Python roadmap

Below is a 12-week, self-paced roadmap tailored for an undergraduate engineering student

aiming to master Python for neuron simulations and brainwave plotting. I’ve assumed you
can dedicate around 1–1.5 hours on weekdays and 2 hours each weekend day, but feel free
to adjust based on your actual availability. Each week includes:
 ?2??e?? ?y??o? ?o??m?? ?o? ?e??o??i??c? ?i??l??i??s

?y??o? ?u??a??n??l?
?e?? ? ? ? ? ?e??e??e?
?r??h ?p ?n ?o?? ?y??o? ?o???p??
?n? ?u??t??n ?o??d??i??s

?c??n??fi? ?y??o? ?t??k

?e??n ?u??y? ?a??l??l??, ?n?
? ?e?? ? ? ?

?a?? ?a??l??g ?o? ?c??n??fi? ?a??s

?a??c ?e??o? ?o??l?

?e?? ? ? ? ? ?n??o??c??o? ?o ?o??u??t??n??
?e??o??i??c? ?n? ?e??o? ?o??l??g

?e??o?? ?i??l??i??s ?
?r??n??v?? ? ?e?? ? ? ?

?i??l??e ?m??l ?e??a? ?e??o??s

?n? ?e??r??e ?y??h???c
?r??n??v??

?i??a? ?r??e??i?? ?e??

?e?? ? ? ?i??
?n??y?? ?E??s??l? ?i??a?? ?s??g
?y??o?

?d??n??d ?i??l??i??
?i??a??e? ? ?e?? ?0

?x??o?? ?p??m??a??o? ?n?

?d???c?? ?i??l??i?? ?i??a??e?

?a??t??e ?r??e??
?e?? ?1 ? ?s??m??y
?s??m??e ?n? ?i??l??e ?
?o??r??e??i?? ?e??o??i??c?
?r??e??

?n??y??s? ?i??a??z??i??,
?r??e??a??o? ? ?e?? ?2

?n??y?? ?e??l??, ?i??a??z? ?a??,

?n? ?r??e?? fi?d???s

• Core topics to cover.

 • Weekly goals and checkpoints (concrete deliverables).
• Active learning techniques (coding drills, mini-projects, Socratic questioning, etc.).
• Forward-looking notes that question assumptions and push you to think beyond the
immediate material.

?e??n??g ?o??o??n??

?o?? ?o??c? ?e??l? ?o??s ?c??v? ?o??a??-

?s??n??a? ?u??e??s ?e??n??g ?o??i??
?o??r??e
?o ?e ?t??i?? ?e??v??a??e? ?n? ?o??s
?e??n??u?? ?o?
?i??s??n?? ?n??g??g ?e??n??g ?o??s ?u??t??n??g
?r??e?? ?s??m??i??s ?n?
?u??r? ?h??k???

Remember: this is a roadmap, not a rigid schedule. If you find yourself moving faster/slower,
simply adapt the checkpoints. Let’s get started.

Week 1 & 2: Python Fundamentals Refresher (Brush Up and

Question Your Foundations)
Why these first two weeks matter
Even if you’ve seen Python syntax before, mastering the basics is crucial. In neuron modeling,
subtle indexing errors or misunderstanding loops can throw off an entire simulation.
Approach everything skeptically: “Do I really understand how Python executes this loop?
What happens if I change this variable here?”
Topics to Cover
1. Core syntax and data types
• Variables, basic data types (int, float, str, bool).
• Lists, tuples, and dictionaries; when to use each.
• Basic I/O and string formatting.
2. Control flow and functions
• if/elif/else constructs; common pitfalls (e.g., indentation).
• for loops vs. while loops; list comprehensions.
• Defining your own functions, arguments vs. keyword arguments, return values.
3. Modules and packages
• Importing built-in modules (math, random) and installing external ones via pip.

• Understanding namespaces and how Python resolves names.

4. Error handling & debugging
• Try/Except blocks.
• Reading and interpreting tracebacks.
 • Strategic print statements vs. using an interactive debugger (pdb, or VS Code
debugger).

?y??o? ?u??a??n??l? ?e??e??e?

?o?? ?y??a? ?n? ?a?? ?y??s ?o??r?? ?l?? ?n? ?u??t??n?

?a??a??e? ?o??i??o??l ?t??e??n??

?a?? ?y??s ?o??s
?/? ?n? ?t??n? ?o??a??i?? ?u??t??n ?efi?i??o??
?y??o?
?u??a??n??l?
?r??r ?a??l??g ? ?e??g??n? ?e??e??e? ?o??l?? ?n? ?a??a??s

?r??E??e?? ?l??k? ?u??t??n ?o??l??

?r??e??c?? ?x??r??l ?a??a??s
?e??g??n? ?o??s ?a??s??c??

Weekly Goals & Checkpoints

• End of Week 1 Checkpoint:
• Write a short script that reads a list of numbers from a file, filters out values
below a threshold, and writes the filtered list to a new file.
• Socratic prompt: “How would you handle a non-numeric entry in your
input file? What exceptions should you catch, and why?”
• End of Week 2 Checkpoint:
• Create a module math_utils.py containing at least three custom functions (e.g.,
mean, median, standard deviation).
• Write a separate script that imports math_utils.py and validates those functions
against known values.
• Questioning mindset: “If my function returns an unexpected result, how
can I be sure whether the bug lies in my logic or in how I called the
function?”
 ?y??o? ?o??m?? ?h??k??i??s

?e?? ?
?h??k??i??
?r??t? ?n?
?a??d??e ?u??o?
?a?? ?u??t??n?

?e?? ?
?h??k??i??
?r??e ? ?c??p? ?o
fi?t?? ?u??e?? ?r??
? fi?e

Active Learning Techniques

• Code katas: Spend 10–15 minutes each day on simple coding problems (e.g., from
HackerRank or LeetCode Easy) focused on loops and data structures.
• Peer review: Pair up (virtually or in person) with a classmate and swap your
math_utils.py. Critique each other’s code for clarity, edge cases, and Pythonic style.
• Mini-quiz: Write down three “what-if” questions each day, such as “What if I pass a list
to a function expecting a single integer?” Then, answer them by writing small code
snippets and observing behavior.

Week 3 & 4: Scientific Python Stack (NumPy, Matplotlib, and Data

Handling)
Why this matters
Numerical efficiency is non-negotiable when simulating neuron dynamics. If you naively use
Python lists instead of NumPy arrays, your simulation can take hours instead of seconds.
Likewise, visualizing data correctly sets the stage for interpreting neuron voltage traces or
synthetic EEG signals.
Topics to Cover
1. NumPy
• Creating arrays ([Link], [Link], [Link]).
• Indexing, slicing, and boolean masks.
• Skeptical note: “Is my slice copying or referencing? What happens if I
modify the original array?”
• Element-wise operations vs. matrix operations.
• Linear algebra basics ([Link], [Link]).
 2. Matplotlib
• Basic plotting: line plots, scatter plots, subplots.
• Plot customization: labels, legends, colors (but keep things simple unless you
need a specific palette).
• Saving figures to disk vs. inline display (Jupyter notebooks).
3. Data handling with pandas (intro)
• Creating DataFrames, reading CSV files.
• Basic statistics (.mean(), .std(), .describe()).
• Plotting with pandas’ built-in .plot() as a quick check.
4. Project structure & version control
• Organizing scripts/modules in a folder.
• A quick introduction to Git: git init, commit, push to GitHub or GitLab.
• Why version control is crucial—imagine losing your neuron model because you
overwrote the code!

?a??e??n? ?c??n??fi? ?y??o?

?e??i?? ?o??r??
? ?r??n??e ?n? ?a??g? ?o?? ?s??g ?i??

?a??a? ?a?? ?a??l??g

? ?a??l? ?n? ?n??y?? ?a?? ?s??g
?a??F??m??.

?a??l??l?? ?l??s
? ?r??t? ?n? ?u??o??z? ?l??s ?o? ?a??
?i??a??z??i??.

?u??y ?a??c?
? ?e??n ?o ?r??t? ?n? ?a??p??a?? ?r???s ?o?
?c??n??fi? ?o??u??n??

Weekly Goals & Checkpoints

• End of Week 3 Checkpoint:
• Write a Jupyter notebook that:
1. Generates a NumPy array representing 0 to ?π in 0.01 increments.
2. Computes its sine and cosine.
3. Plots both curves on the same figure with appropriate legends and axis
labels.
4. Saves the figure to sines_cosines.png.
• Reflective prompt: “What happens to performance if I change the step
size to 0.0001? How long does the operation take, and why?”
• End of Week 4 Checkpoint:
• Find (or create) a CSV file containing at least 100 rows and 4 columns (e.g.,
synthetic temperature readings).
• Use pandas to read it, compute summary statistics, and plot at least two
different charts (line and scatter).
• Push your entire project (notebook + CSV) to a GitHub repo.
 • Question your choices: “Is my CSV format optimal for future analyses?
Could a binary format (HDF5) be faster? If so, why am I sticking to CSV
anyway?”

?y??o? ?o??m?? ?h??k??i??s

?efl?c??v? ?r??p?
?e?? ?
?efl?c??v? ?r??p?
?v??u??e ?S? ?e?? ?
?o??a? ?n?
?o??i??r
?l??r??t??e?
?e??o??a??e
?m??c? ?f ?t?? ?i??

?e?? ?
?h??k??i??
?e?? ?S?? ?o??u??
?t??i???c?? ?n? ?l??
?e?? ?
?h??t?
?h??k??i??
?e??r??e ?u??y
?r??? ?n? ?l??
?i??/??s??e ?u??e?

Active Learning Techniques

• Guided debugging session: Intentionally introduce a bug (e.g., shape mismatch in a
NumPy operation) and time yourself: how do you isolate and fix it?
• Flashcards (Anki): Create cards for key NumPy functions and their signatures.
• “Explain to a rubber duck”: After writing a plot, pretend you’re teaching a friend why
you chose certain axis limits or why vectorized operations matter.

Week 5 & 6: Introduction to Computational Neuroscience and Basic

Neuron Models
Why this matters
Before you dive into full-blown simulations, you need to understand the mathematics behind
neuron models. The venerable Hodgkin-Huxley equations or the simpler Integrate-and-Fire
(I&F) model form the foundation. If you rush through math, your Python code might
reproduce results but fail to teach you why certain phenomena (e.g., action potential
thresholds) emerge.
Topics to Cover
1. Biological background (light overview)
• Membrane potential, resting potential, action potentials.
• Ion channels, gating variables.
2. Simple electrical circuit analogy
 • RC circuit representation of a membrane.
• Capacitance (C_m) and conductance (g_L, g_Na, g_K).
3. Leaky Integrate-and-Fire model (LIF)
• Differential equation:
• τmdVdt=−(V−Vrest)+RmIext \tau_m \frac{dV}{dt} = -(V - V_{rest}) + R_m I_{ext}
• Implementing Euler’s method or a higher-order integrator (e.g., Runge–Kutta).
4. Hodgkin-Huxley (if time allows)
• Four coupled ODEs; gating variables m, h, n.
• Implementing from scratch vs. using a library (e.g., Brian2).
5. Introduction to Brian2 (optional)
• Installing Brian2 via pip install brian2.
• Writing a minimal script: define neuron equations, simulate, plot.

?u??d??g ?e??o? ?o??l?

?r??n? ?n??o??c??o?
? ?s??g ?r??n? ?o? ?i??l??i??s

?o??k??-??x??y ?o??l
? ?x??o??n? ?o??l?? ?e??o? ?y??m??s

?I? ?o??l
? ?m??e??n??n? ?h? ?e??y ?n??g??t??a??-??r?
?o??l

?i??u?? ?n??o??
? ?e??e??n??n? ?e??r??e? ?s ?C ?i???i??

?i??o??c?? ?a??g??u??
? ?n??r??a??i?? ?e??r??e ?o??n??a? ?n? ?o?
?h??n??s

Weekly Goals & Checkpoints

• End of Week 5 Checkpoint:
• Derive (on paper) the LIF equation and implement a simple Euler rollout in
Python that simulates one neuron receiving a constant current.
• Plot the membrane voltage V(t)V(t) over a 100 ms interval.
• Critical question: “How stable is Euler’s method here? What happens if I
halve my timestep? Double it?”
• End of Week 6 Checkpoint:
• Option A (If you feel confident):
• Implement the Hodgkin-Huxley model from scratch, numerically
integrate for a 50 ms stimulus, and reproduce a published action
• Compare your plot to a reference (e.g., the original Hodgkin & Huxley
potential shape.
1952 data).
 • Reflect: “What assumptions did Hodgkin & Huxley make that may
not hold for mammalian neurons? Why can’t I just copy their
parameters?”
• Option B (If you want to prioritize speed):
• Install Brian2 and replicate a standard HH simulation in under 20 lines of
code.
• Modify the code to simulate two coupled neurons (simple synaptic
coupling).
• Question: “When two neurons are coupled, under what conditions
does one entrain the other? Are my parameter choices
biologically plausible?”

?o??u??t??n?? ?e??o??i??c? ?e??n??g ?a??

?e?? ? ?p??o? ?
?n??a?? ?r??n? ?n?
?e??i??t? ?H
?i??l??i??

?e?? ? ?p??o? ?
?m??e??n?
?o??k??-??x??y
?o??l ?r?? ?c??t??
?e?? ?
?h??k??i??
?e??v? ?I?
?q??t??n ?n?
?m??e??n? ?u??r
?o??o??

Active Learning Techniques

• Mathematical journaling: Write out one section of the Hodgkin-Huxley derivation in
your own words (not just copy-paste).
• Peer Socratic dialogue: Exchange your code with a friend and challenge each other:
“How would you add noise to this model?” “Is my refractoriness implemented
correctly?”
• Parameter sweeps: For the LIF model, run your simulation across a range of input
currents I_ext and plot firing rate vs. current (“f–I curve”). Notice where the threshold
lies.

Week 7 & 8: Small Network Simulations and Synthetic Brainwave

Generation
Why this matters
Single-neuron dynamics are enlightening, but real brainwaves arise from large populations.
By the end of Week 8, you should be able to simulate a small recurrent network, extract a
proxy “local field potential,” and generate synthetic brainwave traces (e.g., alpha, beta
Topics to Cover
rhythms).
1. Network connectivity basics
• Defining neuron populations, excitatory vs. inhibitory.
• Connectivity matrices and random networks (Erdős–Rényi, small‐world
approximations).
2. Simulating a small network
• Extend your HH or LIF code to 100–1,000 neurons.
• Track each neuron’s membrane potential (memory and speed considerations!).
3. Estimating Local Field Potential (LFP)
• LFP ≈ sum of synaptic currents or average membrane voltage across neurons.
• How synchronous firing produces oscillatory signals.
4. Synthetic EEG/brainwave generation
• Filtering network activity to specific frequency bands (alpha 8–12 Hz, beta 13–30
Hz).
• Using simple band-pass filters (SciPy’s signal processing module).
5. Intro to MNE-Python or similar EEG toolbox (optional)
• Installing MNE (pip install mne).
• Loading a sample EEG dataset (MNE provides test files).
• Plotting real EEG traces for comparison.

?u??d??g ?e??a? ?e??o?? ?i??l??i??s

?s? ?N??P??h??
? ?n??a?? ?n? ?s? ?N??P??h?? ?o? ?E?
?n??y??s?

?e??r??e ?E?
? ?i??e? ?e??o?? ?c??v??y ?o ?e??r??e
?y??h???c ?r??n??v??.

?s??m??e ?F?
? ?n??r??a?? ?o? ?y??h??n??s fi?i?? ?r??u??s
?s??l??t??y ?i??a??.

?i??l??e ?e??o??
? ?x??n? ?e??o? ?o??l? ?o ?i??l??e ?a??e?
?e??o??s?

?e??o?? ?a??c?
? ?e??n ?b??t ?e??o? ?o??l??i??s ?n? ?o??e??i??t?
?a??i??s?

Weekly Goals & Checkpoints

• End of Week 7 Checkpoint:
 • Create a 100-neuron LIF network (80 EX, 20 IN), random connectivity (10%
connection probability).
• Run a 500 ms simulation with Poisson input. Plot at least three sample
membrane potentials.
• Skeptical challenge: “If I double the network size, does my code scale? At
what point does performance become impractical on your laptop?”
• End of Week 8 Checkpoint:
• Compute a synthetic LFP by averaging synaptic currents over all excitatory
neurons.
• Filter that LFP to isolate a narrow alpha band (8–12 Hz) and plot the raw vs.
filtered signal.
• Forward-thinking note: “If I wanted to emulate an epileptic seizure, how
could I modify connectivity or external drive? Is my network physically
plausible?”

?o??u??t??n?? ?e??o??i??c? ?r??e?? ?i??l??e

?l?? ?a? ?s?

?i??e??d ?i??a?
?o??a?? ?h? ?a? ?i??e? ?F?
?n? fi?t??e? ?F?
?s??a?? ?h? ?l??a
?i??a??
?a?? ?8??2 ?z?
?r?? ?h? ?F?
?e?? ?
?h??k??i??
?l?? ?e??r??e
?o??u?? ?
?o??n??a??
?y??h???c ?F? ?y
?v??a??n? ?y??p??c ?i??a??z? ?a??l?
?u??e??s ?e??r??e
?o??n??a?? ?r?? ?h?
?u? ?i??l??i?? ?i??l??i??
?i??l??e ?h?
?e??o?? ?o? ?0?
?e?? ?
?s ?i?? ?o??s??
?h??k??i??
?n??t
?r??t? ? ?0??
?e??o? ?I?
?e??o?? ?i??
?a??o?
?o??e??i??t?

Active Learning Techniques

• Group discussion or forum post: Describe your network model parameters (e.g., time
constants, connectivity) on a forum like NeuroStars or a class Slack channel. Ask peers:
“Do my numbers make sense biologically?”
• Hands-on filtering exercise: Write your own Butterworth band-pass filter in Python
(SciPy) instead of copy-pasting. Compare its output to a built-in function.
• Coding challenge: Time yourself to implement a network simulation both in “pure
Python” and then vectorized with NumPy/Brian2. How big is the speedup? Document
your results in a short report.
Week 9: Signal Processing Deep Dive (EEG-Style Analysis)
Why this matters
Simulating brainwaves is one thing; analyzing real signals is another. In Week 9, focus on the
analytical side: spectral analysis, event-related potentials, and time–frequency
decomposition. This sets you up for juxtaposing synthetic vs. real data or for later research
Topics to
involving realCover
EEG.
1. Basics of digital signal processing (DSP)
• Sampling theorem, Nyquist frequency, aliasing.
• Windowing functions (Hann, Hamming).
2. Fourier analysis
• Fast Fourier Transform (FFT) in NumPy.
• Power spectral density estimation (e.g., Welch’s method in SciPy).
3. Time–frequency methods
• Short-time Fourier Transform (STFT).
• Continuous wavelet transform (CWT) overview (use PyWavelets if desired).
4. Event-Related Potentials (ERPs)
• Averaging multiple trials to extract stimulus-locked signals.
• Basic epoching in MNE or custom code.

?o??r??e??i?? ?i??a? ?r??e??i?? ?v??v??w

?a??c? ?f ?S?
?u??a??n??l ?o???p?? ?i?? ?a??l??g ?n? ?l???i??

?o??i?? ?n??y??s
?e??n??u?? ?o? ?r??u??c? ?o??i? ?n??y??s

?i??-??e??e??y ?e??o??
?e??o?? ?o? ?n??y??n? ?i??a?? ?n ?o?? ?i?? ?n?
?r??u??c?

?v??t??e??t?? ?o??n??a??
?e??n??u?? ?o? ?x??a??i?? ?t??u??s??o??e? ?i??a??

Weekly Goal & Checkpoint

• End of Week 9 Checkpoint:
• Take either your synthetic LFP from Week 8 or download an open EEG dataset
(e.g., EEG motor imagery dataset).
• Compute and plot:
1. The raw time series (30 s segment).
2. Its power spectral density (Welch’s method).
3. A spectrogram (STFT) highlighting dominant frequencies.
• Write a one-paragraph analysis: “Which frequency bands dominate, and why
might that be the case?”
 • Reflective query: “If I see unexpected peaks, is that noise, artifact, or a
real oscillation? How can I determine that?”

?e?? ? ?h??k??i?? ?e??e??e

?efl?c??v? ?u??y
?o??i??r ?h? ?r??e ?n??y??s
?a??r? ?f
?n??y?? ?n? ?r??e
?n??p??t?? ?e??s
? ?a??g???h ?n
?o??n??t
?o??u??
?r??u??c??s
?p??t??g??m
?o??u?? ?o??r
?a??u??t? ?n? ?l??
?p??t??l ?e??i??
?h? ?p??t??g??m
?s??g ?T?? ?a??u??t? ?n? ?l??
?h? ?o??r ?p??t??l
?o??u?? ?a? ?e??i?? ?s??g
?i?? ?e??e? ?e??h?? ?e??o?
?a??u??t? ?n? ?l??
?h? ?a? ?i?? ?e??e?
?a??
?h??s? ?a??s??
?e??c? ?i??e?
?y??h???c ?F? ?r
?p?? ?E? ?a??

Active Learning Techniques

• DSP debugging challenge: Generate two sine waves at different frequencies, add
noise, and ensure your spectral analysis correctly identifies the peaks.
• Flashcard reinforcement: Create Anki cards for “Nyquist frequency,” “Hamming vs.
Hann window,” and “Welch’s method basics.”
• Peer teaching: Explain to a classmate (or 5-minute video recording) how to compute
an ERP from raw EEG.

Week 10: Advanced Simulation Libraries and Optimization

Why this matters
By now, you have hand-coded many models, but real research often demands more
features and efficiency. Week 10 explores specialized libraries (Brian2, NEURON, or NEST) and
introduces code optimization strategies so you can scale up to thousands of neurons or
longer recording times.
Topics to Cover
1. Brian2 deep dive (if chosen in Week 6)
• Custom neuron equations, creating plasticity rules (e.g., STDP).
• Code generation targets: Python vs. C++ backends.
• Question: “Is my model fast enough in pure Python, or do I need C++?
What are the trade-offs?”
2. NEURON simulator introduction (optional)
 • Installing NEURON via pip install neuron or using nrniv.
• Building a multi-compartment model (e.g., a passive dendrite attached to a
soma).
• Connecting multiple NEURON cells and running a network (high-level template).
3. Code profiling and vectorization
• Using Python’s timeit module or %timeit in notebooks.
• Profiling with cProfile to find bottlenecks.
• Refactoring loops into vectorized NumPy operations or leveraging JIT
compilation (Numba).
4. Parallelization basics
• Python’s multiprocessing vs. multi-threading limitations (GIL).
• Leveraging Brian2’s parallel options or NEURON’s MPI interface (overview; you
don’t need a supercomputer yet).

?o??u??t??n??
?e??o??i??c? ?e??n??g
?a??
?r??n? ?e?? ?o?? ?r?fi?i??
?i?? ?n?
?e??o??z??i??
?x??o??n? ?u??o?
?e??o? ?q??t??n? ?s??g ?o??s ?o
?n? ?l??t??i?? ?p??m??e ?o??
?u??s ?e??o??a??e

?E??O? ?a??l??l??a??o?
?i??l??o? ?a??c?
?n??o??c??o?
?n??r??a??i??
?n??a??i?? ?n? ?u??i??o??s??n?
?u??d??g ?u??i? ?n? ?a??l??l
?o??a??m??t ?p??o??
?o??l?

