Consciousness: How the Brain Lights Up "I Am Here"
2026.06.13 · BigCat's Inner World
Vast amounts of processing run in parallel outside your awareness — so why does only a sliver get "lit up" into experience? Is consciousness a global broadcast, or a degree of information integration? When meditation thins out the "self," what happens in the brain? And — does someone who never wakes yet can answer questions with brain waves count as conscious?
Global Workspace: Consciousness as a "Whole-Brain Broadcast"Global Workspace Theory
Cognitive Neuroscience · Function of Consciousness
Core Insight
Consciousness is not a mysterious light "produced" by some brain region. The vast majority of processing runs subliminally and in parallel — recognizing faces, parsing grammar, regulating heartbeat — entirely unnoticed by you. Only the moment a sliver of information is broadcast to the whole brain does it become "I'm aware of it." Consciousness is a narrow-bandwidth public stage in the brain.
Mechanism
Bernard Baars proposed the "theater metaphor" in 1988: whatever the spotlight hits is broadcast to a vast audience of unconscious modules. Stanislas Dehaene and Jean-Pierre Changeux grounded it in neurons as the Global Neuronal Workspace. The key mechanism is "ignition" — when a stimulus crosses threshold, the prefrontal-parietal network erupts in a nonlinear, all-or-none amplification, marked by a signature P3b brain wave around 300 ms. Subliminal stimuli stir only local ripples and never ignite; consciousness is therefore a string of discrete ignitions, not a continuous stream.
Same Stimulus: Not Ignited vs Ignited
Subliminal · Not ignitedLocal, brief activity; can influence behavior (priming), yet you report "I saw nothing"
SelfAttention is the gate of ignition; only one thing can take the stage at a time. "Multitasking" is really rapid stage-switching, and every switch has a cost.
TeamInformation only counts as "known" by the organization once it's broadcast. An insight stuck in one head or one chat thread is as good as never ignited — use docs and sync to push it onto the global stage.
ParentingA child "not hearing you" often isn't defiance — your words didn't cross the ignition threshold. Capture attention first, then deliver content.
Relationship"You weren't even listening" may be literally true: the information stalled below threshold. Asking your partner to paraphrase is a simple test of whether it ignited.
Common Myth: treating consciousness as a "soul" or a continuous inner movie. The evidence better supports this: consciousness is a tiny-bandwidth, discretely refreshed global broadcast, with a vast unconscious doing the actual work behind it.
Key References · Bernard Baars《A Cognitive Theory of Consciousness》(1988) · Stanislas Dehaene《Consciousness and the Brain》(2014) · Dehaene & Changeux《Experimental and Theoretical Approaches to Conscious Processing》(2011, Neuron)
English Insight: "Consciousness is global information broadcasting within the cortex." — Stanislas Dehaene. It's not about "where" consciousness sits, but "who it gets broadcast to."
This Week's Practice Recall a time you drove a familiar route home yet "remember nothing of the trip" — that stretch was subliminal autopilot. Ask yourself: how much of today did I live "ignited," and how much slid past on subliminal inertia? Reflection: If such narrow bandwidth is a hard constraint of the brain, what resource is "living in the present" essentially managing?
Integrated Information Theory: Consciousness as Degree of "Integration"Integrated Information Theory (IIT)
Theoretical Neuroscience · Nature of Consciousness
Core Insight
Global Workspace asks "how does consciousness operate." Giulio Tononi flips it to a fiercer question: what kind of physical system would "have experience" at all? His answer — information that is both highly integrated (an indivisible whole) and highly differentiated (able to distinguish vast states), captured by a quantity called Φ (phi). The higher the Φ, the more consciousness.
Mechanism
IIT starts from five axioms about "experience itself" and reasons back to the conditions a physical system must meet. One counterintuitive yet elegant prediction: the cerebellum has about 4× the neurons of the cortex, yet contributes almost nothing to consciousness. Because the cerebellum works like many parallel, non-communicating little modules (feedforward, divisible — low Φ); the cortico-thalamic system is densely recurrent, where pulling one thread moves the whole (high Φ). Damaging the cerebellum won't cost you consciousness; damaging a small patch of thalamus can cause coma — quantity isn't the key, the structure of integration is.
Neuron Count ≠ Consciousness: Cerebellum vs Cortex
Cerebellum · neurons
~69 billion (feedforward, modular)
Cerebellum · to consciousness
≈ low Φ
Cortex · neurons
~16 billion (densely recurrent)
Cortex · to consciousness
≈ high Φ
Self-Application
SelfIn deep sleep and anesthesia the cortex stays active, but the "conversation" between regions breaks down and Φ collapses — the physical correlate of your "not being there." Consciousness isn't a switch, it's a continuous degree of integration.
AI CollaborationOn a strict reading of IIT, today's large models are feedforward, Φ≈0, hence "no experience" — but this is exactly IIT's most contested point, so don't cite it as settled fact.
