Dimensionality Reduction in Neuroscience
Current announcements
Nothing to announce at the moment.
Course requirements
- Basic knowledge of Python programming, usage of conda or pip for package management, usage of GitHub, and familiarity with Jupyter notebooks.
- A laptop or desktop computer (no specific requirements except an internet connection) with a working Python and conda installation (e.g., minicondaꜛ or Anacondaꜛ)
- Familiarity with fundamental concepts in machine learning and neuroscience is beneficial but not required.
- Please download or clone in advance the course material from the course’s GitHub repository:
- Additionally to the data provided in the data folder of the Github repository, you need to download data from this Google Drive folder. Place the downloaded data in the cloned version of the repository (into the “data” folder).
- During the course, please visit this website to stay up to date (see Current announcements section).
Important note: Before the course starts, please make sure, that conda is working on your device. We can not provide installation or technical assistance during the course.
Trouble shootings: If you have problems with your computer and/or Anaconda, you can use an online Python compiler, e.g., Google Colab ꜛ. Please, ensure before the beginning of the course, that you can access the online compiler of your choice (e.g., create a Google account) and that you know how to operate it (again, during the course we can not provide installation or technical assistance). E.g., you need to place the data in your Google Drive and mount it in the notebook. Here is description of how to do it.
Who should attend?
- Neuroscientists and researchers interested in applying machine learning techniques to their data.
- Data scientists and machine learning practitioners looking to specialize in neuroscience applications.
- Graduate students and professionals in related fields seeking practical skills in dimensionality reduction.
Course format
- Lectures: Each topic will be introduced with a theoretical background and relevant mathematical foundations.
- Hands-on Sessions: Practical exercises using Python and popular libraries (such as scikit-learn and PyTorch) to implement and experiment with each method. We will apply these techniques to neuroscience datasets.
Learning objectives
- Understand the principles and applications of key dimensionality reduction techniques.
- Gain hands-on experience in implementing these methods using Python.
- Learn to interpret and visualize the results of dimensionality reduction in a neuroscience context.
Syllabus
Chapter 1: Introduction to dimensionality reduction in neuroscience
Chapter 2: Principal component analysis (PCA)
Chapter 3: Clustering methods
Chapter 4: t-Distributed Stochastic Neighbor Embedding (t-SNE)
Chapter 5: Uniform Manifold Approximation and Projection (UMAP)
Chapter 6: Excurse: Artificial Neural Networks (ANN)
Chapter 7: Excurse: Activation functions and the vanishing gradient problem
Chapter 8: Excurse: Improving the learning in ANN
Chapter 9: Autoencoders (AE)
Chapter 10: Variational Autoencoders (VAE)
Acknowledgements
In order to keep the lecture material open and free, the use of copyright-protected material has been avoided. Instead, freely accessible sources are embedded and quoted. The organizer of this lecture is not the author of these sources and is not responsible for their content. The organizer uses the embedded sources according to fair principles for educational purpose.
The organizer of this lecture is not affiliated with the authors of the content in the provided external links, and bears no responsibility for their content. These links are used solely for educational purposes.
The data used in this course is from the following public sources:
- “hypothalamus_calcium_imaging_remedios_et_al.mat”: The dataset is from the 2023’s course ‘data analysis techniques in neuroscienceꜛ’ by the Chen Institute for Neuroscience at Caltech, originally from the paper: Remedios, R., Kennedy, A., Zelikowsky, M. et al. Social behaviour shapes hypothalamic neural ensemble representations of conspecific sex. Nature 550, 388–392 (2017), https://doi.org/10.1038/nature23885ꜛ.
- “macosko_2015.pkl.gz”: Extracted from the the datasets available in the openTSEN package. Specifically, it is the Macosko 2015 mouse retina data set.
- “hippocampus_achilles”: Extracted from the datasets available in the CEBRA packageꜛ.
Further Readings
- Benyamin Ghojogh, Mark Crowley, Fakhri Karray, Ali Ghodsi, Elements of Dimensionality Reduction and Manifold Learning, 2024, Springer, ISBN: 9783031106040
- Frederic Ros, Rabia Riad, Feature And Dimensionality Reduction For Clustering With Deep Learning, 2024, Springer, ISBN: 9783031487422
- Alice Zheng, Amanda Casari, Feature Engineering for Machine Learning – Principles and Techniques for Data Scientists, 2018, O’Reilly Media, ISBN: 9781491953242
- B. K. Tripathy, Anveshrithaa S, Shrusti Ghela, Unsupervised Learning Approaches for Dimensionality Reduction and Data Visualization, 2021, CRC Press Book, ISBN: 9781032041032
- Loring W. Tu, An Introduction to Manifolds, 2011, Springer New York, ISBN: 9781441973993
- Ian Goodfellow, Yoshua Bengio, and Aaron Courville (2016), Deep Learning, MIT Press, free online accessꜛ
- John P. Cunningham, & Byron M. Yu, Dimensionality reduction for large-scale neural recordings, 2014, Nat Neurosci, Vol. 17, Issue 11, pages 1500-1509, doi: 10.1038/nn.3776ꜛ
- R. B. Ebitz, B. Y. Hayden, The population doctrine in cognitive neuroscience, 2021, Neuron, Vol. 109, Issue 19, pages 3055-3068, doi: 10.1016/j.neuron.2021.07.011ꜛ
Past courses
- 2024, October: DZNE Workshop series (1.5 days)
This course material is under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License (CC BY-NC-SA 4.0).
comments