Weekly Goal & Checkpoint

• End of Week 10 Checkpoint:
• Take your Week 8 network (100 LIF neurons) and:
1. Profile the runtime for a 1 s simulation.
2. Refactor your code to reduce runtime by at least 50%. Document what
you changed.
3. If you used Brian2, switch from the “Python” code generation target to
“C++” and measure the speedup.
• Critical reflection: “Was the 2× speedup worth the complexity? If I scale
my network to 1,000 neurons, will this still hold?”
 ?e?? ?0 ?h??k??i?? ?e??e??e

?r??i??l ?efl?c??o?
?o??i??r ?h? ?e??u?? ?p??d??
?r??e??ff? ?n?
?v??u??e ?h? ?p??d
?c??a??l??y ?f ?h?
?m??o??m??t ?f??r
?h??g??
?w??c??n? ?o ?+?
?w??c? ?o ?+?
?o??m??t
?h??g? ?h? ?o??
?h??g??
?e??r??i?? ?a??e?
?o ?+? ?e??r? ?h?
?h??g?? ?a??
?u??n? ?e??c??r??g
?e??c??r ?o??
?p??m??e ?h? ?o??
?o ?e??c? ?u??i??
?r?fi?e ?u??i??
?e??u?? ?h? ?n??i??
?u??i?? ?f ?h?
?e??o?? ?i??l??i??

Active Learning Techniques

• Code profiling sprint: Time yourself to run profiling, interpret the results, and refactor
within one session.
• Mini hackathon: Challenge yourself to implement an STDP (spike-timing-dependent
plasticity) rule in Brian2 and demonstrate how synaptic strengths change over time.
• Discussion board post: Summarize your findings about pure Python vs. Brian2 (C++)
for small networks. Ask for feedback about scaling to larger systems.

Week 11 & 12: Capstone Project – Integrate Simulation and

Brainwave Plotting
Why this matters
A structured capstone will solidify all previous work. By the end of Week 12, you should have
a self-contained Python project that:
1. Simulates a small network with biologically plausible parameters.
2. Computes synthetic LFP/EEG-like signals.
3. Performs time–frequency analysis.
4. Generates publication-quality figures (e.g., raster plots, power spectra, spectrograms).
This portfolio piece can be repurposed for research, future classes, or even as a job demo.
Overall Project Structure
• models/ directory
• Contains Python modules for neuron models (e.g., [Link], [Link]), and network
topology ([Link]).
• simulations/ directory
 • Higher-level scripts or Jupyter notebooks that run specific experiments (e.g.,
“rhythmic drive,” “altered connectivity”).
• analysis/ directory
• Scripts or notebooks for DSP: filtering, PSD, spectrogram, and ERP analysis.
• figures/ directory
• Automatically generated plots (with clear filenames like figure1_raster.png,
figure2_spectrum.png).
• [Link]
• Overview of the project, instructions on how to reproduce results,
dependencies, and citations for any libraries or neuroscience papers.

?r??e?? ?t??c???e ?o? ?o??u??t??n?? ?e??o??i??c?

?E??M??m?
?n??y??s?
?r??e?? ?v??v??w
fi?t??i??
?e???d??t??n ?n??r??t??n?
?S?
?e??n???c??s
?p??t??g??m
?i??t??n?
?R? ?n??y??s
?r??e??
?t??c???e
?o??l??
?i??l??i??s?
?i??p?
?h??h??c ?r??e
?h??y
?l??r?? ?o??e??i??t?
?e??o??.??

fi?u??s?

fi?u??1??a??e??p??
fi?u??2??p??t??m??n?

Week 11: Assemble and Simulate

• Day 1–2:
• Draft a project outline (tasks, dependencies, timeline).
• Choose whether you’ll use hand-coded Python or a library (Brian2/NEURON) for
your final simulation.
• Forward-thinking: “If I aim to publish in a computational neuroscience
journal, which toolchain do I need to learn next?”
• Day 3–5:
• Implement or refine neuron/network modules based on your Week 10
optimizations.
• Run a simulation that:
1. Creates a 200-neuron network (80 EX, 20 IN).
 2. Applies a time-varying external current to induce oscillations (e.g., a 10
Hz sinusoidal drive).
3. Records all membrane potentials and synaptic currents.
• Weekend:
• Debug any large memory or speed issues.
• Ensure reproducibility: fix random seeds, document all parameters in a YAML or
JSON file.
• Quality check: “Can someone else clone my repo and run python
simulations/run_network.py to see consistent results?”
• End of Week 11 Checkpoint:
• You have a working simulation that outputs (to disk) a time series of LFP-like
data and individual neuron voltages.