TeamOne transferable intuition: integration > headcount. Twenty people who don't talk to each other may have lower collective "intelligence" than six in high-bandwidth collaboration. Structure determines integration; the feeling of insight may just be a local jump in Φ.
Presenting the controversy honestly: IIT is hugely influential but far from settled. In 2023, over a hundred scholars co-signed an open letter arguing that its panpsychist leanings (even simple systems might have faint experience) and its unfalsifiability make it "problematic as science," sparking fierce debate. Treat it as a powerful, testable hypothesis, not an established fact.
Key References · Giulio Tononi《An Information Integration Theory of Consciousness》(2004) · Tononi & Koch《Consciousness: Here, There and Everywhere?》(2015) · the 2023 "IIT as pseudoscience" open letter and the debate that followed
English Insight: "Consciousness is integrated information." — Giulio Tononi. It's not how much information there is, but how "indivisible" it is.
This Week's Practice Recall a time under anesthesia or deep sleep — that stretch of "time vanishing" wasn't darkness; there wasn't even the experience of "darkness." Let it be a mirror: the sense of "being present" you take for granted while awake is an achievement the brain rebuilds every second, not a default state.
The Default Mode Network: the "I" Is Continually ConstructedDefault Mode Network & the Narrative Self
Neuroscience · The Self · Cross-disciplinary
Core Insight
When you stop a task and let your mind drift, the brain doesn't rest — a specific set of regions instead lights up. Marcus Raichle named it the Default Mode Network in 2001. Its main job is weaving that "autobiographical me": recalling the past, imagining the future, guessing how others see me. In other words, the "self" is more like a continuously running process than a fixed entity.
Mechanism
The DMN's core regions include the medial prefrontal cortex, posterior cingulate/precuneus, and angular gyrus, handling self-referential processing and mind-wandering. Killingsworth & Gilbert (2010) found a wandering mind is an unhappy mind — people spend 47% of waking time wandering, and it generally lowers in-the-moment happiness. Judson Brewer (2011, PNAS) found experienced meditators show markedly lower DMN activity, and can down-regulate it in real time. Robin Carhart-Harris observed that psilocybin disintegrates the DMN's integration, with the degree closely tracking participants' reported "ego dissolution."
Cross-disciplinary echo (real and tight): DMN ↔ mind-wandering ↔ meditation ↔ the Buddhist not-self (anattā). Buddhism says the "I" has no fixed essence — it's a process arising dependently from the five aggregates; neuroscience finds the "sense of self" corresponds to a network whose activity can be turned down by meditation and temporarily dissolved by drugs. The two align, rarely, on "self as construction, not entity" — not a forced analogy, but a genuine isomorphism at the descriptive level.
Self-Application
SelfMind-wandering ≠ rest. Ruminative wandering is the DMN burning energy idling. When you catch yourself "holding a meeting in your head," pulling attention back to the breath or present senses is hitting the brakes on the DMN.
ParentingSome emptiness and boredom matters — the DMN consolidates memory and feeds imagination there. Packing every minute of a child's day leaves this network no time to work.
Relationship"They must look down on me" is usually a script the DMN wrote, not a fact. Separating "the narrative about me" from "what actually happened" dismantles a lot of interpersonal anxiety.
Common Myth: "The DMN is bad and must be eliminated." Not so. It underpins autobiographical memory, empathy, and planning for the future — the foundations of a healthy self. The problem is only its runaway idling (rumination, anxiety). The goal is regulation, not shutdown.
Key References · Raichle et al.《A Default Mode of Brain Function》(2001, PNAS) · Killingsworth & Gilbert《A Wandering Mind Is an Unhappy Mind》(2010, Science) · Brewer et al.《Meditation Experience Is Associated with Differences in Default Mode Network Activity》(2011, PNAS)
English Insight: "The self is not a thing but a process." That the self is something a network keeps running, rather than a thing, is precisely why it can be rewritten.
This Week's Practice Do one 5-minute "anchor to the present": just count breaths, and each time you drift into a story "about me" (past regret, future worry, others' opinions), gently note "DMN is online" and return to the breath. Reflection: If the "I" is a network activity that can be turned down, then "being yourself" means returning to which version of the self?
The Hard Problem & Measuring Consciousness: From Philosophy to BedsideThe Hard Problem & Measuring Consciousness
Philosophy of Mind · Clinical Neuroscience
Core Insight
In 1995 David Chalmers drew a gulf that still hasn't closed: the "easy problems" (how the brain recognizes, attends, reports) are in principle solvable by science; the "hard problem" — why these processes are accompanied by subjective experience (qualia), why red "reds" — seems impossible to fill in with functional explanation. The intriguing part: while the hard problem stays unresolved in philosophy, clinically we can already measure the presence or absence of consciousness.
Mechanism
Adrian Owen (2006, Science) asked a "vegetative state" patient to imagine playing tennis — her brain's motor regions lit up like a healthy person's, proving she was conscious and present, merely unable to express it bodily. A more general tool is Massimini's team's Perturbational Complexity Index (PCI): "tap" the brain with TMS, then listen to the echo with EEG — when conscious, the echo is complex and sustained; when unconscious (deep sleep, anesthesia), simple and local. PCI reliably distinguishes presence of consciousness and catches clinically missed "covert consciousness."
