Biotechnology for Neuroscience

The course aims to provide at-the edge statistical tools for exploring and analysing multimodal data. Using programming routines (R) it will introduce statistical methods for qualitatively and quantitatively testing experimental hypotheses.

Students will be able to exploit scientific analytic methods to query large multimodal datasets. They will use data analytics skills to provide constructive guidance in decision making. They will be able to interpret the results correctly with detailed and useful information.

Classroom lectures and practical sessions in the computer room.

Written test; optional oral test.

• Introduction to data science
• Programming for data analysis
• Introduction to R
•  Principles of reproducible computational research
•  Pipelines and build systems
• Advanced statistical methods
•  Inference, hypothesis testing and multiple testing
•  Resampling and bootstrapping
•  Analysis of causality
• Statistical learning and deep learning
  •  Linear regression
  •  Logistic regression
  •  Penalized regression
  •  Support vector machines, random forests, and neural networks
  •  General issues in statistical learning: bias-variance decomposition, the curse of dimensionality, overfitting, and cross-validation
  • Principal component analysis 
  •  Clustering
  •  Modern techniques for dimensionality reduction and clustering
  • Graph theory and  network science
• Supervised learning: regression and classification
• Unsupervised learning and dimensional reduction
• Data harmonization
• Introduction to Big Data analysis
• Meta and Mega analysis techniques
• ComBAT method

An Introduction to Statistical Learning with Applications in R, Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani