Learn
Goals and Concepts
Start with the capability target and concept set for this module.
Module 01: Scientific Curiosity and Motivation
Launch into connectomics by turning curiosity into testable scientific questions with explicit motivation and boundaries.
Lesson Flow
Practice
Studio Activity
Apply the ideas in a guided activity tied to realistic outputs.
Check
Assessment Rubric
Use the rubric to verify competency and identify improvement targets.
Teach
Slides and Worksheets
Open the teaching deck, worksheet, and editable slide source.
Interactive Lab
Practice in short loops: checkpoint quiz, microtask decision, and competency progress tracking.
Why Map the Brain Checkpoint
Question Framing Microtask
Pick the best study question for a first connectomics project.
Progress Tracker
State is saved locally in your browser for this module.
0% complete
Capability target
Write one connectomics study question with measurable structural outputs and one explicit non-claim. Articulate a personal motivation statement linking daily annotation work to a larger scientific mission.
Why this module matters
Motivation drives persistence, but technical progress requires disciplined question framing. Connectomics demands sustained effort: proofreading thousands of neurons, tracing axons through noisy volumes, and reconciling ambiguous merges. Without a clear sense of purpose, even talented annotators burn out. This module anchors learners’ curiosity in concrete, testable questions so that motivation survives the transition from excitement to routine.
Concept set
1) Question before method
- Technical: define target measurement before tool choice. A well-formed connectomics question specifies the circuit, the organism, the resolution, and the expected structural readout (e.g., synapse count, path length, motif frequency). Tools like FlyWire or CAVE are powerful but directionless without a hypothesis.
- Plain language: know what you are asking first.
- Misconception guardrail: tools generate good questions automatically.
2) Structure informs, not fully explains
- Technical: structural data constrains hypotheses but does not alone prove dynamic function. A synapse between neuron A and neuron B tells you connection exists; it does not tell you whether the synapse is active under a given stimulus. The C. elegans connectome was completed decades ago, yet behavior prediction from wiring alone remains an open problem.
- Plain language: maps guide, they do not finish the story.
- Misconception guardrail: a completed connectome is a full explanation of behavior.
3) Motivation should be bounded
- Technical: ambitious goals need near-term measurable milestones. Break “map the whole brain” into “proofread 50 neurons in optic lobe region X by Friday.”
- Plain language: big vision, small testable steps.
- Misconception guardrail: broad vision statements are sufficient project plans.
4) Why curiosity matters in connectomics
- Technical: the field requires sustained motivation through tedious annotation work. Large-scale connectomics projects (FlyWire, MICrONS, MouseConnects HI-MC) depend on thousands of hours of manual proofreading. Curiosity about the underlying biology is the fuel that keeps annotators engaged through repetitive tasks.
- Plain language: you need a reason to keep going when the work gets boring.
- Misconception guardrail: excitement at the start is enough to carry you through a multi-month project.
5) The “motivation gap”
- Technical: early excitement diverges from the reality of proofreading thousands of neurons. The gap between “I want to understand the brain” and “I am fixing merge errors in segment 47,832” is real and predictable. Recognizing this gap in advance helps learners build coping strategies (milestone celebrations, rotation between tasks, connecting daily work to publications).
- Plain language: the honeymoon phase ends; plan for it.
- Misconception guardrail: if the work feels tedious, you chose the wrong field.
6) Growth mindset in technical training
- Technical: errors are learning opportunities, not failures. In proofreading, every false merge you catch teaches you about the segmentation algorithm’s failure modes. Every split error reveals tissue preparation artifacts. Tracking your error patterns builds expertise faster than avoiding mistakes; even expert proofreaders disagree on 5-10% of decisions.
- Plain language: mistakes teach you what the computer got wrong and why.
- Misconception guardrail: good annotators never make errors.
7) The connectomics “why”
- Technical: linking daily annotation work to the larger scientific mission of understanding brain circuits. Each proofread neuron contributes to a wiring diagram that enables circuit-level hypotheses about sensory processing, motor control, learning, and disease. The FlyWire project demonstrated that distributed annotation by 287 contributors could produce a whole-brain connectome.
- Plain language: every neuron you trace is a sentence in the story of how brains work.
- Misconception guardrail: my individual contribution is too small to matter.
Worked example: narrowing “how does memory work?” into a Monday-morning plan
The numbers below are illustrative — they show the shape of the narrowing, not results from a specific study.
You arrive with the question “how does memory work?” That is a real motivation and an unusable plan: it names no organism, no circuit, no measurement, and no result that could prove anything wrong.
Step 1: find the structural handle. Autoassociative memory models have long pointed at recurrent excitatory wiring among hippocampal CA3 pyramidal cells. That single move fills three slots: organism (mouse), region (CA3), structure (recurrent synapses between CA3 pyramidal cells). The MouseConnects HI-MC volume targets this region, so a dataset exists.
Step 2: convert to a measurement with units. “CA3 is recurrently connected” becomes: among proofread CA3 pyramidal cells whose axons stay inside the volume, what fraction of ordered cell pairs is synaptically connected, and with how many synapses per connected pair? A feasible first slice: 60 proofread cells give 60 x 59 = 3,540 ordered pairs to check. If the literature-informed expectation is a connection probability near 1-2%, you expect roughly 35-70 connected pairs — enough to estimate a rate, not enough to subdivide by cell subtype.
Step 3: write the non-claim before anyone asks for it. Finding 2% recurrent connectivity does not show that these synapses store memories, are potentiated, or support pattern completion — functional claims structure alone cannot carry. The discipline for sorting claims into evidence bins, with a worked repair of a question much like this one, is Technical Unit 01; defer to it rather than re-deriving the bins.
Step 4: define the falsification condition. The hypothesis “CA3 recurrence exceeds chance” is only testable against a stated chance. Commit now: the claim fails if measured connection probability is indistinguishable from what axon-dendrite proximity alone predicts. The machinery for building that null honestly is Module 20.
Step 5: bound the motivation. The project is thousands of proofread neurons. Your Friday milestone is 5 cells proofread and one pair-checking script that runs end to end. Name this ladder in your motivation statement, so the sentence connecting segment 47,832 to the storage question is already written when you need it.
What this example does not establish: that CA3 is the right place to study memory. It shows the move from theme to question; every step repeats on a different theme in under an hour.
Core workflow
- Identify curiosity question.
- Convert to measurable structural hypothesis.
- Define one metric and one limitation.
- Plan first dataset/tool touchpoint.
- Write a personal motivation statement connecting your question to a long-term scientific goal.
Detailed run-of-show (90 minutes)
Block 1: Opening hook (00:00-12:00)
- Instructor script: “Welcome. Today we answer one question: why would anyone spend years mapping brain wires? Let me show you.” Play 3-minute clip from Sebastian Seung’s TED talk. Then show a before/after of a raw EM image vs. a fully reconstructed neuron. Ask: “What questions could you answer with this reconstruction that you could not answer with the raw image?”
- Collect 3-4 responses on whiteboard. Highlight that each response implies a different measurement.
Block 2: Connectomics landscape (12:00-28:00)
- Instructor script: “Let’s ground this in real projects.” Walk through three case studies in 5 minutes each:
- C. elegans (White et al., 1986): the first complete connectome. 302 neurons. What it enabled, what it could not explain.
- FlyWire whole-brain (Dorkenwald et al., 2024): 130,000+ neurons, 287 proofreaders, first whole-brain connectome of an adult animal with complex behavior.
- MouseConnects HI-MC (ongoing): scaling to the mouse hippocampus at 10 mm³, the challenge of petascale data.
- After each case study, ask: “What question drove this project?” Write answers on board.
Block 3: Question framing workshop (28:00-48:00)
- Instructor script: “Now it is your turn. Take 5 minutes to write down the broadest brain question you care about. Do not filter.” (5 min silent writing)
- “Now narrow: what specific circuit or region relates to your question? What structural measurement would you need?” (5 min revision)
- Pair-share: partners critique each other’s questions using the checklist: Does it specify organism? Region? Measurement? Limitation? (10 min)
Block 4: Evidence-boundary critique (48:00-65:00)
- Instructor script: “Every good connectomics question has a twin: the non-claim. What can your structural data NOT tell you?” Present three example hypotheses and their non-claims. Learners practice writing non-claims for their own questions.
- Class discussion: collect 3 examples of well-formed question + non-claim pairs.
Block 5: Motivation statement drafting (65:00-80:00)
- Instructor script: “Before we close, I want you to write something personal. Why are you here? Not the resume version — the real version. What about brains or circuits or data makes you want to do this work? And how will you remind yourself of that reason when the work gets tedious?”
- Silent writing: 10 minutes. Prompt: “Write 3-5 sentences explaining why you want to work in connectomics and what you will do when motivation dips.”
- Voluntary sharing: 2-3 learners read their statements aloud.
Block 6: Exit ticket (80:00-90:00)
- Submit: (1) your testable question with metric, dataset, and non-claim; (2) your motivation statement.
- Instructor script: “These two documents are your compass for the rest of the program. We will revisit them in Module 06.”
Studio activity: “Write your connectomics motivation statement”
Overview
Learners produce two artifacts: a question-to-hypothesis sheet and a personal motivation statement.
Part A: Question-to-hypothesis sheet (30 minutes)
- State your broad curiosity question (1 sentence).
