Syllabus (2022-2023)
Classroom:
hybrid - Sutherland 554 and Zoom (see OnQ for link)
Day, time:
Winter term, Fridays, 1-4pm
Instructors:
Gunnar Blohm,
Joe Nashed
Tutorial format, Python based. We will use
Google Colab.
Recommended pre-requisite for this course is basic knowledge of
Python - see
this course.
All
teaching materials are available on the Blohm lab Github page.
- Week 1: Introduction (Gunnar)
The research process
Statistics and models in scientific discovery (Pearl)
Study design (power, sample size, effect size)
- Week 2: Intro Python (Joe)
Google Colab interface
Basic syntax and commands
Importing and manipulating data
Graphics
- Week 3: Advanced Python (Joe)
Vectors and Matrices
Functions
- Week 4: Data collection / signal processing (Joe)
Data types
Sampling
DAQ
Filtering (noise, differentiation, integration)
Time vs frequency analysis
- Week 5: Statistics and Hypothesis testing - basics
(Joe)
Descriptors: central tendencies (mean, median,
mode), Spread (Range, variance, percentiles), Shape (skew,
kurtosis)
Correlation / regression
The logic of hypothesis testing
Statistical significance
Multiple comparisons
Different test statistics
Confidence intervals and bootstrap
- Week 6: Statistics and Hypothesis testing - advanced
(Joe)
ANOVA (between-subject, factorial,
within-subject/repeated measures)
Measuring effect size
Multiple regression
Non-parametric tests
- Week 7: Quantitative wet lab / bench methods (Joe)
Image processing
- Week 8: Statistics and Hypothesis testing - Bayesian
(Gunnar)
Motivation and pitfalls of classic methods
Conditional probabilities and Bayes rule
Bayes Factor
Maximum A Posteriori (MAP) estimation
Bayesian ANOVA
- Week 9: Models in Neuroscience (Gunnar)
Models in scientific discovery (Pearl)
Usefulness of models
Parameter search (Newton) and model fitting methods
- Week 10: Data Neuroscience overview (Gunnar)
Promises and limitations (Pearl)
Data organization (format, DB)
Blind data processing: machine learning techniques
(classification, dimensionality reduction, decoding)
- Week 11: Correlation vs causality (Gunnar)
What’s causality?
How to achieve causality
Problem of unobserved variables in high-dimensional problems
- Week 12: Reproducibility, reliability, validity
(Gunnar)
Statistical considerations (multiple comparisons,
exploratory analysis, hypothesis testing)
Open Science methods
Open science vs patents (required for drug discovery)
Course evaluation
(Virtual) pre-registration of research plan (i.e. stage 1 of
registered report). Note, an actual pre-registration is not
required (but encouraged); rather it is about generating the
pre-registration materials, i.e. proposal summary, literature
review, justified hypotheses, experimental approach and analysis
plan (statistics).
Preliminary submission (for round of formative feedback,
optional): April 7, 2023
Final due date: April 28, 2023
Follow
Stage
1 instructions of registered reports!
See
Center for Open Science
for more info.
Additional guidelines:
- 10
steps how to structure papers
- please remember to motivate / justify your hypotheses
- if you have multiple hypotheses, please number them in the
introduction and then address how you will test each of the
hypotheses in the methods section
- the methods need to describe specifically and explicitly
how each hypothesis will be addressed
- you do NOT need to submit a cover letter
- general section lengths: see eNeuro
guide for authors (include title, abstract,
significance statement, Introduction, and Methods)
Further
readings:
Check out Statistics
Learning Resources on the Blohm
Lab WIKI.
Please read this about academic
integrity!