?a??t??e ?r??e?? ?e?? ?1? ?i??l??i?? ?s??m??y

?e??o?? ?o??i??
?m??e??n??t??n ?i??l??i??
? ?i??l??i?? ?h??k??i??
?m??e??n? ?e??o? ?o??l??e ?o??i??
?e??o??, ?p??y ?i??l??i?? ?i??
? ?x???n?? ?r??e? ?n? ? ?F??l??e ?a??
?e??r? ?a?? ?u??u?

?n? ?f
?a? ?–? ?a? ?–? ?e??e??
?e?? ?1

?r??e?? ?e??g??n? ?
? ?
?u??i?? ? ?e??o??c??i??t?
?o?? ?i? ?e??r??s??e?
?e??c??o? ?s??e? ?n? ?n??r?
?e???d??i??e
?r??t ?r??e?? ?e??l??
?u??i?? ?n? ?h??s?
?i??l??i??
?o??c??i?

Week 12: Analyze, Visualize, and Present

• Day 1–3:
• Analysis: Load your LFP data and:
1. Compute power spectral density (PSD).
2. Generate a spectrogram (STFT) over the entire simulation.
3. Identify the dominant frequency band and relate it to your external drive.
• Question to answer: “What happens if I change the external drive to 20
Hz? Do I observe a 20 Hz peak, or does network resonance alter it?”
• Day 4–5:
• Visualization: Produce at least three publication-quality figures:
 1. Raster plot of all spikes in the network.
2. LFP time series vs. filtered band (alpha/beta).
3. PSD with annotations showing peaks.
• Write detailed captions for each figure as if preparing for a journal submission.
• Self-critique: “Is my font size legible? Are my axis labels clear? Would a
reviewer understand the plot without reading my code?”
• Weekend:
• Wrap-up and reflection:
1. Write a short project report ([Link] or markdown) summarizing
objectives, methods, results, and next steps (e.g., adding more realistic
synapse models or using real EEG data to validate).
2. Give a 10–15 minute demo presentation to a peer or record yourself
explaining the project.
• Forward-thinking: “If I wanted to extend this for my final year project or a
conference poster, what would be my hypotheses? Would I compare a
real EEG dataset against my synthetic LFP?”
• End of Week 12 Checkpoint:
• A complete GitHub repository containing:
• All code modules, notebooks, and scripts.
• Generated figures in figures/.
• A concise report or README explaining how to run everything.
• Video link or slides from your 10–15 minute demo.

?a??t??e ?r??e?? ?e?? ?2? ?n??y??s? ?i??a??z??i??, ?n?

?r??e??a??o?
?i??a??z??i?? ?e??s??o??
?u??i??i??
?r??t? ?u??i??t??n?
?u??i?? ?a??e? ?l??, ?u??i? ?o??l??e
?F? ?i?? ?e??e?? ?i??u? ?e??s??o??
?n? ?n??t??e?
? ? ?i?? ?o??, fi?u??s?
?S? ?e??r?? ?n? ?e??

?n? ?f
?a? ?–? ?a? ?–? ?e??e??
?e?? ?2

?a?? ?r??-?? ?
? ?
?n??y??s ?efl?c??o?
?o?? ?F? ?a??, ?r??e ?r??e??
?o??u?? ?e??r?? ?r???r?
?S??s??c??o???m? ?n? ?e??v?? ?e??
?d??t??y ?o??n??t ?r??e??a??o?
?r??u??c?

Active Learning Techniques Throughout the 12 Weeks

1. Socratic journalingAt the end of each coding session, write a short paragraph (3–5
sentences) summarizing:

• What you did.

• Why you chose this approach.
• One assumption you made and how you might test it tomorrow.
2. Weekly peer check-insArrange a 15-minute “check-in” (virtual or in person) with a
study partner. Each person presents:

• One challenge they faced.

• One “aha” moment they had.
• One question that still bugs them.
3. Mini-projects & coding challengesSprinkle in short side tasks, such as:

• “Implement a FitzHugh–Nagumo model and compare its spiking behavior to

Hodgkin-Huxley.”
• “Use PyWavelets to detect transient oscillations in your synthetic LFP and mark
them on a plot.”
• These keep the journey fresh and force you to adapt your knowledge on the fly.
4. Cheat-sheet creationMaintain a living document (Google Doc or local .md) listing:

• Frequently used functions, code snippets, and “gotchas.”

• Brief neuroscience definitions (e.g., what is a ‘refractory period?’).
• This serves both as study reinforcement and a quick reference.
5. Reflective “Retrospective” every two weeksBefore starting Week 3, 5, 7, 9, and 11,
take 30 minutes to reflect:
• Are you meeting your daily time-allocation goals?
• Which concept was most confusing? How will you re-address it?
• Where do you see this skillset taking you in 6 months, 1 year, or beyond?
 ?a??e??n? ?o??u??t??n?? ?e??o??i??c?

