
Needs include bodily imperatives and emotional imperatives (fear, rage, separation distress, care/nurturance, play). Emotional needs are biological, not “fluffy.”
The midbrain decision triangle (superior colliculi = salience map of opportunities; PAG = current needs; midbrain motor field = action) prioritizes one need at a time bottom-up.
Top-down executive control (prefrontal cortex) can inhibit the bottom-up priority temporarily (e.g., “I’m sleepy but I’ll keep studying”), but only up to a point; eventually the bodily need wins.
Mammalian Drives and Altruism
Mammals are born helpless; separation distress (panic/anxiety when separated from caregiver) and the caregiver’s nurturant drive (often called maternal instinct) are species-specific emotional needs.
Altruism is not uniquely human; it is built into mammalian phenotype: care for vulnerable offspring, play (which requires attending to conspecifics’ needs), and group tolerance.
Conflicts between drives (e.g., rage at a beggar vs. care for the vulnerable) and between self and conspecifics generate complex motives like guilt and shame.
Individual differences in drive strength (innate) and developmental trajectory (environment, learning, memory) explain why people respond differently to the same situation.
Psychoanalysis studies how conscious self-reports (preconscious, verbally mediated) can diverge from deeper, non-verbal motivational layers (unconscious), and how defense mechanisms veil impulses that conflict with self-image.
Neuropsychoanalysis: Bridging Objective and Subjective
Neuropsychoanalysis draws on both neuroscience (objective study of the brain organ) and psychoanalysis (subjective study of the being of the mind).
Neuroscience lacks access to feelings (cannot be seen/measured externally); psychoanalysis lacks experimental falsifiability. Each offsets the other’s weakness.
Behavior is driven by avoiding feeling bad and seeking feeling good; leaving out the subjective aspect makes brain function unintelligible.
Freud was a neuroscientist before developing psychoanalysis; his “metapsychology” sought functional mechanisms underlying mental processes, compatible with cognitive neuroscience.
Freud is the foundation of depth psychology; later figures (Jung, Klein, Winnicott) are developments on that foundation. Starting with the elemental form is methodologically necessary.
Dreams: The Seeking/Wanting System
Solms’s doctoral research (1980s–90s) studied brain-lesion patients to locate the dream generator.
Damage to the REM-generating brainstem abolished REM sleep but not dreaming; conversely, damage to the dopaminergic reward/seeking system abolished dreaming but spared REM sleep.
The seeking system (Panksepp), wanting system (Berridge), or brain reward system (mainstream) is the same dopaminergic motivational circuitry.
This converging evidence (lesions, PET, microdialysis, single-unit recordings) falsifies the “dreams are random cortical activation by REM” view and supports Freud’s wish-fulfillment theory: dreams are generated by a motivational system.
Neuropsychoanalysis links the subjective meaning of dreams to their objective neural mechanism.
The Hard Problem and Why Starting with Perception Was a Mistake
The hard problem (Chalmers): explaining how/why a physical mechanism gives rise to experience. It is real, well-posed, and soluble — not a pseudo-problem.
Starting with visual perception created the hard problem because perception is not intrinsically conscious; the mechanism of seeing doesn’t entail feeling.
Starting with feeling avoids the hard problem: the function of feeling is to be felt. A mechanistic account of feeling (valenced, categorical, prioritized, extended onto sensorium) explains its phenomenology.
Mary the affective neuroscientist (vs. visual neuroscientist) would know what feelings are like from a complete mechanistic description: valenced, subjective, qualitatively variable.
The “bottleneck” of voluntary action (musculoskeletal system can do one thing at a time) explains why only the prioritized need is felt: feeling guides choice by signaling progress toward/away from the set point.
Dual-Aspect Monism and Physicalism
Physicalism ≠ materialism. Physics explains matter, but matter is not fundamental (E=mc², fields, information). Physics must account for experience as a natural phenomenon.
Dual-aspect monism: mind (first-person experience) and body (third-person observation) are two perspectives on one physical system — a functional, self-organizing mechanism governed by a single set of laws.
The system is autopoietic, predictive, Markov-blanketed, and seeks to minimize variational free energy (free energy principle / active inference).
The subjective aspect arises from the system’s intrinsic perspective: it has a point of view, values, and a generative model it must maintain confidence in.
Markov Blankets, Nested Hierarchies, and Felt Uncertainty
Markov blankets are nested: cells, organs, organ systems, the whole organism each have blankets. Survival is delegated upward; the organism-level blanket is the one that matters for the conscious agent.
Malignancy = cells going rogue, acting independently of the organism-level blanket.
The generative model is model-relative (coarse-graining matters), but the organism’s blanket is the relevant one for its consciousness.
Felt uncertainty — the system’s confidence in its model — is a key physical mechanism for consciousness. The system “palpates” its uncertainty to guide action and model updating.
Artificial Consciousness Engineering
Solms’s current forefront project: engineer an artificial agent that has feelings, implementing the formalism developed with Karl Friston (2018 paper reducing extended homeostasis to a formalism).
Feynman’s criterion: “If I can’t create it, I don’t understand it.” Synthetic feeling is the litmus test for whether the causal mechanism is correctly identified.
Success would demonstrate that the brainstem-based, affect-first architecture is sufficient for consciousness.
Corticocentrism as Main Opposition
Primary opponents: corticocentrist theorists — the majority of neuroscientists who assume the cortex is the seat of consciousness.
This view persists despite evidence; progress depends on “the death of the old guard” (Planck).
Within the free-energy/active-inference community, disagreements are intramural; outside it, they are allies against broader materialist/functionalist assumptions.
Empirical Adjudication Methods
Consciousness is first-person, but science uses multiple third-person methods to triangulate:
Converging lines make the brainstem-affect theory unreasonable to doubt, even if absolute proof is unattainable (Popper).
Personal Motivation: Death Anxiety and the Search for Meaning
Brother’s brain injury in childhood revealed that “we are our brains” and brains are mortal → nihilistic despair in youth.
The only worthwhile response: understand what existence is. A glimmer of hope that understanding might free one from mortality’s horror.
Theory confirms: when the body dies, the Markov blanket dissipates; the second law wins. Life postpones oblivion but ultimately fails.
No change in belief about personal extinction, but gratification in having made progress on the mechanism of being.
Buddhism, Nirvana, and the Limits of Model Change
Changing the generative model (e.g., Buddhist practice) can alter the self-model but cannot change the prior preferences (phenotypic imperatives) that ground the model.
Even the Buddha had to defecate; embodied creatures cannot transcend need.