- Narrow to a specific circuit, region, or organism (1 sentence).
- Define the structural measurement you would need (e.g., synapse count between cell types X and Y).
- Specify the dataset you would use (e.g., FlyWire, MICrONS, FAFB).
- State one non-claim: what your structural data cannot tell you.
- Define a falsification condition: what result would disprove your hypothesis?
Part B: Motivation statement (20 minutes)
Write 150-300 words addressing:
- Why connectomics? What drew you to this field?
- What specific aspect of brain circuitry fascinates you?
- How does your daily work (annotation, proofreading, coding) connect to the larger mission?
- What is your plan for sustaining motivation through tedious stretches? (Be specific: milestones, rewards, accountability partners, rotation between tasks.)
Peer review (10 minutes)
Exchange motivation statements with a partner. Provide feedback on: (1) specificity — does the statement name concrete goals? (2) sustainability — does the plan for maintaining motivation seem realistic?
Outputs
- Question-to-hypothesis sheet: broad question, narrowed target, named structural measurement, named dataset, one non-claim, one falsification condition.
- Motivation statement of 150-300 words, including a specific plan for sustaining effort through tedious stretches.
- Written peer feedback on a partner’s statement, addressing specificity and sustainability.
- One revision made to your own sheet or statement in response to that feedback.
Assessment rubric
- Minimum:
- Question names organism, region or circuit, and one structural measurement with units.
- Metric, dataset, and non-claim are all present and consistent with each other.
- Motivation statement addresses both why and how.
- Strong:
- Falsification condition names the specific result that would count against the hypothesis.
- Scope is sized so the first milestone could be finished within a week.
- Motivation statement names at least two sustainability strategies with triggers (“when X dips, I will Y”).
- Non-claim correctly identifies a functional inference the structural data cannot support.
- Failure:
- Motivational text without measurable outputs, or a question no result could contradict.
- Motivation statement is generic (“I like brains”) or omits a plan for the motivation gap.
- Claim language exceeds the evidence class of the proposed measurement.
Common errors and how to recover
- Your question has a theme but no measurement. “I want to study learning” names nothing checkable. Recover by forcing the four slots — organism, region, structural readout with units, dataset. Small and answerable beats grand and unfalsifiable.
- Your claim outruns your evidence. You planned to count synapses and wrote “this circuit computes prediction error.” Recover by sorting each claim into the bins of Technical Unit 01 and rewriting anything structure cannot establish as a constraint statement or future work.
- Your scope is a career, not a project. “Map the hippocampus” is a decade of institutional effort. Recover by cutting until you can say what will be finished on Friday, then chaining Fridays.
- Your motivation statement would survive no tedious afternoon. Recover by rewriting it to name the specific question, the daily task that serves it, and the action you will take when motivation dips — a milestone, a rotation, an accountability partner with a date.
- You read boredom as a diagnosis. Three weeks of merge errors feels like proof you chose wrong. Recover by re-reading your motivation statement — that is what it is for — and by logging what each fixed error taught you about the segmentation’s failure modes.
What this module does not cover
- The evidence discipline in depth. The three-bin sorting of structural claims and the worked repair of a research question are Technical Unit 01.
- Research norms and the hidden curriculum. How labs actually work is Module 02.
- The datasets and tools themselves. The flagship projects are covered in the case studies: FlyWire whole-brain and MouseConnects HI-MC.
- The anatomy needed to interpret what you see. That is Module 04 and Technical Unit 02.
- Testing your question statistically. Null models and claim calibration are Module 20 and Technical Unit 09.
Content library references
Teaching resources
Academic references
- White, J. G., Southgate, E., Thomson, J. N., & Brenner, S. (1986). The structure of the nervous system of the nematode Caenorhabditis elegans. Philosophical Transactions of the Royal Society B, 314(1165), 1-340.
- Dorkenwald, S., et al. (2024). Neuronal wiring diagram of an adult brain. Nature, 634, 124-138.
- Seung, H. S. (2012). Connectome: How the Brain’s Wiring Makes Us Who We Are. Houghton Mifflin Harcourt.
- Dweck, C. S. (2006). Mindset: The New Psychology of Success. Random House.
- Lichtman, J. W., & Denk, W. (2011). The big and the small: challenges of imaging the brain’s circuits. Science, 334(6056), 618-623.
Quick practice prompt
Write a 3-sentence hypothesis with one metric and one caveat. Then write 2 sentences explaining why this question matters to you personally.
Teaching Materials
Slide Deck
Classroom-ready deck links for teaching and delivery.
Activity Worksheet
Learner worksheet aligned to the studio activity and rubric.
Slide Source
Marp source file for editing and rendering.
course/decks/marp/modules/module01.marp.md