? ?e??o??e??i??
?efl?c??o?
?v??u??e ?r??r??s ?n? ?u??r?
?o??s ?v??y ?w? ?e??s?

? ?h??t??h??t ?r??t??n
?o??i?? ? ?e??r??c? ?u??e ?o?
?u??k ?c??s? ?o ?e? ?n??r??t??n?

? ?i??-??o??c??
?n??g? ?n ?h??t ?a??s ?o ?p??y
?n? ?d??t ?n??l??g??

? ?e?? ?h??k??n?
?h??e ?h??l??g?? ?n? ?n???h??
?i?? ? ?t??y ?a??n??.

? ?o??a??c ?o??n??i??
?efl?c? ?n ?o??n? ?e???o?? ?o
?n??r??a?? ?h??c?? ?n?
?s??m??i??s?

Tips to Stay Engaged and Avoid Burnout

1. Chunk your study sessions
• If you have 1.5 hours on a weekday, break it into two 45 minute blocks:
1. Review & question yesterday’s work.
2. Learn new material and code.
• A short 5 minute walk or stretch between blocks prevents mental fatigue.
2. Ask “What if?” often
• Always question your own code and results.
• Example: “What if my neuron never fires? Is it a coding bug, or is my current
injection sub-threshold?”
3. Celebrate small wins
• Finished Week 4? Treat yourself to something—coffee, a walk, or a short tech
podcast.
• Recognize that each completed checkpoint is progress toward a higher-level
skill set.
4. Reverse teach
• Every few weeks, teach the previous two weeks’ material to a friend or record a
short screencast.
• If you can explain it clearly, you truly understand it. If you stumble, flag those
areas for review.
5. Keep a “parking lot” for future exploration
• Whenever you stumble upon an interesting idea that’s outside the current
week’s scope (e.g., “What about incorporating realistic dendritic branching?”),
note it in a separate section.
 • This keeps you focused on immediate goals while preserving your curiosity for
future projects.

?ff?c??v? ?t??y ?e??n??u?? ?o? ?y??o? ?e??n??g

?h??k ?t??y
?e??i??s
?i??d? ?t??y ?i?? ?n??
?a??g???l? ?l??k?

?e??e?
?e??e??a??s ?o??
?efl?c? ?n ?r??i??s ?a??s
?a??r??l

?e??n ?e?
?a??r??l
?n??g? ?i?? ?e?
?o???p?? ?n? ?o??

?a?? ? ?r??k
?e??e?? ?i?? ? ?h??t ?a??
?r ?t??t??

?s? ?W??t ?f?

?u??t??n ?o?? ?n?
?e??l??

?e??b??t? ?m??l
?i??
?e??r? ?r??r??s ?i?? ?
?r??t

?e??r?? ?e??h
?x??a?? ?a??r??l ?o
?t??r?

?e?? ? ?P??k??g
?o?
?o?? ?d??s ?o? ?u??r?
?x??o??t??n

Final Thoughts (Forward-Looking and Skeptical)

• Continuous iteration: By Week 12 you’ll have a solid portfolio piece, but remember—
mastery is iterative. Revisit your code in 3 months, and you’ll likely spot new
optimizations or more biologically realistic tweaks.
• Question your assumptions: Throughout these 12 weeks, whenever you simulate a
clearly unrealistic phenomenon (e.g., a network that never stops oscillating), dig
deeper rather than glossing over it. That’s where real learning happens.
• Future pathways: Once comfortable with Python simulations and basic EEG analysis,
you could:
 • Transition to real EEG datasets (e.g., BCI experiment data).
• Explore machine learning approaches to classify brain states.
• Dive into more complex neuron models (e.g., multi-compartmental,
morphologically detailed).
• Long-term vision: In two years, imagine presenting at a conference your own cortical
microcircuit simulations paired with lab-recorded EEG. This roadmap is just Step 1.

?r?? ?i??l??i?? ?o ?a??e??

?e??n??r ?x??r?
?o??i??o?? ?x??o?? ?o??l??
?n??i?? ?i??l??i?? ?u??t??n
?x??o??t??n ?t??a??o? ?a??w??s
?s??m??i??s ?o??e??n??
?efi?e ?o?? ?o? ?r??s??i?? ?o ?e?? ?r??e??a??o? ?e??y
?p??m??a??o? ?n??s??g??e ?E? ?a??
?n??a??s???
?h??o??n?

Above all, stay curious, stay skeptical (“Is my model actually capturing something biologically
meaningful, or am I fooling myself?”), and enjoy the creative process. You’ve got this—one
line of Python at a time!