We know that certain brain activity correlates with conscious experience. When you're awake and aware, certain things are happening neurologically. When you're under anesthesia, those things aren't happening. But here's the much harder question: what's actually enough to produce consciousness? What's the minimum configuration of neurons and activity that gets you from "electrical signals bouncing around" to "there's something it's like to be this system"? A review in Neuroscience & Biobehavioral Reviews surveys the theoretical battlefield, and fair warning: nobody agrees.
Correlation Is Not the Same Thing as "This Is What Makes It Work"
Finding correlates of consciousness is relatively straightforward methodologically. You compare brain activity when people are conscious versus when they're unconscious. Whatever differs is correlated with consciousness. Done, right?
Not really. The problem is that correlation isn't the same as sufficiency. Maybe we're measuring side effects rather than the actual generator of experience. The brain does lots of things when you're conscious that it doesn't do when you're not, but that doesn't mean all those things are required for consciousness. Some might be consequences, some might be coincidental, and some might be necessary but not sufficient.
The question of what minimal brain organization actually produces consciousness is where neuroscience meets philosophy and neither field has a definitive answer.
The Cortex Gets All the Attention (But Maybe Shouldn't)
Most theories of consciousness focus heavily on the cortex, that wrinkled outer layer of the brain that seems responsible for most of what we consider higher cognition. Makes sense. Damage the cortex in various ways and people lose various conscious abilities. The cortex seems like where the action is.
But subcortical structures are increasingly entering the conversation. The thalamus, brainstem, and other deeper regions might not just support consciousness by providing inputs and infrastructure. They might actually contribute to producing it.
The review evaluates evidence for various positions: maybe consciousness requires cortex specifically, maybe subcortical structures are sufficient on their own in some cases, or maybe it's all about distributed patterns of activity across multiple regions. The answer might not be "which brain region" but "what kind of communication between regions."
The Theory Wars
There are several major theories of consciousness, and they disagree about what's sufficient. This is genuinely useful because different sufficiency predictions can (in principle) be tested.
Global Workspace Theory says consciousness arises when information gets broadcast widely across cortical areas. Local processing isn't enough. You need information to enter a "global workspace" that makes it available to multiple brain systems at once. Under this view, the sufficient conditions involve particular patterns of widespread cortical communication.
Integrated Information Theory takes a different approach entirely. It focuses on how much information a system integrates. Consciousness equals integrated information, measured by a quantity called "phi." Under this view, what matters isn't where the activity is or whether it's global, but how much information the system integrates above what its parts could do separately. This theory makes the interesting (some would say troubling) prediction that even simple systems could be conscious if their phi is high enough.
Higher-Order Theories require meta-cognition. You're only conscious of something if you're also representing the fact that you're perceiving it. Under this view, first-order sensory processing isn't conscious on its own. You need a higher-order representation of that processing.
These theories make genuinely different predictions about what configurations of matter would be sufficient for consciousness. The challenge is figuring out how to actually test those predictions.
Why This Matters Outside Philosophy Seminars
This might seem like an abstract question that only philosophers care about. But it has surprisingly practical implications.
Could we create conscious AI? If we don't know what's sufficient for consciousness, we can't answer that question. We might build systems that are conscious without realizing it, or we might worry about conscious machines when they're actually just very good at processing without any inner experience.
What about patients in vegetative states or with severe brain damage? They can't report their experiences. If we knew what was sufficient for consciousness, we might be able to assess their likelihood of having inner experience based on what brain activity remains.
What about animals with very different brains? Fish? Insects? Octopuses? Their neural architecture is different from ours. Are they conscious? Which ones? If we understood the sufficient conditions, we could make more informed guesses.
These questions feel abstract until you're a clinician facing decisions about patients who can't tell you what they're experiencing, or until you're a policymaker setting animal welfare standards, or until you're wondering whether the AI system you're building deserves moral consideration.
Still No Final Answer
The review makes clear that despite decades of research and theorizing, we still don't have a consensus answer to what's sufficient for consciousness. Different theories, different predictions, ongoing arguments.
What we do have is a much more precise formulation of the question and some testable predictions that might eventually distinguish between theories. The science is progressing, even if the mystery remains. The question "what does it take to be conscious?" turns out to be really, really hard. But at least we're getting better at asking it.
Reference: Bhattacharyya S, et al. (2025). A review of the sufficient conditions for consciousness. Neuroscience & Biobehavioral Reviews. doi: 10.1016/j.neubiorev.2025.105816 | PMID: 40816661
Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.