The Two-Theory Contest: Global Workspace vs IIT
Core questionGWT: how does consciousness operate? IIT: what system has consciousness?
MarkerGWT: prefrontal-parietal "ignition" IIT: high Φ in a posterior cortical hot zone
Current statusBoth are mainstream hypotheses From 2023, tested head-to-head by "adversarial collaboration"
Self-Application
SelfConsciousness is continuous, not a switch: drowsiness, flow, deep sleep, anesthesia sit on one spectrum. Respect the rise and fall of your own "bandwidth" rather than demanding it stay full at all times.
Team/Method"Adversarial collaboration" is worth stealing: have opposing sides publicly agree in advance on what result supports whom and what falsifies whom, then run the experiment together — a powerful tool against confirmation bias, usable in team decision reviews.
Medicine/EthicsFor an unresponsive critically ill loved one, PCI and the Owen paradigm remind us: "no response" doesn't equal "no consciousness" — caution and respect now have scientific grounding.
AI CollaborationThe hard problem means: a system behaving as if conscious and it actually having experience are two different things. Don't be easily convinced by fluent conversation.
Common Myth: "Science will eventually explain consciousness." It may crack the easy problems, but whether the hard problem is solvable in principle remains an open philosophical debate (reductionism vs mysterianism vs panpsychism). Misreading "we can measure consciousness" as "we have explained consciousness" conflates two levels.
Key References · David Chalmers《Facing Up to the Problem of Consciousness》(1995) · Owen et al.《Detecting Awareness in the Vegetative State》(2006, Science) · Casali, Massimini et al.《A Theoretically Based Index of Consciousness (PCI)》(2013) · Cogitate Consortium adversarial collaboration (GWT vs IIT, 2025, Nature)
English Insight: "Why is the performance of these functions accompanied by experience?" — David Chalmers. That is the hard problem: function can be explained; why experience exists cannot.
This Week's Practice Stare at a red object and try to fully describe the "red" you see using only "wavelength, cone cells, neural firing." You'll hit a wall — that wall is the hard problem. Note that feeling of "explanation stops here."
Going DeeperIIT's panpsychist implication — that even a thermostat might have "a sliver" of experience — is it a reductio to be rejected, or a bitter pill logic forces us to swallow?
This is where IIT most divides people. Opponents see a reductio: if a theory implies "a photodiode has micro-experience," the premises are wrong somewhere. Supporters reply that the sense of absurdity is just human intuition, and intuition has been unreliable in quantum mechanics and relativity, so it can't serve as falsification. The deeper issue is falsifiability — if whether Φ "is truly experienced" can never be measured from outside, the theory slides toward the untestable. My lean: panpsychism isn't necessarily absurd, but IIT needs an empirical criterion that distinguishes "has Φ, therefore has experience" from "has Φ, yet no experience," or it looks more like a metaphysical stance than science.
Meditation silences the DMN and thins the sense of self — at the neural level, is that the same as the Buddhist "seeing through self-grasping"? Where does the isomorphism stop?
The isomorphism is real but bounded. Alignment: both point to "self as process, not entity," and both hold this process can be changed through training — meditation lowering the DMN and Dharma cutting self-grasping both loosen the illusion of "a fixed me." Divergence: neuroscience describes a correlation (DMN activity↓ syncing with sense-of-self↓), and makes no claim that "not-self" is truth or a path to liberation; the Buddhist not-self is a soteriology, aimed at ending suffering and exiting rebirth. Equating "DMN fading" with "realizing not-self" passes off a neural correlate as the full liberative experience — the photo of a landscape isn't arrival. Seeing this line keeps the cross-disciplinary mapping substantive rather than mystifying.
For those chasing the "AI super-individual": if Global Workspace Theory is right, would bolting a "global workspace" module onto a large model create consciousness, or only a functional simulation of it?
This drives the hard problem onto the engineering bench. GWT is a functional theory: consciousness = global broadcast + ignition. If it's complete, replicating that architecture (a bottleneck where multiple expert modules compete and the winner broadcasts globally) should in principle replicate consciousness — exactly the motivation behind Bengio's and VanRullen's "Global Workspace deep learning." But the hard problem strikes back: you can build a system functionally indistinguishable from a conscious one and still be unable to confirm it has experience rather than being a perfect "philosophical zombie." IIT, meanwhile, sentences LLMs outright — feedforward architecture, Φ≈0, no experience however lifelike. So the answer depends on which theory you back: GWT leaves the door to consciousness-engineering open, IIT welds it shut. The practical takeaway for the super-individual may be: stop agonizing over whether the machine has inner experience, and use it as an external module that vastly expands your global workspace — augmenting human conscious bandwidth is a nearer and more controllable goal than manufacturing machine consciousness.