Brains and AIMiddleLesson 2956 min read

What a connectome is

A connectome is a map of every neuron in a nervous system and the connections between them. Learn what such a map shows, and how scientists study it as a network.

✦ Ask the tutor 🧪 Lab · 25 min 📝 Lesson test

Speed

Lesson in motion

In 60 seconds

What a connectome is

A connectome is a map of every neuron in a nervous system and the connections between them. Learn what such a map shows, and how scientists study it as a network.

1/4

In simple words

Think of a map of all the roads in your town. It shows which houses connect to which lanes, and which lanes lead to the main road. A connectome is a road map for a brain. The "houses" are nerve cells, called neurons. The "roads" are the connections between them. The map does not show the traffic, but it shows every road the traffic could take.

Your brain, and the brain of every animal, is made of cells called neurons. A neuron receives signals from other neurons, and can send a signal on to others. The meeting point where one neuron passes a signal to the next is called a synapse.

A connectome is a complete map of the neurons in a nervous system and the synapses between them: who connects to whom, and often how many synapses each pair shares. The word comes from "connection", like "genome" for the full set of genes.

Three kinds of neurons, in simple terms

Kind	What it does	Everyday picture
Sensory neurons	Pick up information: light, smell, taste, touch	The shop's staff at the door who notice who comes in
Interneurons	Connect neurons to each other and combine signals	The office in the middle where decisions are made
Motor neurons	Send signals to muscles, so the body moves	The delivery staff who actually carry things out

The first complete map: a tiny worm

The first animal to have its whole connectome mapped was a tiny worm called C. elegans. It is about a millimetre long and lives in soil. The common form of this worm has just 302 neurons. Scientists sliced the worm into extremely thin slices, photographed each slice with an electron microscope, and traced every neuron by hand across the pictures. The map was published in 1986, after many years of careful work.

Since then, scientists have mapped larger and larger pieces of brains. In the next lesson you will see how the whole brain of an adult fruit fly, with about 140,000 neurons, was mapped.

A connectome is a network

Computer scientists call a map of things and links a graph. Each neuron is a node. Each connection is an edge, with a direction: the signal goes from one neuron to the next. The number of synapses can be stored as a weight on the edge. Once a brain is a graph, we can ask network questions:

Hubs: which neurons have the most connections, like a busy railway junction?
Pathways: what is the shortest route from a smell neuron to a leg muscle neuron?
Groups: which neurons are tightly connected to each other, like a WhatsApp group?

Python: a tiny made-up connectome as a table of numberspython

import numpy as np

# MADE-UP example, only for practice. Not real data.
names = ["taste", "relay", "decide", "mouth_motor"]
# W[i][j] = number of synapses from neuron i to neuron j
W = np.array([
    [0, 8, 0, 0],   # taste  -> relay
    [0, 0, 6, 0],   # relay  -> decide
    [0, 0, 0, 9],   # decide -> mouth_motor
    [0, 0, 0, 0],
])
for i, a in enumerate(names):
    for j, b in enumerate(names):
        if W[i, j] > 0:
            print(f"{a:>7} -> {b:<12} {W[i, j]} synapses")
print("Total synapses:", W.sum())

Note

This grid of numbers is called an adjacency matrix. For a real brain it is huge but mostly empty, because each neuron connects to only a small share of all the others.

What the map tells you, and what it does not

A connectome does NOT show

The signals moving right now, like traffic on the roads
How the connections change when the animal learns
Chemicals that wash over many neurons and change their mood
Everything about each synapse, such as its exact strength

A connectome DOES show

Every neuron in the mapped brain, and its shape
Which neurons can talk directly to which
How many synapses link each pair, a hint of strength
All possible pathways from senses to muscles

Real example

A student in Pune thinks about the Mumbai local train map. It shows every station and line, and which stations are big junctions. But it cannot tell you how crowded the 8:15 train is today. A connectome is the same kind of map: it shows the lines, not today's crowd.

Try this. Draw five circles on paper for five neurons: one "sweet taste", one "bitter taste", two in the middle, and one "mouth muscle". Draw arrows for connections. Can you find two different routes from "sweet taste" to "mouth muscle"? That drawing is a tiny connectome.

Watch and read more

Lab

A small made-up connectome in networkx, with its hub neurons, a shortest pathway, and a test of what happens when a hub is removed.

