Neuromatch Academy aims to introduce traditional and emerging tools of computational neuroscience to trainees.

Notebook systems are proving invaluable to skill acquisition, research documentation, publication, and reproducibility.  This series of presentations introduces the most popular platform for computational notebooks, Project Jupyter, as well as other resources like Binder and NeuroLibre. 

This course corresponds to the first session of talks given at INCF's Neuroinformatics Assembly 2023. The sessions consists of several lectures, focusing on using the principles of FAIR (findability, accessibility, interoperability, and reusability) to inform future directions in neuroscience and neuroinformatics. In particular, these talks deal with the development of knowledge graphs and ontologies. 

This couse is the opening module for the University of Toronto's Krembil Centre for Neuroinformatics' virtual learning series Solving Problems in Mental Health Using Multi-Scale Computational Neuroscience. Lessons in this course introduce participants to the study of brain disorders, starting from elemental units like genes and neurons, eventually building up to whole-brain modelling and global activity patterns.

The Neurodata Without Borders: Neurophysiology project (NWB, https://www.nwb.org/) is an effort to standardize the description and storage of neurophysiology data and metadata. NWB enables data sharing and reuse and reduces the energy-barrier to applying data analytics both within and across labs. Several laboratories, including the Allen Institute for Brain Science, have wholeheartedly adopted NWB.

The International Brain Initiative (IBI) is a consortium of the world’s major large-scale brain initiatives and other organizations with a vested interest in catalyzing and advancing neuroscience research through international collaboration and knowledge sharing. This session will introduce the IBI and the current efforts of the Data Standards and Sharing Working Group with a view to gain input from a wider neuroscience and neuroinformatics community. 

EEGLAB is an interactive MATLAB toolbox for processing continuous and event-related EEG, MEG, and other electrophysiological data. In this course, you will learn about features incorporated into EEGLAB, including independent component analysis (ICA), time/frequency analysis, artifact rejection, event-related statistics, and several useful modes of visualization of the averaged and single-trial data. EEGLAB runs under Linux, Unix, Windows, and Mac OS X.

Neuroscience has traditionally been a discipline where isolated labs have produced their own experimental data and created their own models to interpret their findings. However, it is becoming clear that no one lab can create cell and network models rich enough to address all the relevant biological questions, or to generate and analyse all the data required to inform, constrain, and test these models.

As models in neuroscience have become increasingly complex, it has become more difficult to share all aspects of models and model analysis, hindering model accessibility and reproducibility. In this session, we will discuss existing resources for promoting FAIR data and models in computational neuroscience, their impact on the field, and remaining barriers.

The dimensionality and size of datasets in many fields of neuroscience research require massively parallel computing power.  Fortunately, the maturity and accessibility of virtualization technologies has made it feasible to run the same analysis environments on platforms ranging from single laptop computers up to high-performance computing networks.

In this course, you will learn about working with calcium-imaging data, including image processing to remove background "blur", identifying cells based on threshold spatial contiguity, time-series filtering, and principal component analysis (PCA). The MATLAB code shows data animations, capabilities of the image processing toolbox, and PCA.

This module provides an introduction to the problem of speech recognition using neural models, emphasizing the CTC loss for training and inference when input and output sequences are of different lengths. It also covers beam search for use during inference, and how that procedure may be modeled at training time using a Graph Transformer Network.

This workshop delves into the need for, structure of, tools for, and use of hierarchical event descriptor (HED) annotation to prepare neuroimaging time series data for storing, sharing, and advanced analysis. HED are a controlled vocabulary of terms describing events in a machine-actionable form so that algorithms can use the information without manual recoding.

This module covers fMRI data, including creating and interpreting flatmaps, exploring variability and average responses, and visual eccenticity. You will learn about processing BOLD signals, trial-averaging, and t-tests. The MATLAB code introduces data animations, multicolor visualizations, and linear indexing.

There is a growing recognition and adoption of open and FAIR science practices in neuroscience research. This is predominately regarded as scientific progress and has enabled significant opportunities for large, collaborative, team science. The efforts and practical work that go into creating an open and FAIR landscape extend far beyond just the science.

Difficulties experienced in understanding machine learning techniques often stem from lack of clarity concerning more basic statistical models and fundamental considerations, including the various regression models that can all be subsumed under the General Linear Model.

The importance of Research Data Management in the conduct of open and reproducible science is better understood and technically supported than ever, and many of the underlying principles apply as much to everyday activities of a single researcher as to large-scale, multi-center open data sharing.

The workshop will include interactive seminars given by selected experts in the field covering all aspects of (FAIR) small animal MRI data acquisition, analysis, and sharing. The seminars will be followed by hands-on training where participants will perform use case scenarios using software established by the organizers. This will include an introduction to the basics of using command line interfaces, Python installation, working with Docker/Singularity containers, Datalad/Git, and BIDS.

This course contains sessions from the second day of INCF's Neuroinformatics Assembly 2022.