Bing Brunton on Connecting the Connectome to the Body | Mindscape 352
YouTube · YsNcg8cc2xA
Quick Read
Summary
Takeaways
- ❖The 'connectome' is the wiring diagram of the brain, but its definition varies in scale (cellular vs. mesoscale).
- ❖Mapping a full cellular-level connectome for complex organisms like humans is currently impossible.
- ❖The C. elegans worm (300 neurons) connectome was mapped decades ago, but its function is hard to understand due to reliance on chemical and mechanical computation.
- ❖The fruit fly's full connectome (150,000 brain neurons + 22,000 ventral nerve cord neurons) was mapped in the last year, offering a more tractable system.
- ❖Fruit flies have 'cell types' and 'jointed limbs,' making their connectome more directly helpful for understanding neural computation than C. elegans.
- ❖A 'pruning' study on the fruit fly's leg-controlling neurons identified a minimal circuit of three neurons (two excitatory, one inhibitory) sufficient to generate walking rhythms.
- ❖All animal locomotion and many other biological processes (like breathing and digestion) are rhythmic, generated by 'central pattern generators' (CPGs).
- ❖The brain is not isolated; it evolved to control a specific body with its unique limbs, muscles, joints, and sensors.
- ❖The concept of 'digital twins' for animals involves simulating the nervous system within a physically realistic virtual body to understand behavior.
- ❖Behavioral fidelity in AI models does not guarantee biological fidelity; a worm's connectome can be trained to control a fly's body, highlighting the 'digital sphinx' problem.
Insights
1Connectome Definition and Challenges
The connectome is the wiring diagram of the brain, but its definition is debated, ranging from individual cell-to-cell connections to coarse-grained brain area connections. While a full cellular connectome for humans is technologically out of reach, simpler organisms like C. elegans (300 neurons) have been mapped for decades. However, C. elegans' reliance on chemical and mechanical computation makes its connectome difficult to interpret for behavior.
The rough idea is that we all know that the brain is composed of cells... So there's essentially a wiring diagram, so to speak, of the brain... The difficulty comes in in terms of how do you actually define the units... We've had the connectome of the C. elegans worm for so long and yet we still don't understand it. They do a lot of computation not using that connectivity matrix.
2Fruit Fly Connectome Breakthrough
The full connectome of the Drosophila fruit fly (150,000 brain neurons and 22,000 ventral nerve cord neurons) was recently mapped. This larger, more complex system, with jointed limbs and distinct cell types, is proving more amenable to direct interpretation of its neural connectivity for understanding behavior.
The one that has come out much more recently... is a couple of efforts... to map the full connectivity matrix of a Drosophila fruit fly... The brain has 150k and then the ventral nerve cord has an additional 22k... It has jointed limbs just like humans do. And it has enough cells that they're actual cell types.
3Discovery of Minimal Walking Rhythm Circuit
By simulating the fruit fly's ventral nerve cord connectome (initially 4,000 neurons controlling two front legs) and iteratively 'pruning' unnecessary cells, researchers identified a minimal circuit of just three neurons (two excitatory, E1 and E2, and one inhibitory, I1) sufficient to generate the basic walking rhythm. This model-driven prediction was partially validated by experimental observation where activating a specific predicted neuron caused leg tapping.
We started out with 4,000 cells... She went off and did it. The answer was three. Three cells. That's the minimum you need... Two of the cells are excitatory... And one of the cells is inhibitory... Our model said if you if you zap it with a laser it should make the leg tap... And it tapped its leg.
4The Embodied Brain and Digital Twins
The 'embodied brain' concept posits that the nervous system must be understood in the context of the body it controls. This leads to the development of 'digital twins' for animals: virtual models that combine a simulated nervous system with biomechanically realistic bodies in a physics engine. This approach aims to predict complex behaviors and understand the holistic interactions between the brain and body.
