What Thoughts Are Made Of

Jan 30
3 min read

Neuroscience, Consciousness, and the Quest to Understand the Mind

The Bold Bet: Neural Networks

Churchland believed that our mental life can be fully reconceived in neurocomputational terms. To him, the key lies in neural networks, whether biological (our neurons) or artificial (computer models inspired by them).

By studying how networks of “neurons” activate in patterns, Churchland suggested that everything from seeing colors to recognizing faces could be described as movements in a vector space, a kind of map where sensations and thoughts are plotted.

Dorsal networks (feedforward): great for classifying objects (like identifying a face).
Recurrent networks (feedback loops): powerful for capturing processes and sequences, such as learning grammar or understanding causality.

In his vision, consciousness itself might emerge from these recurrent loops of activation.

From Neurons to Patterns: How Thoughts Take Shape

Modern neuroscience suggests that thoughts are not localized “objects” in the brain, but distributed activity patterns that span multiple regions at once. When you form a thought, sensory areas, memory circuits, emotional systems, and executive networks activate together, forming a temporary coalition.

This idea aligns with Paul Churchland’s proposal that mental states are best understood as positions in a high-dimensional neural state space, rather than as discrete symbols or sentences in the head. A thought, in this view, is not a word, it is a trajectory through neural space.

Recurrent Loops and the Feeling of Consciousness

One of the most influential hypotheses in consciousness research is that recurrence matters. While feedforward activity can explain perception (“I see a face”), recurrent feedback loops may explain awareness of perception (“I know that I see a face”).

Recurrent activity allows the brain to:

Hold information over time
Compare current input with memory
Generate predictions and error signals

This resonates with the idea that consciousness emerges when neural activity becomes self-referential, when the brain not only processes information, but continuously updates and monitors its own states.

The Skeptical View

But here’s the catch: critics argued that the evidence was far too limited. Small “toy” computer models were being interpreted as if they could already explain human consciousness. It was like mistaking a paper airplane for a Boeing 747.

Susan Greenfield, on the other hand, focused on brain biology and described consciousness in terms of dynamic patterns of neural activity, what she called concentric mind patterns. She was less concerned with computational models and more with how actual brain circuits generate experience.

According to this view:

Consciousness is graded, not all-or-nothing.
It depends on the spread and integration of neural activity.
It changes with brain chemistry, development, and pathology.

This perspective helps explain why consciousness fades gradually in sleep, anesthesia, or neurodegenerative disease, rather than switching off like a light.

Why It Matters

The debate from the 1990s feels strikingly current. As AI grows more sophisticated, we keep circling back to the same puzzle: can machines really “think,” or are they just clever simulations?

Understanding what thoughts are made of is more than philosophy, it touches clinical neuroscience (e.g., how consciousness breaks down in coma or anesthesia), artificial intelligence, and even our sense of what makes us human.

Conclusion: Mapping the Architecture of Thought

The study of how thoughts emerge, whether from neural circuits, chemical signals, or network dynamics, remains one of the most ambitious quests in neuroscience. While we can now trace patterns of brain activity that accompany thinking, we are still far from fully explaining how subjective experience arises from biology.

What’s clear is that thoughts are not random sparks, they are grounded in the brain’s architecture, influenced by memory, emotion, and context. By unraveling this mystery, we move closer to bridging the gap between the physical brain and the intangible mind.

Understanding what thoughts are made of is not just about neurons,

it’s about discovering how biology gives rise to

meaning, creativity,

and consciousness itself.

Source: Stevan Harnad, What thoughts are made of, Nature, Vol. 378, 30 November 1995.