~25 min

The problem

This runs in your browser. All data is made up for practice.

Run the starter. It builds a small directed graph of made-up neurons.
Finish top_hubs(G, k): return the names of the k neurons with the most connections (incoming plus outgoing), most connected first. Hint: G.degree().
Finish shortest_route(G, source, target): return the list of neurons on a shortest path. Hint: nx.shortest_path.
Make a copy of the graph, remove "hub_X" from the copy, and set still_reaches to True or False: can "smell_1" still reach "motor_leg"? Hint: nx.has_path.
Write one sentence on what this tells you about hubs.

Practice lab

Edit the code and run it. Python runs in your browser — nothing is sent to our server.

Ctrl + Enter to run

import networkx as nx

# A tiny MADE-UP connectome, only for practice. Not real fly or worm data.
# Each edge: (from_neuron, to_neuron, number_of_synapses)
edges = [
    ("smell_1", "relay_A", 12), ("smell_2", "relay_A", 9),
    ("touch_1", "relay_B", 7),  ("relay_A", "hub_X", 15),
    ("relay_B", "hub_X", 11),   ("hub_X", "decide_1", 20),
    ("hub_X", "decide_2", 6),   ("relay_A", "decide_2", 4),
    ("decide_1", "motor_leg", 14), ("decide_2", "motor_mouth", 10),
    ("hub_X", "relay_B", 3),    ("touch_1", "hub_X", 2),
]
G = nx.DiGraph()
for a, b, w in edges:
    G.add_edge(a, b, weight=w)
print("Neurons:", G.number_of_nodes(), "| Connections:", G.number_of_edges())

def top_hubs(G, k=2):
    """Names of the k neurons with the most connections (in + out), most first."""
    # TODO: use G.degree() and sort by the number
    return []

def shortest_route(G, source, target):
    """List of neurons on a shortest path from source to target."""
    # TODO: use nx.shortest_path
    return None

print("Hubs:", top_hubs(G, 2))
print("Route:", shortest_route(G, "smell_1", "motor_mouth"))

# TODO: copy G, remove "hub_X" from the copy, test the path
still_reaches = None
print("Can smell_1 still reach motor_leg without hub_X?", still_reaches)

Output

No output. Add print(...) to see values.

You are done when

top_hubs returns the most connected neurons, most connected first
shortest_route returns a valid, shortest list of neurons
Removing hub_X from a copy shows that smell_1 can no longer reach motor_leg
The sentence explains that many pathways depend on a few hub neurons

Hard questions

Try to answer before you reveal. If you can answer these, you understood the lesson.

Q1Counting connections treats a 2-synapse link the same as a 20-synapse link. How could you find hubs that take synapse counts into account?Reveal

Use a weighted degree: add up the synapse counts on each neuron's edges instead of counting edges. In networkx, G.degree(weight="weight") does this. The two measures can rank neurons differently, so it is worth checking both and thinking about which question you are asking.

Q2The shortest path uses the fewest steps. Why might a real signal not take the shortest path?Reveal

A path with fewer steps may pass through weak connections, or through inhibitory neurons that block the signal. Each neuron also needs enough input to fire. So the route a signal really takes depends on connection strengths, whether they excite or inhibit, and what else is active. The map lists possible routes; activity decides which are used.

Please sign in to continue.

Questions people ask

Is a connectome the same as a brain scan in a hospital?

No. A hospital MRI scan shows the brain's shape and large regions, in pieces far bigger than single neurons. A connectome maps single neurons and single synapses, which needs electron microscopes and, so far, has been done completely only for small animals.

Why start with worms and flies instead of humans?

Size. A small worm has a few hundred neurons and a fruit fly has about 140,000. The human brain has about 86 billion neurons, so mapping all of it at this level of detail is far beyond what anyone can do today. Small brains let scientists build the methods first.

If we know the whole map, do we understand the animal?

Not fully. The map shows the possible routes, but not the signals, learning or chemistry. It is a very strong starting point, and later lessons show how scientists use it together with experiments.

Are neurons in AI models the same as real neurons?

No. AI "neurons" are simple maths inspired by real ones. Lesson 5 compares them side by side.

Lesson test

5 questions. Get 3 right (60%) to pass and complete this lesson.