The brain does not live in a jar... It always controlled a body and it always controlled a specific body with these limbs and these muscles and these joints and these sensors... We and some other people have been calling them digital twins... it would have an simulation of the nervous system and the interfaces between the nervous system and the body... situated in a virtual reality environment that's capable of interacting with with things.
5Caution: Behavioral Fidelity vs. Biological Fidelity
Achieving realistic behavior in a simulated animal does not automatically imply biological accuracy. The 'digital sphinx' experiment demonstrated this by successfully training a C. elegans worm connectome to control a biomechanically realistic fruit fly body. This highlights that powerful machine learning algorithms can produce behavioral outputs without biologically meaningful interfaces or underlying neural mechanisms.
If you use enough deep learning... it is perfectly possible to get a worm brain to control a fly body... You can get behavioral fidelity without any biological fidelity.
Key Concepts
Connectome
The complete wiring diagram of an organism's nervous system, detailing all neural connections. Its utility depends on the scale of mapping (cellular vs. mesoscale) and the organism's computational reliance on neural vs. chemical/mechanical processes.
Central Pattern Generator (CPG)
A neural circuit within the central nervous system capable of generating rhythmic motor activity (like walking, breathing, or chewing) without rhythmic sensory input. CPGs are fundamental to animal locomotion and other cyclic biological functions.
Embodied Cognition/Brain
The concept that the brain cannot be fully understood in isolation but must be studied in the context of the body it controls and the environment it interacts with. The body's physical properties, sensors, and actuators are integral to how the nervous system functions and evolves.
Digital Twin
A virtual model of a physical system (in this context, an animal) that is updated with real-time data and can be used for simulation, prediction, and optimization. For biological systems, it involves simulating the nervous system within a biomechanically realistic body in a physics engine.
Lessons
- Approach complex biological systems with a blend of skepticism and experimental validation; initial doubts about connectome utility were overcome by new data and methods.
- Utilize computational modeling as a predictive tool, not just for fitting existing data, to generate novel, testable hypotheses in neuroscience.
- Embrace interdisciplinary collaboration (e.g., neurophysiology, computational modeling, robotics) to tackle 'holistic' biological problems that single disciplines cannot address alone.
Quotes
"The connectome roughly speaking is that for for all the cells and their connections and the identities in the brain."
"We've had the connectome of the C. elegans worm for so long and yet we still don't understand it. We do not understand it."
"The brain does not live in a jar. It always controlled a body and it always controlled a specific body with these limbs and these muscles and these joints and these sensors."
"Humans are really terrible at at reasoning through what happens with feedback circuits and and recurrence. Like we can go forward, we can follow a path like A to B to C to D that we can do as soon as there's recurrence... Our intuition for what's going to happen is really poor."
"If you use enough deep learning... it is perfectly possible to get a worm brain to control a fly body. You can get behavioral fidelity without any biological fidelity."
Q&A
Recent Questions
Related Episodes
Joe Rogan Experience #2467 - Michael Pollan
"Michael Pollan and Joe Rogan explore the profound mysteries of consciousness, from the intelligence of plants to the existential threats and opportunities presented by AI, challenging our anthropocentric view of the world."
Dopamine Expert: How TikTok Is Physically Rewiring Your Brain (Permanent Damage?)
"Dr. Anna Lembke, a Stanford addiction expert, explains how modern abundance and frictionless digital rewards are hijacking our dopamine systems, leading to widespread addiction and anhedonia, and offers strategies for reclaiming control."
Essentials: The Neuroscience of Speech, Language & Music | Dr. Erich Jarvis
"Dr. Erich Jarvis explains that human language isn't a separate brain module, but an advanced form of vocal learning rooted in specialized speech production and auditory pathways shared with a few other species, with profound implications for understanding cognition and development."
LIVE: Fed Chair Nominee Kevin Warsh Senate Confirmation HEARING
"Kevin Warsh's Senate confirmation hearing for Federal Reserve Chair exposes deep partisan divides, ethical concerns over undisclosed assets, and fundamental disagreements on Fed independence and economic policy."